From 1c903403adfa106f98ca37f7fc89396a04c16d5b Mon Sep 17 00:00:00 2001 From: <> Date: Sat, 18 Nov 2023 12:57:42 +0000 Subject: [PATCH] Deployed b22988b with MkDocs version: 1.5.3 --- assets/arete.diagram.light.png | Bin 185607 -> 204659 bytes assets/arete.diagram.png | Bin 214978 -> 237525 bytes assets/arete.drawio.xml | 276 ++++++++++++++++++--------------- index.html | 2 +- search/search_index.json | 2 +- sitemap.xml.gz | Bin 127 -> 127 bytes usage/index.html | 17 ++ 7 files changed, 173 insertions(+), 124 deletions(-) diff --git a/assets/arete.diagram.light.png b/assets/arete.diagram.light.png index b9d2e7d64f02efb0c3997c3ba6c195f54d4249fe..c2d2a4590aaead2f3204aa93ee5b9473e5a5ba88 100644 GIT binary patch literal 204659 zcmeEP2OyOH8;=xPq`gFu5OK?0QTE;|RCG98#&JeSqEy-{(NH7}>KAR5N_%K8X;Ntq zY488@-o0Plog@E3NZs$3d%f@b9nbT8p85TLXWLqvbnei%Lz^~jI-8jq+P7)b*0xO> z(Q^{*;mU?KmAl%AEEy4IB9_Qv{57)aDP%HX!x;0!xV{ziUa~ZI9T1AN%v*b z14Gq=7-4V;ejmv4X8JH0-rRc#8U&)6rly)E)lrk6NYo>0!T-n(QamUm?lpDAprmmq3zcFEW z`Lp0jf`+CBd^F*bF_X>kVZmd@nwoGIw>d@_-Iv=nCxV}c`9u>3s)>!GpQo9zwVq!P z_nrs_JCw-^!bd<**HkBRe-RcH%-~-2Wrq2Mdtz6KMv5AG&>Hq38sPH*UxMg?_`S9; zL^bFTHakpwUTmy`sWwJLXIFbSCu&fLqZ!%NPVjB$oq>@iYZ!PuBXu}14cu3vPYM5C$`NPHt8BJBD9 z=7!slP@)LT5UtLl#ou}84|gu=GH#*)+zzmeSgZDD@VhyP@~b{so&(u$bs3JlWg9|bR=R_PXs>yEs z!v%>1h81oT;H?s=NMGY{*t8LYtij*?s!99*;DRG@O;ViAJfnzIFRzJej#NK(P=J3V zK}%mv3#Uuqu1$k9G z1`NNp(Kzt^dw?vQg1F(`XcWXPxW7p}3W1+^;AC-mrYTd;JjBM*!A>*U$Slx5$lsad z;_vUK*4XOQ$A>}j5~@zUkyO?weFAoNbJvnYWUnDDNx~HXGDz!ROIG7sB1xUhop01= z(^is#bEZkeyZ>wCYaTYQP$9+b{Mr!q&pccKENvCipPnxm1RN(H7gEjY5>x+TN+R-m zcAA>aOiW3buwh|<>a*6O1r_4FoME?0Di$Pf+!&RhK}{tKsOrW(iaxX`pJe^h^jH9oJ7OX3TZJ74Wv4GxC+H|c?G@cMWaxfl#i_h zDEo7Uv4o#68dM|t7(5$?#SjoZ=G;U=EoiSuAW|uMGyqS8 z3DwdA1dVF;*>6NGQ4(I0-&Q9G!GtE4G7w6W)`l@;=NQ*$xH>6e( zP$0rwKjh<)E#wjNLN>bAIF%e^tUL_Uhq$gMq2`gssDCG!Oyw^gArtZMsL7`7B9jF$ zvueivKL?hJ%tFdI$S6&XQ-rM9!a~|YvxcO`#(eQ^^|yr|FEFgC zuc;lI@w@A>`lSwjNIgr4u_n!cMgu-(L^8u%Ak2ix2llfo{EP)|xZeyhuK{Cs>{o`E zZ6T!M$%pF2;8bJOC;?n{fT!UOCdJV(+)vBL-qV#4R_aJV9yD&f5e3VX>tCL#3Il6uviS-VUPaZ2`o>s~Lgm!_HKx81 zEVrQJeW+0ie<@rf+?Z-%(nS_1k!l3#(n5up%~`rUBW$DT&i0((!A_)HwH*<$t1%e2n`*wV5`5 zn3v@DrG=!Y!M(2HZ8qjWa~Jx8jztL!Aue#WoW^fh?3gP8WTl|;5yVf;A%P*(mm$|- z)k=N4eaKK1Kqa4F8jrj#oY2uAQ<0t+#P zwb~r?xN!c)4V1sE$baox)uZAj6Zn(HWMq}`q_vt{<$+M@5pd@Jm#9ZVReH!HMa8hq zk9xvtMlpEfDch`Pv+LJ#*1}Wiyqpntv}Ip(w_Q+y6$ii-%qaG5=n@Ck5H%NS!1990eNc z-~FQ^#zkR1Ej+D@n*43dLgfGrx^tIc_sM9+T1ustP-fTs3t^iYBAWkPo`s446A27a zKwug`M#=nEK2fmA-TV@$Fh<1qw*c!MUk3)9VsvG2cH2`Fj z_@iRgnrya^U58vzJqK@pjo2-WUj_H`qo4#`!|^LclYjMFm4A<#{B3@1l<))@#guRf zubQ#{FFb+zT5YR$GOBID-(NL7|KIzY?m{=Ac246*KmIvU7hHKxrfI3uc>i$yFC>Na zuc*i*0l}}m1HTOejTiwja&}@E2N6RZiBT4z=Jr}+d?MW%{yYqU>`7I;u9|E8&+#}B zZJvV6FBE-mIIboMZ0JBiN64Tif18hyifGViJCVYl`-kqMy{B4=!V&6!pqiY9%aOH8 zDflHqUGb_NJa8!`QfP?hkkHh>_8|!r8I$3F7(7&s8g1H=DmZIoScJ@}hG^MxT9<%y zKmZ&YR?~idJbTH}LxZ_ziTN-ivF!*1j-KLKlYuSC;HuLxJUh>M!>Mjw_^J)yQ0ySE zM%j&yMCIRkicH0qoARW(F>QgB;_;@s(KW>2e}5lPv;jh+!EV*^jB3VFRiRdmd50F} z>uU)hV72pk>K}Be1OftCg!T27P-qnMsM?&N?zd>rB{jG^E?^v0?SoNWN`Dz@ZedbN zB=bj*iKto}fq4B(DN$gvQ&=y;9}8&Oh7plmwWm1_GF1a;(NYQr)N(X_!|Bp$()jNM z6a)^MYJ`$wl3+1Ob6lDz)aOQl|Au2&0ku+dWbD^PCoN2^5HzadI>-_eAOR6*5-1mKTVTO%^OP%90uIBZOQ|D})5MGr2J3zbr|3uX&rY)yfDW+S*CEn;`e zT6v^a&Tgz!LX8k!kIND?X?U6dDP>gL-R#l^_&EL<$`J&^4(NNcqugICP8P79stzHk zQS@4v@*tpX9IC{>F?JiTidUmnTa01UlrqsZ@^C$>k7|}cvsE8;^5I&N<|heyC(V|0 zgt|ywkT=xGu_l3kfbY7=Q}e;U)|BgOqnZxj3pnkP84$p;=R|AFxa#b6O@Wh<8pHb# z*l(r&7PB-;xq#q|Rdv6m#EXSS3`lr&(m!LKR|x`c<5ZdJ&CWbWThIP8q@y5Bh|E*; zz1fittwZ`}NT-E7iRMqbIv6$lc1x2yB$K+IvD7L{e`IA=OY;1Cr8END zQjt-30%c6m=tPtp-qcw>15W~Y$rwApA& ziZTe2u4-j>vm+f^Pf?pJ`GPP7*BE!>KhzR#!!V?0hU7l^bk&unWV_nCJUh$gxb zLEhhA212bvS|HSq9t>-GAfyvkhowE9bT5C5LP6d(;b8&HAnZx(U`B5`+utUb?!^ql zj1&!Z4eWeIeprN)O|iRyV#3z^RHwL5G_K+x20M&_pSQv(w)l+=Eb+(jDl8uIA_JI@ zx`^aWhq`E;37<4k0>ZFh8*K0ji#qZZY3&tCIqJhi@+!5!IlqLs@Ld#XrN^?FhO2D$O zJtk2t_CU}O9;WyMAC-uMH$K}$WOMUOH;P;En&v3$_E$`Eq0@~zUfXmxSuG@VL{$t_ ze<2B+*gV8V31nX4Qi$plBJW>r45bOtT9_Nws)Jg-4r*{Y!C{a5{W%^pOBxoxc=+6toeWyteemISJRQBhrt znkl3TbBJ3Zl@Z2{f@BGv3&Zn;xO%0i^O={F>e{t=nx%M*htCD!LsUFf3CSQlbp(bi z`mym6UM9xIq0J&6)!^nH&Yk>7MbrSK;!*e-xIMt5@S8kR3EkLQ7<&o;Ref&@LZEnh zQ#r{}tf62!n*sUbpRsd-MhjcBM-trqpN#VsasleUNUqx9@`6n^hYC5N{EEvRADp)G zE7U^}cq@SF68uApXR4i6QH@^_*fk`-0u3=TRt0!*I-@3=Ex*=fm3wOu@}Kdmz!?;< z*&hEP7y>(A)<4{mXn61uXWw7-pgC}}(7KV%$wg%a1b*{gq8%E`Y9312KPP~Z9Mtx>Dxm21bR zpT?sJgB8dh*AYD8mIU?L9OI6MdT3sFF%uo%@>eiZIQ4L_|EH)&!C?}&Wr*lh&t^wG zjOXB^Q5*FdJ+{F${c6$EXu4Zi5QNZrh-&@Wa)?K2g-$tNbqndLhP5uSfRfn!O zW$HOGss158YzG@pnjf1L;iK>29I7Uac-&N6wdGg%y050Srj5H+&6X8x>tt=gldR)* zmVYH1gcx`RCBUMP&F8Mm3xwzbXG!HRoe12opa(F0gWwr2$k{Mpd7VB66HG69fFAyf zK&H2M0ECA0g)K?@DH_z*7&_kr)Mj&p@5=hm35n;8gf7KTe%Di~ah)?D7 z<&o>UGLoN9s+hzs99AmwFjdtrc&_m?S}Ri_ra}vac@VLAFcpsM4aQU!Y`Q-q2pEOS zsr;Z+b*+`HIAL3Y27g5iQApgRd9al*8@E2O$_Jn-xRt3OkHGUG8U#jG55{R~;lx1U z9CQv`v%@N3+e_3ZQ(-Ih?F?GEiif3uz3RqQBy>P}v*Rk{cQ%*`bP8e`M0o{(vD~Tz z>ws4Tq~V%9UI{(Qu|5GypY9(<2a^hRKJoX5AFAxBR@1#k;3A=<0AepxRNVZ43wLa5 z2CjN4EUv*{G=l&DEL9CMW(Ik;f*T*Q)yREBElqVzJOF_+AQ^46-252uml27EWk3Kt zdSEcbk%Bx!v45;sp7sn_3ypx#6VteSyF0bIpIm=IxJ^ey>$eeSqN29u$C<+FW%x92 z!*ZsrAJZWi7P_=5WG=te#&skLe|CY44D#m3b;6cXG$PlTg$B@rylv@WeyyBW7ax`+ zQ0PKN=M6SL&J$MJ*@&EHZ(`QUa&==ley}5+pxOY&`i(D#v0pCl)C}w?-`9%njQv^Y)bS=HnZdYP=$D{!A6*V;T)zVU||xe&9(SM=rIUf7HAYL?O<)|zzk!AA~oF_$>HK# z4Sa_-LqtJt4P4vuC}5&!$&)b= zRvIRzK5AZe(QFT2noDQ^D@sjRErEc&&U4%-s&TZWdoluSS)t4@CWM)ZhMugjFgUw_ zs9&urG61_%zv}o3o})`0ZzHEOx|+uIF+~* zEq?GIY2i_N9)hCSFE2TppjRjiJ@|~FKQy;(=|OZh3(h?;wuK)}Ojwpo>sAfLqn>KC zxG4f#?KBIETd0Q86qa2TfQ;IAiWa^tGixOFfRP6X3v!(|?s7G`LoOygexXyMZ> zEM+J#-AJo8xx=f3j|eePJ(8QNYHnyn*|D=t2)@R|2#?xY@g8lcZwFy3Nyqaf~e5_S6jkP(ZXj=ScBDZ&9Cr?u1rx zaXl*j8-Yr#Q_mwqb%%kU5An}vRV9Et6oE7ere7Yqsm6&wI-;KkDc z3?B@@t4gVbKv6aA_eP&wqox9I!-u~~j0M)j_|IJj87+Nz0dQ%9j`38*fj20@71)=ZR_u-t$E zWn0vK^#W^Ht5z2BB>b@IN^A;vGv(-76EZt$jzMyrx?dX}{0b2_80rd3{0~=Qg(IhE zk)Xi{qI(8Fl_NpJGYWH?89wxIFwlSM?y4tvt_joULJ_NBt?1?fL^OdSJxcY17c`Mk z&CAM7&w_4*t$Mi<-N3;;97~Z$L&GJs*+B~t&<_JZA*K9H0b0D7p{>}`u&gYIEi^o` z%frge1Y1yKZ|4PNRx{>aR9aOPz>k=6L8!pr|9$l$XSa6Pgf-g=H1F5t(Wx9d1%Kt;j#u za5ZMa*5_~zTTpN~EULlGdt>-Ntgzu6d(ByLgIx}T5Q#>ZA9ianGnj$J<$htW!i#w3 zK>%s~Ji^;dY*XCysRhuQi~guNy30dmzT87oWOqZ>q##&##Gpff3YZ#OSO)3Xa9>|W zD7Lx?`{u)9^KrQbD?E%V9REkW^&jPTO?2)qVtBNbGY`UR5^*D{$>4jqt&v?sR4~bx z66NI>;b`v^Obiiv@`eCaTZ={-cfBQ9o$#|_geG8{leMT-8&&df+s~zyyzKV#>R&M^ z*nxh;D4!6|iDnGT@JVbhmbs;i8ZZ~eU$vQxzkT!@RxrZ9+)|N1{PmGME$l!yiU0%g zZ!=FJ64cS|7l_efn#&vq*p%<(S*EghAUrgmqXHb2-;Cgk(77q;nTc}1R zHUU%<(urn^YSkSfy&za$%iW!z71ywB@v7k(v3d%inc!L-mJimmRi;OwYG6BD2&@-q z#WgOVfL#oo2L=S)hNAzcb0j-`}l469fR9!ZHNE?Td273ab7A zeuWOEWya!Ypb9v_Y#jE8*pU%fR;S?-3RF~?Q&W@80pV(6k=k_mA8yIOrA@<@5K_^B zDTqE=8tOuhDMN+zd|65X=J92)0=c`4U@V1I>JW_-HLxQagnx)|G_Y}Cqya2k_~o7W zZBPk>?Z^ugaxY*vJizK3(Dny@H8U(LrKR?Ry@bs5u(4*mUq3gofb z;4$$2JbUB?^-pj>WRqWmLqd#%^A-t@P4nnBR2#|5SAVsbPPnMoAOL2bF)5I0AtrWU<%Jx(Z%?`*{_q!A37 zHI9226-**$vw(WtqvPKNcnS<18$Ay2Aa&gD%@5%lGTvN+_o7KNfVD@pR08mNDQ2Eg zM5>q9L^VgMA3G?(Ka!xOuZ9$A3o98AJW`xMg>5Ksd`A#n|D{hvL1q@xYIAT_YUPyL zyplvUDJZPkermj&5^&Fzan5V%|r zQAh|i`P&f3b@Pb^dTcfwtDrkXhQ~{*%^2$bufq+z z@8^UbxE-xar|?LH2?}u3{>&f6HH311^-5K)B>i*(2)~^+s=+wlU!^N@Qa1+kIt*>+(I1mg7kb$ zt*I`~6FnFXA$EprHICra?%E(9FEFyo7b0lzkW1J`R_Gd=5df7x5e%+tfEQRGCy_Vf zQkO3L?1+GaND;s>GFr}!T5140oPw(Wqj?2cdM+0k3zP9D0| zKFfdP!`94QzgRyU5Wox$1qS)mVXgGgUhmjDp2Jvw)(wrRbyFFf z9qZAcr)c5t#JPejBsYRl27`ZIgC1Q&%)=`gb)ohjQBF9zp~j>mP4VsW%e~sd=gNld4fU6J4XDtakN<4?s_`B_ zKVs_qCyUg>b|m#zP};y&nT7y{QSo384~8L4%8Ox5&?>7#ii5`n!=@E0Jhr5JG6HN_ zp-gNyIPiZwSz%#tm;q5A?IK=9$pGw5eY%%FwrgsTw?Wl`*F+!AA9x16R4;((i@hw2 z6%1De47{ogF#BaNRjaWk<+-<9#2jZHz8o2e=OcvKStVLCA{i%>W zum9TOdJYS*SuBY3R;wIC?1fnuRxED@{}M9JZ@rmxFh0OK;yjE$05AjrX6%Ywk6|Hg z+Q_vrGt_g8bbr1|JhHn}ubEPklKUkR-`-jGLB(pFRzC?9tGus+GBSqBIkoHOv~H-} zoN0sh&z2j%LOLf*a({7~C#MeHJvMt~(%FFJx1T?couy=>|Nhzs>xf4yLtpM;-cuVw zl93Y;_>3224$TPbv$c(668~qYXm{HJY123{zE7bwT#lILsVVRpnT)u5WL;4O{!av4 z-qx-A+(rDaPHNMxd#+$dWb&4@X_vez!(|?Si*aH@Ga@?+bS^1w6%RZN+b5T0diP#-!D55; zs;salmk6FmC1yp-(%4x!#aq2hoLJZM=9d==S<~VsUfH-PZA7?vRQ~B?(r`|yT=to+O3ZL%v??h zT%l6D)bG=Jc+>o8E69Uo@Xk_QB51?$Vx6 z-%l@g-%ZvvSSA;{W$f-Pj$6jL?q5v_bkn;&Fu~#e@%hRb_MdJ{_<;4Mv;%Z)%;T-< z9R7?emMN57%Pv2Cs${$G($wsT7dI~DdW_cVeb+!rM`5k;eEZP`r4w?y8}@jw;FRk= zenzp*n=?eoPy^c6*uprGr<1OUcA&?2omjB+#m(KSQr&LMojrTeygtMnje}>F6D?Md znBE^N%HaJymJ@xuv`a2_C-?7-wfwkSM%*ReqcYqH8lre1%d8vo^7{EZ0!r+o5~MxF zL)zL+AD-~|khG^(=tIk>mwO)d?`bRXYVO>*OU#xGQOpecRLNl8DcDn~t*3i?Z@9&c zU4g1|dktS7M|kq?D18YgAQHB^uSakNw8B-i;(g)BgekU@Ufo$byPHzZehY~c(|uJu z4#Y^ExoszPr7*tj)$ch^E^knp_-y<6Q4VWjRL5*_l#emhjh{E!|Jvp?ag`;i3%U&! zk*N>|k^P=`I&CmEkdMkDS$$88+K`v0At@F%dME9c2;HyeDEOLC7aYxk)<)gvqBQ>BVd)EfO6S7KG4Z*n^j zv*$*?cZb9Wp4{qn;=!Zzsq5n^FF#v!8APX|Gcw>g)Tncs;HSHi0pYY9Og}j&S zr8L&9e7j2Un-8E-J?jJw(@%4Qt{SMwg#^z0k z|9oes{}|Vu6PG^Sr9Y!X<*{BKru(Dc0YvOv409MaThdl{dE{$qm;gSV2PCQ-#(Hne z4^b05p>sprwY#k12c5$0Ap6=yEB6Z^*Ro}RQL}q3G*IG)uZm&7rROVsSX};4DD5eF zk-mog=!)m;`Fodpekn1JSeh}+=oX>FC?ZMAi8*RxsqA=!aSwyO@9_QEbKQNY?h;+gd z;BmqyQT9jZ;QB$jB8;5!dlSq!V0z1{6rgAu{YZO|yM>p>_W&^JO9>eI@ulS2xJog` zdp&GaA7Judo8zyGpDL+K3V(b-Pv5{`yw75t$q{|Dg3`jCT{~O8>uq0NF zk9~37s=#Grscsr>x_xfTwo8`sh)L@U%$E(*xqRXtstfnMle%iSRrJPg*yp_Z1&2Mo zfUWx9KX!>b)*w0Q&<|;bQ39I-_@xoRqCqBSx?nS1xLe{Dm1?);EBPopKX93i z)jZMD8&Z}mk;f!?cUNG?er^;?tbqeYZQM76KJ&;+yT9YGS~rW8^0`?>u6CEoH!3`l zjdq;2y=Q7%#%7t`X;z}&zp*V&{Gg18`54?s(=(w{KZ<3=_Dw!vla=Q`%2AwG`SfMi z2r1f_tMtX2a-N0m*k{m_0~`afAsEa(m5s4~+&+l}j~jy?UDH;DDWCUjkX4fR>0(nt zuZe?Q=j=$`vAjdy;$7X;`lv;Ie)mkN+Yn0)%OsGbts6J%W4#TMgLk#x7_gW-9T?0! zsNA)?(@Oi`_uCe4zqGIC#kV_-9Z4?9ix?m1cWtxlrTvp9WslplEg>Nx|3(^So>;FM zq1dsPdh+`hH}~BFwjbMdTL?z2Jd7i5Z*)n*Uj22RZp`L!*RK?1beb%(cu(+HK(Acy zGdEAF>dl-v)BIwl(bem6v2xMde9ynU;?hgaO=0r;=ZgYLwq?KaJHP6<^r%Uf7&oE6 z3Q2&)-E$?d2IS(!C5{^1P9|=0Gy7ORdVE$lIoG+97SGx=KPh9g;<$jKZ&%5~le_BZ zY0n=x+I?^MGatiHa{Y~SB+O#Q)xQ%;Tugw%G=E?l!r$?b_SU^=sM9l8JiyaA<|R6 zbKAYFuptL0!av@Znjs72TPnK4|wDc1*jRxph7)`gOW?ss$FYt!yWogSCJ|A9a0COgHU=jb7ibsT0T zb{M(SG)YfWvB-H_zhg^24C`kC%r`+F9v?D})MYs~n0a#2;m;D&+-~T44&Fk&KCbU6 zr+00Oii5?UZKZOM7qnnH4E|Gb5|Kl(S&~%M)t9o)DtuabZpoQ;V`;aPaD9*zHylu| za)#4`{+zyvh@P9Xp)icL&i2{6#V^-Xc6oa|6B|5yO2+~;$N1~e`eCzcx+re@J>|gl z)9mxFjwmXzK4!KI>l%E%u7IXJhL2VP3@Nn{%M6k|aPX0GWsbq+E~nk}JM6wQlu3LT zzEQR#$D{+P9FAEY$DGR3u;=9@MSOK^OdZ^_Zc`-BR7fnxv;ybjf^i^$GoFD;zyb*g z>e=^$cj+WOO(AD1*sHrmvDsfwbAh{)cvhn+dZ%+r(dkW@V-FR!>>7LBO>;4{B^lU-ya*E zUov3WtK0jOW_1~`7+j;GwF^hBu=P4Jv$OqnAFESI^3y-xJ~p>k`N>6t?Q`9A%vMqY z_4ge7{@G$qPc;=97|U12hK23uKQhAd<(>ViJyaZRUrb)9N7ec9^$okCR8M(aeqP$> zsqWx+DD~3{N|sSH&!XO96s4TX(3}1M%wVZxiznm?U5}MotP}#{bYwr71+y?pg#mj% z8RUtrcfq|>0UJDJnCrHUY~}qki{-8T!4s5IxhN&3;xeR1 zP~?kqt47*y8Ap@s(4kT#wtpvJv8fn~oezJs8SJT;%;SP5*F@^o0jBvD}U?WvRN1Mc4>MYKPff(t6Zx3_{E`* z&Y1-V=SBBf6Zml|_^nrd0A+Ku)t_DaFVt`RE!W<6lp^EgVhO(I(gMb`ZQsduknPG3 zT6bdxzj<^%)oGKn@}ev2t@;@`RTiR7qGx{u&2+|NGt@#+L$iy!hUd4K3cRx?%=PQ zRx3!A;PI+D`}vliUYZlc&TZGey~;Id+db@Zdw12yofX96qBagLF3VJ==b3EQE7`_y zirSy2&!QKFV4zL`P`~T$ac%+hd!fX*>lelc4Om%zVP08O<(SuEwh|W&BnB>)mD7Ip z-|jp1?p=Eay(Rjb*?9Pb^t4n`OaS=Vl}|V2=4fb1+k1s|W+hyCoDT95UK->jlVPe)62H?#pwy;P~g zm&}0u&RM2iSBiGHAJ_ZHvOJp(qEfQC*Zv!8kv?k5mDMD$U)xA#L9aJnaTj^dF_79z zXGF4P2Wgv0_9Lq8);UZH22GTg&mGGo%+DPw}}6l50! z)mL&VT~j3J5^E$xI=oK4nr=+V*XTEAwVVj;6o!KR zwR|u0#M!;!Jr96Krb?eN#o%g=(-H`Dna<8EFG^L?Jyc>WA+A07{pQaDWaQ*ZMD@Rs z7N!i@RWL(Azngqg-ocKm2R|1#@Hud3y0``STVlSHcQ0;^+26%x{F{eoou-K=DkprF zr;AYwlVaMQ8ZxJcYRWzT_Z@~xSSR#894bCs?&yuj7uW7@HvovBPrE z@XiSz^Ii70Ep^6p0mG?~hDCH>34&>UyYFe~f@p!Mmn3cHI! z>3C?Nz+Z>j7t=y{W^L|0xw7qemx0N1_Ya8bcW#iYj7CD6bAI|GLb4`Lx|WpL-%M%q znOB8&U>q%w@ruIUVA@W!yUptO1;mA)KfKC48ox<&4e;huUAL_J z?|ZBy`=^Axdg*uV!~gD6Mt`JGx&+=F!*-=9!<#+i;}W`>I=y_nS_ur7|B5Yi^Ik=czO{cn^}ghj zQ`;FuL++ngILTGcPx+jDhFsQ}pcNmo?bn&BM?E5Sk+?fD{Au#Aarw4eBU0^j_sku3 zQ&Hpx79u_9e||ZbQ<0h?k*l^0TsGm9Eag_2-TJfRT;{lH+jO^Cu%;iNP3gFeTgLC} zaCF+H_I7!ZuRVLLzm}W+savF9jC4w3`gpR&at->l4xyw>ow4`V#IeKIxuk1c8i{2d(J#MqtTUVGKYG^r$ipPjY1@>~JM2wcy5GJ#cbiXVU60UD|EzudTcrK*AH65Km@Xe8{W;={(zg-i`wq5OfAewD zs({Vsr}R|L`<^;$@R=Q&Hshp+Ea`bmGwpo87t_k;zt^MWF8p|1BK7-f!h=ACvWTx% zm2%@#HMSEZKlXsJA9M?{PJyvYh}!L5aB}{^gNye)icQ_Pa@IV*OFJpS2h)!49O5$8 zyHaM+@YFeFqGeM94e$8*#~$`6U!uwKDl+K&%`&mO?Zbn{w6-!8-Qk5ci+n}{LDnlt zhrh`icZX*BGR^I|PUzOe4Wyp0vp$6v_)d7(Z?%Q{V1(F~7oSYQzrVhTO7GZ7+lF*sVi`PZB3GUQ&#q-iCF^{Vf#`X- z;?{vO8dy;1O|HBz=0g~ZN)(U$vhPcmPhpXdf4rPDEyDBq>LYLMH&1YmpD?W5&Ht9y z&|mQ1=1-T|yK!c5!VAqY3DY}1o)EIVgP8b+Ct)GmD}h_1Pq{9FIc^V)x%b7mb;9(* zY({6b$x4*cV7Ud&(MnziN;6#CI2drx|@&?LO3*G&4-APbbHNv<@4S`z&8-bG!gvltFpy-sV=@b#oKjflB%ga(B!k9w%$0J_4d+1Gx_|3r}keD+!^E?U#jfpQ*Izqo~@ejEh+8?=5M9%d}s-# zXn(n2Hr;72i}{=!%oG)spS9zWvo)t*nHD*yh5?{wu484F9dZ6RGRlGcKL&w*B&m-UnTx9Ro^3L{K!b^Vat`N zYqdvxjoB6R_E^9XsY?U2OBsDCt#iK=gg<=RMON9}z--IHqi1AtpPgKIAjvKz7wq#> z6=g?We7w+ZQV5_3v2d?pvefCfEC_ z`@|*06*dQBzj+@{ken$=7#jm{`Q#Vk-HKOv;Fd<1pw$5K3wK@?lgrF}m2e}?%l6@*?lz;Qv%En9r_ylXggOj@Xe8Er? zoOH%s#rtvGz`bc97IM~}xnclhGM3GNxpZ}&bDx`akP|tN#Y(k9-$Sglv>dZkV&bXV z_)ATVV0hUQ>EkiYs{{?a8rN$9M@d0%3LxB^thHJ61crJ}*pm}!)ZX}$6)Nzg`_BFx z-&6|&gqcpIy*cv)d^i$1AN6Q&up@Lf8*&>eJ<7IYMhW(*U}H}n(CrVgw75sujjL&s zQD?Vf0qY|3kTiHu77ImaUU{B~0Tg^JO<;DyERKsnXZhmK@KOy-3f^oPU}6n@9Rdvm zi5B^w*%8M~wZcK{4)D_BpaTdgN@F2s#_5g+7C*7rbZQSseQ$keKzd$40^VQ@#+Ekf zs!B%iQoa4O;7Z;i(3hif-^dcH4!tAuz&kFmca)k1d9_u}>)rzJCoo7GU;W z5!C|FJfJV*;TbLxaR&P}1=_x=uA>A4nlKj&f_BE|Pfi4jVnqgGNX5&ZSf^1IT`+TA zItv4&BSw#sVNUFa0PL%xLE?N3U7LV0z(~EVoLnfp{iD1{bjH`=2RTe#?1GEXiBgu^lkS&#(n&y6DC6E*SnA6xIzP< zfm?b@1MwM%V_|YTKl}j})a?XS6%08`3ec3psiPnjbNWQjFkxC^GR3EVcOYKPUmajYlD$`EP-w8d9Sq5(&RgjNm z136U92Y@?m6t@|Jp2^Ts49ki(7@rM|IWTi}3ME}&GR|YOy?osRLp0)Gn5mhA8Pj2Q ze1K6H?|VG>h%jpkyy(`66g9xQWmp!#tk;4goM0KeSPFYl8}A5k$|tD9J07|XUjW8A zXCAf9pfgK$l=OlIm7&wiJ(p8W@y?j`fSa?Yty%^2k0b&KS?&WC%*P+JPX%s#TIu@= zdgG3bsAI1g27_>v94-&?r`Yij9Y!39`9iN;J1#&7QZOI-xBEYd69%BYpF+Rwr$|!U z;B8j8L7VRTNLYo1?nDqzbG6Fnsd$?OScv~aZdVgbH__on+4kpNqS-NoH(Bm*>8}DK z@x_E}CUG+GB8SU}7>eJ|2{{9u4Z^%fjv2vy($We#f2HWeAxuq7hXzJ&+{VeKV9d!= zrw|tp4RlI4wH`Hi0vh>reHwcNymSjNAbXWO=DTq|9#|urREwej$*5QZ%rI_l8txrs z1K%H|F6#!!#h0;x7k`}v9r&@C>J7%=(*eY`7iVgWEy_jYm; zOlA_1)8lURiZ%_p+j;vQN(q6nsDshzQ5Dev0h?nWCPfG?)!qG{zoq3o2-X>kwrzX+ zns`6NDlDwu|EV4cfqrN8jla^0R14iG?LC2V3QVal&EyWvgLmPg0Y*!31fn z(4E+Eqm}sXqKDH*hHY^AadE|^(u;06{R{M^i4~)6@0(;J8~s?m&&Y_)GTO%1vmG{M zfziwUR9T+z5d8TeH1-3H%&1+Q@xqL*2Xp z)yD3!e0p`$BbmKeblrLLGF`(%+q_Q=+ReJ_Jo?e;r4~MCmdOX++!dZJ<%}kR->}G# z+<5bY;excTB_VGQZt+i%Q66ur;BP!Pa^r|ImE>C$A+-DP+hoNIuRtYL`ODk;>~{x^ zhv4sA3$L%A9;L6dNgpM7X9JYSDBFyIV7l1kEyfFF)uxL}Mby}0UZk7Wa_#bwuH8KM z6`k0WSnzt{7-sS0l(PG5CtbQO|8|lcP^h{ zo<3^7Wg7Kjo~>SL`_r6U9%f63U?LNDMjhDZj2>3JacPZ#xa+~#DTm`AMLP_f2WJn{ z)p5hsk~`1nOP;6z7ahyk#>=~kn)fS$IQ=@Kxjos>uWN3YwcGFFAjOkA@^qff?9g-A zP;n_)NBR_0NyuAV6sImCu@+`THC_zN`ekpQ6%Ti%i>)EYe*KuxDNK3{#PuBq6-f~7 zb?Mq$5{D+dlJ9E7z9%_bWr3_t&U?L!NZ33uVJk8ZT@B=>Na{+HgKoLZRss_~WUEQ1 zH8Ee_zgTM0p}hFNvA&5)r~O3AHz%5_tcFrKRXUw6)4%_)ofEIxf60InxFf2dgt#UX zhP57QT1pn82MkKxz4OQIjWVjx(I)5*${eJ4r}BKCXO~})OTLUrl(CUET99yGFGf#W z&(JB`&S}V*ucw#GE98Vet#qYLDNp-6KCV3dzbMxYn!7hRm}uLX9dpH?Ppju+HxY>qoxkipy=c2qr}X!4yLKyoaQDR3)f>u=%o=%n)YE%%H@96?Jvq1S zd|JUZD50=1{8m{syb}IXdLFe6a<-yM1cE&j?5u;T#@RV~`SP#DoOhK6pDW)XOP> z^a{26-qf&R66FT=?XF+w(hzaK0 zgmOvK86A2gZy0Vcv(rSWt5D1{J~LU#MivX5Pi%XyQy=jY3kE#1U+Kok@?~AfhCq z_}CIiiRG85mFa`uLCv7L-oPtdE79jMIACHTsJ+ z?@!hL0Qi)%(DT@w>{pPYkJN9XV!97-LQn0x2V(5>PxnooIg$MXg zbyfC%lUNVR4sKan2sakO^qinwXIHxEqN*I{(RhG&qNvg;!wKT{*c zb-Rz*iyJ!`Fh7+5f}3m)LapuTfm6S(IV{oj@a+8_kpCUy+TH`I2m4TiGjzKT-mmi= z5eiHOaEZaPFl!?@$mjd0Vm12YYu&DiC2#tYzJs!Jimg>TQ{E~)uq1#Pa+}=u?=5(H$zog78=oH6W{oPH za=1VJ)xjSJDqZyau4X%!%xJUTY9)o<3FVyGJ~cHoECfwcY6 zqdOI?diY9qXrhut-xo>aB$6OYv&pf1%ZecxrF4CE=fuI0A#%oxI}BBrnjiCtovHp` zyY4bZg&lYLobBt>-s|Wqxj7`MD}d5h+&6-X+!U;nsBos^pbjZSUu&%T(K3D1cu+ta z3M{3^E*WBFzh;K?#DhP+PtMe|)!h@_BX1O@i9Zg3xxG7X%(*#YmKLRzXE#;8c$%(F zhz*qLUeHI$W_8#en@P{(&6&1aA~P1xQIc4&#$eXz9n8zg8%{WF?`eB2P*Ex4>9zk< z2Tpmr2tq)nhmRatnkp$dWP&v1`aq`xF;OKu{Zk*DTx8%L^V&%9sNCDlA9N2bNwgi6 z@EywrMrBnHJDKTASBkTGC3Js38Ujr=P;s#Ch90})jvtULQ)Ttnmy%b%xhrs!zBUtQ z{|t_KgH5KB*tM6phmBm25b16CeZ`F7wAWb!7SGk#1%(Ou!!rAKwXG17pC?@wZu&GY z@FwNc=V!Uy#Ja~W8#!sqyuM`f*@F^pmc1)1EHo5V88*yPQZ3=R{1dT+o-z!%VN#xw zU)Bz>TqXzEh%vItZ`&PmS#R0r(~}M9$&dBfb3h+%{4w}}6n$Fm!S8!J$!cF8T&ysX zYhYHyf_UCJXxv{Oi+)F@m(JLBV%p{QW`|~}x!xprFSdL_9OCV>eZv;#hxfclj>g&4c}Lp7_&ZkEu?c?L6D2+*+*^ znNb=ZFOiA+#cd=NI6nC_aD)6M5=@Lvel)uLqOZ*!{dEsc<#bjw2#(%&`ld>P&WMxx z;s((#PA^p*d24leKSoUK@U4fSc_Ew&N@gmJ`M&JoH{Q2+;XvkLA1#ob;a}y>BGAiUIE;EQE8|*Oum1YdstS;T$?uNK*0R- zLq%p7_;<~jCDtLgpwfH}ee$m3o-bzWPivRXE>Vt8ng%89c z2HaHIWI%a*F4O#V<)wlf-||MT4LHAAXWEXd(smkYPZF+_oRSe*-MUN+ zAM$nhp4aL&_p7C zQ2%*`akko7d)*5nx1RM`w606`*6~)o7mX9&5*&Boa)F=O&7g~!Gd8;PnVFeaVuzm%c6Kl{OfU%Cs)1YVA4P#q9xg`b;>BVT;#3bG8XzIpo0Quv)|i_9 za6Z8^cgcU6Ba1}b*R=24_avbCRTnQ$&ns*4R$SdQ#%9%`2OIUYr@Sxmt-LzME_^yMNgD0kGOWhDu$F+N9vzTADydhylO=F zVn947b=T;xFmlS?Da?CIyyjduzF@$}(@+y5;|}h?u7KNnuNJ4oy+Q8YO7PSov5Jex z&LX}urWaclX2jvi5YAZ^C$}1KQa`S~}y#e^9>D_3Snjhtv33=K38%OHU7m&rIAK{$bbI8Rt`! zH(9+-Fo%+d9Jy6d-#))v2Q?IxCkHR=9otoYw&CZuPdc+wJEJI!78dGxc(v@hY2kSZ zs%l2?9XnIg#LRZNrUzm^b$Y!^PLXOT(Mfg{SQE;u1wJPuQb$gDX?Sp7^t+t2QB&M< zod3%?M?ah|JNdaLS^aFqXHX@Wg*9zdo{8jWe^{46Iki+ZRX((0yG7zTX~1c*4Tb-}B9{#69S=qRet5)E7w@N*|xGITEabt)PL2uIc1HI=D0r zE1H-#s>{Gpn`39r@26!wC9q_h5%?L8-(JTuUN}(c0F`eYS^6K$3stYMhe$j5(D%<> zuc{I*>7nURdJ(svKP~f|`r*|B6kINJ%095uP$JUmO;5!hTGzigtkG{5 zt}bIg?WSsm+`aD~?@MJn|F`1lrIfhOfcsgjy9XwE*ujdGG`W*1xGI9+B4BTM(BS~c z6c+}G8cPlhlPC-O_Bk9=HcRsk%Wso5mpC?e!Ge^#v0tpPYC5b03W`n4;zipzUTPiHltU@@>?fKwW!kME|5-D4duS_W^_01LU$q z(?XVH{-6x$Egyh_n<_>sL_fZ?Zf?Q5Yh&jqYL5WhJ=r&Xiu1RrU>9PQ!iqcOGNi9* zMOAvg>-%2g9h9NHm^3Z!Gd!yZ1$UBW3U?;I(!mzUsLXw3qw_9V#(#xg@)W2^oG8`p zy)JXj3~EfzviQsEtsFd0MN3_|u*SgYl7vE*?tV~8UtW+4XLBVvDGuY_%J)6)eZ=an zpo3tOC>qTi7AZ{|UQjVDnt5umqRLU$kMG~4B!)&#m-Dkuqb>@NNDseh+M~n$2`t%el0yqX(%6c(N_&8j?xrsz8M!a%VlJ40DudG_q_+%I)7ijT(E%*; zGc-~7K{DS$cRpAl&YNB2rA-bUIy5o2>wCid0VACyp;ZWQtTXIv*pu+cDvfG4M)%Oz zE4u>6etNm@mE$xC3Bq&63Ib!!=lwgy&By8_NDZ1hP}|#Vd&P_yGqMjXvI+*DWbqKI zzIQ#s&T%TLK}amIh^%+V<7JEEhIb22jhi+`yevi{*I>mPsKqXV^)reN>n+W7pO-PG>0*n?9-L42*eQO#`!JeF4_VnQ_L~fjDyAOq z|79JR&+9I&^OEX?{65n{tTZpGI69c5P+AP7GqziFE%u6^Ub@Q#h44&a8o<8z-Zyln z!i2tHHm!G6>2k$FGGJt+T~LD2_5R}*Y<{xNd5^Ya?yQ7_jp>^lv+wQdBR^g7uw9?U zPbo(Opysyl2uQwmr?lIk;;#=$>MJ_unV#qY{WkZQ=bj$fd-+MFcc1KUezKS$O7g#E zDcLV=wDbPx@Y{PAbT5p*pcDNz>B0ruQ)#sAH~J-d59c`MJQ;*hSW<0P&m|78sqsC= z8ZgRt+1`#FdqsWFyVE)f14e`;ffy!=aSy`Ov#}7->BIR797tFrq?aQnhbc@Acv+k$R6_z46?;eWL%#5H&jFDK+KV?ML3vEa5*B`v2YE;8m^x z*sC9bKp@MK&99I7r{IX``G)ozS;r*n0V1oUd%KThp@DjOI+z@;zD^;qV@63vS6zTBjCZfZS7g8b@RA0=xi|iP3*2Vz%pDzGGeXtVuL zn4eMq0Cl$6e9-2%{wo8`Px=2fWOm?9+5VdOIj=Z`1Z|hizf3ImGVs{s9`02)XsLeF zZLBs*&eZ6m~2=pddS~W>uA(IoANk9vhP+Pfy5lXFKYX(NJ<>-->TIAdY@PI8~ z_tffn?ue8nb~w!_b#Y-K>xiTilFh=|+s_1GvLR%Ju9PNkjd+NP*f_?cVIcSh6Po>1 z4%L=VbMZh{?%m{!F}~M<-kU9U=@EKst^J@WfuCFN$j1GWzN9mi_MGpto0_SGtfuwz zr4)D5zWcoLjWfl&w1e+gK=c8|wzsTk*+1aQL<)$Lz-fnsAra}GxICU%%0Z7c^SLpGkw(}iEJj8=AmV4iSIQPW>&M#Ed=r0WPVp>ERhc^@$ zy`-F>hZ5X3Y9^hq8jo($^fn76;f%+FpTM7BdxeRj&anC)uNtOZWU40aefL^`|z&<<2NZxwy>j&BR<^%RBTs#Id}PQk^ii{Q5M;E)U+e((T?q~|rCSDj zxbS0<7g=D`X6`cq!u9_Y;h=u=_#T@0gb-ts-D487F;lv(qoz1O z$-moy!QutP)6>%nD9QbI%3`z7Nn$=%0XR_sSamDGoRk6Zkc%4tQ@sZ;*Ysyt5vN(7 zU}mqNCn_N2P?L2*ZCsKW!bjD;O5Y%G$9Nx~nv^E^A$M#{Sge*g;37x0*BAWvG&dCl zsoZ91ijOU$J!s-s^phczx)E2uOZvU!t2tM=1xDzS9lv}_kicS*LPU1VUE!Jy7)OvNXOJC}Mu+_G^Hp^kZKD zr3l9LBUCZe3y0^WMDZAD5Kl+O#;y35_^;m+oPPiNom)p+@Z&VqXYFqxt@fBy9PGFGzQs~Pajd}sLk_;Yxhs^8hb+3D}U)+D9}(tTz;pKf}? z{$Bq~oe3#7?mA>{@Ob`NTdPvAYBT9KtDrPig6^3u3U35^)e`o76o+YS6&x8*)>))6;_l`u7ih{S!Pa&B zT*Ce#EleLCc04R2A}TsG3NiVx5Hbj?o9x^~_ikpv7rGY~c%c;6y~Fmzm-}Ew)?{o6 zUw4P%u^o3wwIkWZyH^Il>^AfX>Vi>wO0urQNvFZ1Y=vq$=3x~4x7$VBYkI2^#pqe* zY|fe6ufF*oab>$emUKZ=zLO`I-@li(FZ4n-w}cn%)=&}{%fRe{0e_SsE;%%xSue^s z>laWgb5g+L1SBk*e3H@kG?`z64H8%|kb zDJB4c#o|ufrzCNE9Ql__AWK2`WWl+)E-4g{R=WB3X=xis!+(KZEfUAQe#mYr1t~u;46h8?Q;f^o?Xl-0JOB!CDkjzHJ|e-F85L>wrtNMisK6Ac;}?K%~;v&xnbxefDb3#=Um`(=8lsFCG7D^f~{CBNTEO;uV|qaV7Hk0lqqsW2Ng5Kt(XLF_#{T*i??;qs$99WW9AQV)VMpI{4*9k{-5Tn!LxQ zyHNIj5*->L2wa2o@|N)vz1BcHe*%@5hFrSJ@{621f31+k;*nS%p3%g2jy*ZY8v9)# zue?J1UwbO{GL>OaRmgVXA2iOBv$1zY?@f(Tp>D-&mu|w12Jl}irK_e^!dcRKzJ~b(Q7ez zt!(4iGg=CRDy8h{A5O@b+@FuWM+$NZ#2m0)NI zo|9wk-EMhjH^mV|C3EmP#jfev_yu*hi|@^`sNtb-MjPj8#5W+VHh6Zqz#rcg#ZYDQlCvx@ zRpRUwTLkm=F^8a&;Br5Z+SV`pR|P}eTMitVU03?>h!TYZ5ap@NxhW!^O1`1$bSH0@ zmc33#A_eXPmZ&)kw9Q=2rJD0T%>lFVcN_!VC_m{pl(_LeGnB*{g6J6AY_u04UNL7G z9Cl~!*n1>_>ppT3en}KkuthF!~`jyC`myk*QX5s&&RL&SC zHPDm2Pc<0>U$pA@3vO(~Ubr}SQ5 zMVkq(Wjj!!9$#-4sF?Pd=W2d~{NSz0q+NDW>YjU|(plT!lV3EGcgj}~UxtWlU@IMjup6@0YyMB|mtX!RYy_B!gijO_0 zw|_02f>{fQNkz~Wa9{{-Y=UkVX^IqjzLQq)>{2H9{HuHf1L@Sl=KD-t2aZ0?0v{lz zKA^7v3qnv{@=Tm8yY%fua9$ohT9)_kNh9REy+#y%9S6gB$!k2=WE&Yc>n3*bLC647 z42^mML?kUWbpRrKj|m6^8@2CKDI@4);?w*1rkB`>5H?$-(8EC*iAEtSyh(&DPo*-u zwOJ!GDr&q8?PFKiP4rr)KNu+c_l_0i(TJu(kn#<%n;6-?K4@BPX}O{8K*J)@$!AU1 z7$H_>D~InTR1zb%{Wi?cn@IzcWqURYsV+n6={$1?zaDX1%XWoGUdA$A7xH=Jp@JBvre^w%^d(Wzh2FLw)(O4o=5BEt0=}Y zm8hQvV*m`e=s z%oU71@f)QEeny3@F*y&sOIpu{mNpm6TdkrOT#}EZq^|G+$1PhGps0}sGSV?B`8=#g zS$6a2QE;Js{fR;$?~$XUrS6I-%(lN(oeXkMM`+4Dp5<_UNaqixJBjjwt04(^AgrJK znJU2rj|Wz0Ic(-K9EMjOP%{teZHfo;sm)HadNkB z%;fGmhPOVp8Zj4y{VnyTC$szGHM5b@lmg%A{L|ZHztM*g>24@27i;+=O^Z0lpA& z@vBjc?d4UGZ-66(&Ls0^e%7t$ce!Q(ipOZ&pWac2a;yel6)j*zi)q4skD8T(d5tYq z4e)$f0`;J`-D9c=xY#=dSwoW`?nuWg1~ikH8G@1q$y8!S(VzP7E*I5OlZinC)mIoM zEBKa!isOWKu!j$9kKAK2J)vaAO-YHZY|G_VTCj5n6pAJHqtCp2-}0D>hUa)^@^rdX z7^^j-^g-ttQjgj9?Vk$UC^rAgGOGy-NWyC&UDW;D) zMS9mkC57vAQOtrf@K_6S5=f%-F^P-r>!mHOsv|j|kXt8^Ke&`=7}sLqd1u?3E0bG2 zl#sKnToq2E-JEt^nvskHyt=HD^FaV1A|t6f?YBU(QVwXKg_hlvh~-l#SphAkS7v@` zaNbH!FmN)6AbGuK|S!8~h&ihwsK9Z**b4M$xH%)fwT`Sp>}<5KFAx zR`!w5rosmRX2jka1sCEOE3X$zr}$#juo`jJ!ii(d6*sgaYLb z?_O(lb$6g)DvjH5KL7B8E$p}scvKs>_pKuc@4u@{=sHe-8Myh|v!jJhSn7~8Bn5?=7%hdOS|qG{cMEWnT7YJhw&-IH zb}^O$oV*6NHQFO$tpKF9EkgU)&umRsnUBHAz;?vE((aMx)nw8=O(kX=6Gd+x^2fIP zN^G@ea|!$B6<1Ck>-gc1sxU7%&ZH(;B$D;9fR-T z(q5QGlohQaO7X%*%p=xMCkvpB+X9(oJQhPJj8im2Yn}72U0;5;zM4z0-o9nD=Uybq zy!5IwuwHW?O5?)AoGh`zfS9|YpI7M&q~R$?FabeQmNE~3o4|vL>HS5 zvzVn2C+67h5fwSDtHcWX$$jjVP+u+`l$=V(uUUiIO#jo`^1D$6oBU~&K=Y_=y|Xvv zuKT2j;5dO#Y5fH61L3R|kFAnle8o-4JGM)d2-UC}pMJvN@;RUwx#0KyzpMUBrkDG{ zQ4i%eBXK_7XTe%sCSGXA>#1VQ$1LkIKNdjbo{TJ*R|?C+q(@LX9^fH>!F}cvbkEWp zdm-i>hQMP$kmrIxh*~aTpS{Y3sJwC^IhxIR;{4YAQtND&tR}-teE-%zcNEg(7?mWb zes4}t(1Gu8GJCIoSYZ*cxB+<;(VMDB=nkFH9%i}%Kz@}1KGv@vCfMSy6kzP4Cejpe z4Zl>GTDFV|nq3SY`PYu$GS@ic`mEN$^Zl2~I^;$-k53Ac6v~ zWMXmK+?IvW9#E0e%My*jp%^QR77AXFW+AFe27B1sNO|V&=ksv(+AA=`H$ONnFLuKxsL`z7 z$hP$sFtb#!;AooB@EJhGWCMkhakLRV#jp8R>wcbd{`-Bi_&+Ue#$0~D*kIxFMiY}I zHoX;C5yC`Zh!R1j_1{OUI+Zq3ZdOD_nyv@NgpO;V7ZzEv~u=ZHh-`7zHj6s86rqoKG| zOH!!IgyOO=^Uv=TA$_5IBmPMPs1DY)mRA>WXfH_)oKCX$n8`-bTCP$QPkmRA zq(`BR;3$0pBd8pU6<8z|d9j}-3ML#jO5B*=M1zsrwMgYlhb|z~HA%titd&Z*s|2J@ zanPwEGL12N+u>m02Jg))v}uq$(s1cV$BK!47V(Y+Xq@Dd9+i#JPDw}FW0rl&Q4bO6=&tK_FA3#6%TD1Kh!bekO zGyn@zwgAj6;)OK;&nuqVP}~OO?93+&C~5KJ#(T~Tw6%v1eq6x+#(EI=wRn67l#;d2 z&u|K?4)iiKRJ|4pshW3(>cpM~hhKI>C;Q_biPRLCY@CkiHW)*7^qJ9!-1Y{pH?Xq% z@zn=UHYwG&@QtEg3P<^D#1IO+Fp3e6(x6po04x^<1hY)@S_~?3`PMOpLV4-Akwj6K z_XdZZ7G#!gGEPPl|NdcWsvXRGbd+*`eT=FcgOI5fyK*+h88_ZzO(^!suul?PhVvDC z-G81L4K?aY@KMuOqzLbdPgG%K^?}wkeXAN6NOj}*q9{iGdYD$2#tHhRrj4QF5v{#{ zd&4|Ne4mnbNKM58qq|LBmm(H5uZ@h=6#R1gyLnJ=J#=0o21t)lB1z|8rj&k}9}jY{ z^nZ~WtZ&!Dkx{*Aoxx-*p8hHaI}SCRb(Z4fX8_s7_lLmWdiVcpF=gH+|w`hqHo`Qs|F!KOhc<3+4a!%{hjMXl&z+tJ zwtnFFO7=m1XDSw=M2i0(%@!seUMmx-JXZ432kT)jC-d&_6+F_XlWF6W)L)LD3@sv* zO(9)}h`N6Dr>&kZa4O|~h_;=5kzH=n<%HDtWpW zdNsnEdM97Z_xt=`@%~)gz*=xT*_^>(-I@vk-({b|-&cQa8TX?F341wz? z#0L0uA3TkKQ@JQM;=)sB16n zFsdP^4OXi+6<7*CjIY@$SdW+JQ-p<&s3yNSOMIg4QvS|N#2FRMooOVcG^t(9f^i#Z%EF(`#GMf zx^`J-W$^N)zwzr?Sh0lP>G42C)o7Q!m}_=;EZ)s8kC?}zn@t7gShYYJgEesb37+U2 z{4jj*Q^^@4;JV>H79|LmISG2!H$iBM<_&62xY@zI=|j=ZzFR+kh@goS>_6SoY@_t% z_5QQk^R^)?z#yO_KnK~4ymnizzjU_XCejl48S@};BaP=TIMd>xL}Nsk3YO}aZVs(a zsdf+2XRvSh;eXL(1ev9;Kw&i^D0#J+g!CnJ^vln?&$U>NMvWE)6XcPqvANyZ#23xU z``V!5n57R*X+QtK$QdSIqYnKiv)8WuK*tD1zwQhB4Uu#VyphHzh5s)Hf3Ff!OJ9 z6HGCy)`RHM_G@3u!`hs?@2b=2>DcQ0Mw0n1O*YotjrQPi#^PF5xj5esL`%2O3FhTO zn>D;}9L-pk)%ICXPWXV6e0R3-<>RW0L=8#9XSM6^(b}r6dxjSu3h-mICY^)^lm`XR z2uZlp!@KRf5Ql%2eGPWm^wXcuaMbjE4@wzoZ5OcJ;C^A()b{bdQBAkXVdYiZ?+M%d z&JkR^tVH0d3iFmWQy!wimp$1m`moK4imFc3CwJ{NyCygkJb+)!?227nFxc+>tlq56 z@niBzsok{2@=VGz^C&HvN{Hv-{F*ZU7T1PLE|iJ^-W`~Wh}r)po7v>Vuc%*DfOM5p zho^Fs1;Wzxy!wrpJi@yJI*txmBPeud{{H@2Gx#)$P<*mpNl=#fvngsGbp2atg zxsoDQ{=x3sbgm^b8S>|krX1RyQ9epS{{{=ZKtCfzl|qe1Ibg^sM4!whRCb^Z67)qu zA6U0Q_?4&8?tYf|r(^H^1<8#n(^zf4h+n8Fu_FcFA1~pT2y|m(V=L0V3-epY}2+=JgdIwZ!|1r@Z9c$X(cCOi+ zw9t3Ec74eB(Nr*avctsoBBzca;WcI{9qAn$ z#d)THmw%juL96O55gvD7%(yocvYe6@j+cM!Fd$>`$=u%Nq?+)BDvO|NSdw3Nip>-} zOM{B_bNXMC&rUfX=((SE{ePAk zGmwO>lZo)C(V_Kp+GE6Psg;N~whcR?{{Uo3Eaa!R_L*Rr_=sib8xZTUuX z^n%GDpHO0+k*K8=E*49u3I{$V<*R4Anh3J1S@d*dTg<9i9wmI&2WZdJ5iIW7$R$et z!n%kvw&QBoDBVUQUEIgQl|0r(AOR0h@yy8X8DF(MnFxk(S*Ftr{Vj2!VqQ}W&Gi<~ zBqts2g&`1cCqes!Ik0H`$(b}q&nJP2&8a!89*CrNPZ`j%#0m|^dtpJhC-3&wY~v{M zGy{u0G`O_Re!|UdauF=l*ySutOmapY4zzx0LF*%E*-r`TZs(dz8+epAJj_I|_g4o8 z&-Z_5)095MrDSnk=+Mk(TKAW>>jH#y-RXT=7WO`HD*twWXIkNDAW7z<)!NHsqtmNM zsQiV{mgvXLfdd4St_q;*lbsU&r<+k~$K%NY7Jd2p_0J_tuq@F)FrLQKg^8Tn{LZkE zdw#!^z3J{yHF)v#@Jnc!-GQR?P3XV7F2L@|_X8)=VbmpjBwipW37?wi`ibN!K)D1+ zAjZlF8mkXX@xrf4U^KzYe+f)-ZVZC9P7Pbu-@IB3o^9G4`VBoI?9-TieKCag=j4jM z$90;uL|<6q#C)g34GnNU+%zaBQRw~)RP;-!`z!s6X*ydVpH1UtKt7eKQT*usYVcp3 z_otT5bufYhe<4A>kau-dY;3~;_AwJ}wgh%r09;S-3$Z2xFv<9j9DS8*xBH~SLW8Dr z3V`tjOM>};>j&Hsk5v{E0ocjk9eCB0(D4=9c5Qt$7=!A1#KS!*vcIvykB*LRx}dCy zgX^ssxAKv@GJ6qqk&fET)UO+TK=ZX;31Yo8)) zLGUd2@AHdI%<_k{&y{nzvVjP6(%8M9@z2?|zWXKI1m-wI5yn&srwXgFfJ?${fj&WV zyThTzh29#!IA;B0P7bGWeaJ~nq(BL9W-5V&K>dhi%mBa^I`aq+EtmXYD2Iv$KHwOM zpq?4Tn9}(76oXBbJA&Ipdd^*bA})WCQ+?n-VW&NXJy zLEm;g4;?R}o!^=NO4AV?K8rfvV%K|>2D-fNBunI+es%%qTxEv_q#nO}wqCcvX6H*; zL@l;o<*|QUI}dG-#ex}gklxRACUN=B)t|JC;(?wO>ZsYbJdDH|uw0NS*0>N$0tAo9 zSddpWD*1FsB|Jfc-wt-9e6#=AFC-i92)tv$FJO34_GX_NkK_pfyiN+IVklod)Evtx8TZo;uTJ3sAl zZ9?Z%BI(_JJE(?@bF$aYXu(lUu7}7085pLpa3Ck^^IA1>f@OM#!>5JhFS*q2p_5ZQj{sT z5i+&QM?8l(2YMe)zibYF0|0EUzw<@&<2dfPC*&A;;gbI^Z@r&Sf624XONjRr=w%w~ zhrFTbF%&LI>s_lM+GaRGAsfT=(ig|wcVKX5mx_ksAL|)1r-Ptw2^xXAu#u3!CGZJ#8WxER)W@)qS7<~b<9CAl$iWz(xbJ~9gB1uTH4gY)7j&YnM-pMp#C;KQ z8{x2b{a=vNAb;2m&r+sT?fQV?na$hr@k!t{oe9nv+_!Guc7Ml+y44)dxCVA*N)$TZ zHq4>PdY}6xpJT@P;j&0s*(P*)ES_-YCI9{P2~(KhF-RgK77mkec++)KU1s}dI2Nn8 zd>xM}hFJvZ_btbA*8+L}?xIM>vn*uxHSkvX1EKm1dQ`ikX;H<`lFpI2f3#njQ+}(7 z%TzEC==4Q_!hh@rvFBb*Z}mRMQ=7H>+TiM6XCSKImf5B+$zPkLUhFK5y0xtTP#zru ze3;2;nsk_W0opKTdX&c`QuygY5f(PGu(nbK)I)4W<2hjqz+e}2`kwg|jEnVRTLz0jtgF-6 zK-OpAfCs_COQFp}&ezg{p6*fT-;Zt9sbE+76B>;N7$pMk#UNm`Ftm~c5(7h4oA3s7 zQYzUj?@)Qa%yDYrG5E8(kDFr|f$yhLq-HX3)S_3!?S zC!4wu4mcmK`Q~BD4EE8%DA}U|SCG^%{9?a${4M_rx8%o-K1Zqi1m36A%T8nkqFhq4 ziMu)0s7Fl4|HFy8_C7S^bJz24(AIrvE3hxbd^Y#^yX)`P$v;~^#u@($)2n1c4IIn- zN;jGQXF*B-P=R8=WY!QBjS0L=Z3yNS4P4wWt^tFEp?%~IM)fy^oS%(0>L^>^OmhSM zN)Cyzl!d|v09n0<$`9p2eEv_cZd4Ek-uVdo$uU_>CCwWN=_tJMFfnbv$H8=7=<5;3 za>LrPyb6+koUJfSB#AVq?#K-7*CTHpQRtskYcv{l+Mp4QTp|6Bmx9BK&{TNwtF^j> z*XT~bXgb-CgOB!97r+!356_ppKi7R31h1=JB{Tp6qB;k(eh&Y56vr7uiyG~dSLWSr z*O&!cGHy4rQj6Y`o}52dRCLDpT$!IcG@LnSGf8)sb+LfGdZ1;Bx%s*XLZ--)mL&a| zT$hzbU7-9*E0xe(T@3P{s|pjJ9H!|zH!1n~um>A%c+> zUMF)RM8<1$%qvw`_vmYws7=h9Z(1KJ#BB6INx$^8c7MZ(8jZdP*HlnJeR_$s45ZP@ zO!vd3A^3;!t^|Gxsuf{gJLC#ntx)ozWQjfxz8ee!ESBlNYms<56DY+|Yw?QT#5DZA z4-pc(D7+;k5+#<>quW<(pBjlJs)sAUHDPne)O@A(1Ysb$*6>_r1>!-@xX#LlCsl7_ z7bTQ)o<|M!&onUn`sldKMpgnY;cINe-_^4~ZKex*j{ko=ud|7+k~oK$QBC3Nkj5>_ zI@uS6^fMhsQjY%~jjHcMK1);4hP^Z{9F7u`jhlQV5fUW*e|m{^xkv9IEmyg}{2^1A zEWnp{Aq9I}?tL8g=qCF(&5%$d59crx{y37`FB&)k3_p{U23#GiKgzOIw zOXju}i6|3#WZW4TneQ zqF)3TQuR5ia{tA+`#e`HcbZ{!i%BAP-m)5q86siayvD|%efQN3nzVB!g`reD=7|}e zE%N#Mh^p7s=XzOQ%_5%t`(0;mI*r9h7DvI@2wC2x#+EO-GT9$hiv03J6#P>Og)*+g%QdjCSFlXVc$c_pChsSEQsYDG$}7c>DpB68cm+LLPU~7 zVhdFr6mtxQ%Ct6#ej4MwX0RC1w|?L9+>(k`0rspD-_;Yq_D|DM9$O|NdY<1D-q9&k z8q>Ylw02L*ZMd4=!qzG6J`R^$k4!7hVZh{;iTZ4Zfen=`E_N~CK5!ZF+9D=++ihPSN!tw|jI7c$g0h5Degh6=Qi=uZ&B9!2mNB6H_4t}IVDAWZ}; zpI?7_(;F@-tKvxCFXizeo=B&yv!k$?!~(K4Hep~?Dp^XSQeKGi<#U4gV*AVDgJsSG zU4`KhBCXn%4wF(+zYZH(S(cC$@SN?>&}tOf*~eFJXdoZDSP}LOg*&CYAU#cc^)(}_ zX2hFs#T?54IXSK(?40@)HYD|Cbcr35q!!R51d5wxoJH(&iq}wl`*pX&B${;*lS}wt za(i7JDPg?du=dw`vK}Y~bZ{R|VA#QRZ)C)0r(2~G3(zUl9~YuBQ0jnEYslEq6OY;YtXhrZwqkKi&Vqn_fs|9|HoP>+%t_!W0o`Yaf&MsA6inE)l9@POc|y9H~#1 zJY;2PnEhDSDb@Y_-gGEC2ta6S{mH9HuPwqQY2lZpcv<(n;a(f-?ee8A>9t4(AaC0+ zNgkz}FpXMDl_6GS11(9W!z22MQl8{);(i6%8Oz2m9LK8o-+ju%2@(Q{-SE85UxKbn z+`eBC%_A~D6Wi52BP02IyoSfWcR1yWqF_u(tWy$U(XXjO;S;q0K;G~We`Xi;;$v94rzLt z*zl-COF@qwABJVbVccZ3JLywQ2OmOnruL199+Hlyak#F_^{f9E&^b2~kS8e=UPup# zc6!CBiCl|w(d9HM^m0k1ps`=5Ev1= z9tRT)RoZowJo=LoB3@>3Ca$~o2uSNmzk1{E;#6WHCiVQ!yMQV!BCJd;O?VB}rF0aT z65E81ir_OzQ$~!xy-S4JE_AWBH0eXFDI@dHb!_iS5b60 zI}Dw~fZ9SabqzCEqUKk?H8pPv+=8#WB91& z&+z(PCyiVVS5TZ}EFUA#1rrGgpyQAuXna7x3Bkphh&fRO29!g+z7x-xI!#4??%8}; z3^CRmeL@9Q4od}S;;pBklkw?aq)T5Z5`A3odk-oL!~tDaDyijndBLmo>ag6V#$5C` z5-+E75dAz2vwZ?AFDKe~VAY@}{F^!PnBqw2^VK(U+_}MFosgVUaeVM5OdD3%?Nt&p zwN&w^Am!*+&5~gu{R9XZ)mVDKt$6fRGJZ#2Eq?#q2<#dHp-}^xSkl>-Y`XO@XS`=- z7esG9T0XW?ZwY(I`UflnYvS-l$|*-FIILVM^5MwLXR7eYmjcc6^N1a1PN))RYM-=+ zBf`o+84}H_9j8@Gz-Uk)xd|zftea)-_X-Z@vm7OR`H{x@5R;I__NyaVu?Zty+rMtC zK5p~oY{{qbJa$Ag z$KU4ua+TqbYo}#s7QLl}VHXS}a+!6NuzzYJ72;v2n!+huUAy+a@LYH<88T3>xmBpc z6e{eE@~+VaDjJ4jvQ z?HeEBQE_DS(+S|y-O;w7!5A}or4L!CFk7n8gu~l9seJA0EeOfq(ViRJe{~r@jt^?x3t&F#^1Za1J^V`mZ4|9pj})VRRY0u2iA^xz8)_;)obpYg*! z7Pm*GfFrj0S6Yd~_tu|ogHNWPzWSqw`nN(`o>@-f#n-c}be?q4Q&K_}*C!Hr;?d>B zlcbB5pNzC9w}$u=-RWCGRv8Q0^#+E+%%Es*%Z{=uyJ6;s2`Yv!qFzvCZ^Iu*J<$B0 z(|baDj3}6{R_J@`S`Hy=t&Oui*LoWlO=-5ICIREOo!~#?sXl~%59mhZeIZq2n>z=? zZzxqtpYzf8JW&@5hzCU~xr0H|LuwJ|3jSrVyq!Gvcu$wm$BC!dXp3^hrWgOQNP`Nc%dBY}<%fhUoE0_wAMffi`O*(sel@bFDa2hSbXUH^muxd( z_#}I3!l~7*AXl${>BA}Onv?(LoO;1~zvuk#EdNU7h`i5q^@$Jl8 z_ON3uXUN0`R7$ZaglpLBu(Aj}x@Afs(TMe=@1s>p54}lEFI;e9_Xf~6;*X0C6UaU7 zJg?kz^xnI#*FOpOA{JVlTP_hrH6*LbW^%0CUm8zj2LDE1>`?iqn%$$vr(>@jEX$#u zb-rf3ex#YqtDREc~4zufqoGpIDwiY&q5B15`m&5gdazaSypnINCBrlnNR;)exKDK08z# zhv_t7y~bA;Z3bI;lHJ%vpRkV9Db!xBO3*PY4LlbLA6;5b7VE~iH z!nk3U76y@YKn|<+cs#o~=ii;L^hgVihrDu;a3_px56`}wiZ!PO#)wh6Qcl3qe-|7$ zT_-E{d-_Fpw@PVwnT!YShsyCLNCL075!8N?2IMsm{E;&V!V2FFrLJ;~?fi4Wtez1z z;H+&@815w>QWTnoTaw<39jYThz4a+)q{z72r_t{aeGaBu@8p$V-f@)}PD` z%R48wSXnMm(K9hZ7G@7U(5mJ-u+@`3{qmeQ5>u81|B{L`c}lLSs|8KT9dXh4xprH6 zR_Toukj`W(W?{47g@kU+(vv^2WoAXa=8HG8&j`u)4NEEP=j^vfXH-0Tsm8RMt+$6j z4bdw^`H-4O-aljq*;;;YeSXaYUek5GuRZAP1$X)b_COpk~D@b%)x3 zhnQHuke8AyzYu@0^(?Acr<4oxLoPSBmkSx^o_FL85E;uG49*kS^6f3`QCGn8L;2uj z(>d#%8h`8a#c4}IihD~^iU#RjZECKtzEI?sM^5<3TpOM|+#B#!2kS_sv`^%hTq-vq z%TWsWp&l|n+O%ES%Yacn4Q;aQ+GYugk!nzWn{)OCz@R+QJFJotULxk;$ee>@LK!8tRQvr&&?>8rJL_rS@@D7v>hY{ zwFVcrI}FgBA8&_$3*9iDvYt-C-Yq6cV>Mv@L9pex`LEYJ^sNsbSCA2d`Yh48XMd1U zf)g8W(5#uT&u(#t%UlfF?#21~H!nG2unbsiKIw9Ahx5$&2`w6dMGF5(Lx7rvuC$o- z#ca|XHcijCtv#lTGfcnnwq{IswR){%wt^1k9`8Ou82vYp&eS+jI;t?Xq4H})%jNb|5n&ew7FNag8g<@y9`ffD7xeL* z89$uxqY`%-p(zU4d}TQs1(UaY-grg*Ugk!6>J7Q)cF!=6+cz+s?)`BTbfz2IaYaad zti)s=n=_-)s7+xKG0XeMZf#Ufj3_>@QrOc58trVF!VR0C#0?vbS^oG^eU*&IHdj?C-YoA=o#vo*2(r6R1x(W=X)wl}%r%Lz`BO$l)+TV25XZ#O& zA=&5%CBtNXRF6U_I4z{s&0^Yxdl1Rr9sIEnkmws$y#Y;)r|5v@ z#}`fN?BO<$w@{_;GFA;~E19!fF>n0K^dQ-lu7O>LPy#4HC%8*VsvM-8)x0M_YeU`) zBvp(_#6mNX$-DW%tfI4r>mge>#BIf6JCQ>HcB32y2pz_-hNK#k{jzf`T(8mdiT!kw;3q&nu|=z z6x76&(kJNbtkir4MI7wk0}tlG-J#?ls?rVhQ=X(DKE3gS>MI&0`T;o$G5VYf{U|! z8wmBZKbq#Zpe1nr?~FglX}2!nFGb5C){ECB*eFS_Zmt&PTw(x-Ba`hEnRgjsj( zwEz$(n{uc-6@gPcJ&hAnEPGqU6@B76IKwa`o}%B{=&k{1Ghm-PFQo z@1|^HoI*fB+ut7Z?ukFTY44pufMT=WTNc$*#Q(c_@7m3nG=5Rek;O@(Wanf7|kF9o3S)H;o}z%xyS| zpqRH$-w^fj*~4(HSQKwBa0hP(aK`plAj36Bf}YVy;1jZ}GH8J7n`!`q^(~WG9cEtv zw~0{9O5tt_Z5YXCE({(aW8wVPqNT=WJoE>J$7t_9WaCAT=T&^iPM0uxbK1(6`mknv z7Ek~lOMT7($Gu_SMbAq+NEP%xZqBtxxl6ZEgUiY-TaR_plis|+MVGJIx)V`nD} z)FiQL{r%{gJ-cf2gW|gPFQG~Yv>aOXGZ=%`%yOM_-979TXkttn1B#^qm;0j!0FQNH^-xY+xl_S_TQJLiaCBC9Q><*E-p1sc*JLv-Lso#JwE^HI(~^$XO$AI zV^nS>v8_2n6H16HbKUvk> zuZX?t<^!#Ru&TBBI}(<=mUB5(mC~SGNlQy)y?8``<|}Sc4r;?$rAgm?+C$aVs_wQI zoo8D2`wM3Sj!oCA5&iF6t9tQD6uEEAJZV)(Na#52bx_S2G;Uu#;$JwN0~;$CyI(W1 z29_o8w|c7>Rkn1H4}i3()Wy!awQKV%RMFWm99!D zge49-CFJ-iaTKa_x#RG!7$fV&tDx4+M8bDF?;rZNy@}c)lGvvF`>X$iQ8dV3)4|y? zP86BmbIJ%)oyi%scl+9WhUetg&MX~!b?d_`7IZ4TP>o8RJ1Z8|b6!n?c5UeSkdq7i z#|ph^q%o|i%4n1PpLj`j7T8Rw2|Ms-Amjn~cc2fzziPEZ%5A{%YC7@ZIl5;<+J*>> z_SSg5!mb~LV?JlcR?yjeaP!ALVnq6`K-BQF^=0H zFl+1xNU+N#51d`_gnLo(n!lmbFqx;c#33N|`Tc5{&Qr_gDZjt;yA2l4)rytiE7PuD zg*78Yy)z2bec)~4;bIEBcZ7ayVd*e?w4p*|TwX}!v=Fd^dDnWHSB-`L1R_O|!TeO* zLFVc!1^-q#`Kh8I=y?saZ<1NYJ}2PU{&fdu#z&#o$RD?<*TN~%A>}RB`zfH0#sTZS zX?^G5q$I`#D|*5aIR*j9F&tC!^P zGT>#OkN+cCvzKMpGQovC0cW771;lpB)BE*|N@=02%DYoAd_XfBPRSv5MnnsAn!r3? z3BEuej)Zyt2}qVy@I0xOjxK*7$*~%elH(#&QP-`Iv)F9Pn%e!r=W%N>9p7dJ`E0l7 zTbySsIXw?;t_{DU5Rk@8g<)g~F+U(=@lJW!-U#|X4l7}B5LD6UnPktwpDdPtDL|~N z5H0=bQ!R(1`5dkETTyGBwBP*)aW9*h#Fl8!Lw%X-#PEOp7%WYW9Qn}-o`!aUhBKQi@|kcdC)`{cTk1D0>^B5fPX>Li~P(g-8G7-;Z7D9t0i z(7~osLUUcRAeMS{5B0K%`m@{vPIft1WV}_^q5-E#zgRWiYRD1s?!T}A<(GgxUd#dr zWmHY%K~zD9z!cFC5P(hHzO@A3EY;36uJ_?>f+PCcDMg;sMbH5mwJeb2il)=)0+Ztm z7&Hh@Y;f(_V3O2w@mD%ca7E|hK*)zc7gXQSklOM?KpLQ*r~yH>7f@p6Lax)LMA0>X zv^WIbbDtPzfkLJgyawQ^omc}ARe*Aw|9fX%7sx%lL`TJ-I=RcCn!%oDBrSR+j7DYi z;|`&VMSDHpUN&?n*Rdc4=2|ae7X;9`jkS!G5Fu)+4()0?AzUuc^i-Y(>_}bdI z9~{%JjUDI)pmAV}@3h`3BI@ zSd^`S4oU)WQ&edX)u57E1gd+gx&`tSKb*kZ8v{nLPk#dq4K>%VNemJ(xV}a9cc;K2 z5vW#n7Bqn8Gc*ZrQi7?DkKG=!LJVBsSRE{>&7YL*?t0cl`1e4HKf%yQoH{X_?=uko z>ZxL!BV{*+rerTdMcXYZ%bf>s}-I}~a_G9d;f*1x5RD(ar{rEv|c}ez=ST17<-T!2Dye!OfHj1?HZe--r*Eo?C z`<|=k1%))X7LS^HbC5ZYH1N?#7KNY;@Um%LA^>NM|AWkeauJEt65yg}JB^eS0d56V z6{e?mq^nc^rjT`Op80!MO`zpMd5VuS8N>-9ehmi0KK8XfKwm^9P!G?+!Oze!C9Mt> z(m@3Py5|re&En1?4HsoIY{3b4)S7k{T8JTMyC8ANvih+C0fmqJ6-Ccr2~iTfLi;I) zAF+`2UBdWCl``wLu)8{n4ZM;7Mo>qdg@DI_ElAu0RY6-ef}SENL6HDk!z|9U?+_3x z^np=*&!uuCwsN5frv(~&bcNhBXF)^^NQ?-euvf-iF8EpWsOP?4QX(>PB=!viP`?#= zv))J!Fit3^R&;Bo8yu+@qAuOZ$^cKoE*mc08n73M)9%`+o7A~OItE}@0)FDz!uQk} zBpj&Q{5L0(fN8yG=TH>A-73X-EPDVnQsXmsfwzcD8Kr7xZs$Hc#!eKws~W#D);4_sWR-4xO`OVzQVasnkSVa9pS0$jQ_BEsAm<|`NWm3fTv&B( zC91ekVzL#i4hwko0`|ICWPyT>UGGdwa5ktA)prkKIvc_-TRneuUE0K>_W_zF%F`)g zA5SLo$a$e1t)f{TIiE<)&)JJ)9*^1qZ&J4!`P!^DX3(NhS9>~4eR7bkbmX0){x%zb z*!*=5?i+uW{I%fEGQu{^qNc-?3Mpe^3}XEcM&s}l2s)x7qi{e5A3$CMf_ZIQ?+tiR zWQ>yVD#nMmMoKh$jsQu4);)9z9Ubw*&QVWc=jbpHXu3AuRaqh!P06=A>*=c+JqQjn zK2#mO+_VEZt7;p+2#c5wAw*ZTN=n2i>L{5(RrlH~Yx8nXk~SxaiDf2%4h{l;}!_< z)F4WSIR80PvatG0{my4dnF=E5i0+Oc;D{>BC>csOkgFKu_c~9VgP6(y`|?-nBrx)FYP#%9}P zO%ov7euk(2&e>>BPDqK}fZZGjQ&I2ux~7)rMgyV{FSO(-u+J5Ql86L+4FwFL#2a`E z!`P;c4#5s{p`$e*|F{m=nY=vnui14TQh{4k(Lm=Ne~Z`W zLwBlA=pYww5#42`>4+0Ea{BA-_oM8!v7Yx$y~0k{`D`sAMQ9%4fGAbvK!DlSg(J2m zl!=ua+k@~aOA>CNy$#9z)={jA8$>#@C3}h2Mh7zu)(4Lc-Zbo%``)803uAeOnx>A& z*jIw>KL!XrdajF|n6XzTqOt+I*jc{{D2Y2$wGj9kf*&EUrRc5>_IJ)QbzuLI7kikVR)JL>l!(ze^i4HDS-ah`}Rgk7Exo4Ziq z4>Tt!=pm70H~xl?mFa;^wUz?3^}b(Ch~8YdHrEy&zPRr$JP~vVu0&dH5>6ge{h=^E9(9sI0-PjUr>_TK@KaM#a%Mn!pFiGEvI4pQMb+5$vFlaU7ZI+?MI8#mdJ57F z_kaLh&y8xXqnzp__xbROc;vbKeTmSKhKz?dHCQa9a6*|y&|!5|B}U1IsbFEdVa;iA zRnhqE;TL~i!W8k26;vss&j>2rQLR8Bk@Dgo^(>{p;LZGdjaY<`UdYC!o)1$C$KT2T zeBVd&Pe2CtNfZtn1BzvKgEF=ma1~}KaU_|dsu$)Z^iau%x&*}e&Lr~9NN`nB9O5+J zdveVBBv7Wh)FKw9sF3cHD1BL};C8Z$2(C1y6r9t+jt z&=qa{QsLIgb60CG=ogrDlL-%H?~fIa&K8KFz3G$>tWs)&XatR^9epc%Kso23x)*8k znm=gL&5R9}VZTk%z?Z>q5EJ%ig-#iP+(L@LZFLp!kAU)1Jbq(VFA*&w6=53Ysjx4I z?ruLP`{MO$JnFy*+=JAsZGMFfcWNgK-~*#}XxY%;;`Rf1#u>3A-GiKi@^vM_^1l4)NWrEtB6@W|O+jM|eTDn_KbPbj4X%elEnF&349F+21faO_95r>xd zd^*^w)L^T^4A0wia9ed5Y*o*;TIvAu2*YwneG5>P-Jr>KH7#6t7Mr>x&4wYc5ng*hb1E@Nl z+kzVMXzu}J5CS&WkVDFb;iZ{mvjR2$vvZ(p3Lmr#y}T8U zNp+=gOsvyb>0~IkYMgvWeWRpbxJPSZ_1JBs=?R}EGW2^^YL-Ot6;RV)H!gj}as7`o zIq=UiY4>*%Vg@d_=?N9WIv?g>R#!vNYvPz%^UEAK&{ZpbqdNM_>X+9qp8fjFF3^gg zjVM9rA2fpW%kpM`s1P*H*N-loOK0*g{CZ4Pf5)Rx+)3rMO54jxriFC3w)kc-+JFt0 z`@fs$k)~bjcn7&%f3pA3nMDE=t?~(%?+1hubuy8&Ml*ibEE_a(lh5f;#r;e&{2m{# zW*ExUcQ|6}v}GH*IM$hIEh`pw<@fi>Z1dlZKXB#Gw0PKH=#6y4bBK+&0 z<>31Z8ig%aO7CV{#&;fDKafA_^QY-E;OT*^XgB(uiGij{Y#nj=6P{uZbGj#4qs?GZ zg18u^kIJDN+Y-h2TGW?E*_WHafh!tbVLY)9N8l@e!vnc{*cNSWfSn|E)lsYa{1%=UB02px8EOcnY}`x z#gF*iI&P{mG<0_qQ&E?6*v_;MTDJa)BfLAy--6Ex4V!&iR+l#iO6y^!;r4Qmfp#fL z?=5a~&xx?YO$UyRzj`heWB2VAt}HJrSSny|B%Qjw^_fw0n@rb$XKC&$Xbv!AF4Xz!`XK88&fS{IWD9^x*Sz2cEVnfC1858jBF4!)Bm9G z!=Yz5N3Ez{%C#edV&E za{(GrFMg$4JyA4596hvBHP_jJ{Hm^DhKxH*Um-89_51?3#!90g{00^>&}OPr2K6e_LIS_s9rA_-Xf}kc(-2#{Ez3{De9Bv3kcZFr_PpBJD#7 zb8)Bsq1*D=mDVjcZwAaT-zV=gn5Z1K?F72+n_1geCN#$Jh6No3z{o&Bu4(?82{AQ{ zB9h1k(zOK5mbg(=$^}QUy6aVI6#bok=Yl^?Y4693@1z%Z;P|_kY~x}fxJ;#XNa@bg zL90exrGXhPHIX~dxDNxLsEl|A(Sg$3L-o8~F+51*+&{za`0*!g#p&Ip_K(e0r_`Go zd32v*LcnvZkfml^=RvECsDFcm05T^v1;wHia)Jh|K&;~F8kg4uo+eYMW=JNoL)!)@ zT}R5>>a9PPW3>{ecS%Kz;ZxF-+5P)e(Gj4;KYUg0{QHli4E;-y+2@P37ooNf~Jls^8cgZ)G_${+uuRg_lQEC-am{^$DQy3i7nRpLS;$@gL9o zX86KIq+0BCiBge8-!v;XsRFTv5KLu{X~y?- zVC13q;6VvByCW$WjBV(>q&=!M`2J0V!Mt@s^uE(_fr7izm5uHMN9>Zf_rOxbD z?88}giGKeYBm4Bsg9fW4>wEangh2@1mAKU8JhT%6Xz?UU0w6 z;N>;@o&+oTQX4qP|%egfjnIxvoXGe*bAD6paY$8o5D;#wELczlQJznq(zgi-Iy z+ix?s$Ps4WcCoY!{Y*uJSQ3@SU3HH`Pq!4$#T8}g(hG*XK9T#Q zB2&44LVFgBCD7vZHcDlze46PhV4F?vRK30zSq%Bu<7h89~w6O$?RoImw z<}E=tfLQQGW|M8sDQHBgg)tl1yy&{bwJ&&XYohuu$ysDo!T(VVAX#y{-&bfzO;VuD$G+?v%;{v`}MBtJrXIsugTy5qW~)ewi8i? znELT9QOOJ|Vlk~;yIMbXF5`~!AhEx@^N0K0^zGqbo+M5bAd}EC=}jR+4%gDw*yn~e zeo|imDcbj6^+4NzVyHymgspY--V|hou~;MZp0DlU$$Gi9$pnW=rzP!u_=4AHvLF@W zm#LFTIfyPxcxqyR<}&@fAokm|IebYc< zS=aboIbEmg_zAsTDeb8A0QL_o-o}#t&g%Yc`?khMVUyF{CSNEPyq?@I$+CfzvIq|F z7(sv{=~Y=k)~ww8n8GFsyZiTK57Sxw{-g{**ZLQZK>($kIRtGpY6CutJs>#DK$Y4# z%z!Y>tIlNsUR!a`R;u6-opuO@0G9rRkz*ThHORA$MKvhYIpVESprx+0=>fB97>r ztEcop2Nh8%JPx)L8bZwK&V zm0lC%#sxSC7aI8Q0-P^n(8c=(A2Nai%bw*Iq5Xf!1Pf*H?LFvi_@{x)$_Zl$p-ddm z%i{G*TZ9nOwE!I+l0HKq#;Mi_0?R09MPPcIzhad;#eD&b$+X^66>DQ_-soLA^-i5r z`;HzlwGg7Zgr38lf>ciLmfbr87nv26@R~e&G2_9pf7ggDWEUOxF!1!ysfMabWFN$@ zTOO@{UP#}J(@q0sLWwM1LR+2%6vmdW$wI*N&h)XDaIfvl4*(d@ncxGQ`}TxlFqrrx z3ixYKVSEKz{JgL+0Q!$MqN)l4dSBGW5AQ%qML72^+J6#F$Gzu_h0cjhxOL0*SQBT= z$<4Z=-(J#w%|$NClmQDs>E^^FwKZhPpY})i5P~~nF=R}tQJxNTBCnpFO)*C-c>HYx zSI5}%+vGJeFvxi0BS%Lod#OG!vmjG>|DYizxBktN*Qg?B`27>eoY9>sE}&eksrugJ zD+tc2u24dC1J22Aph=jR+OMYO4iWGw#>b$07$fZ$m@@^EVTOHK!cIVg$?m4E?~cJ6 zF#awrm_<`g+^-k^I)uz&xg88884?_%=&XS1M}{r^iSlK*{PY4!8zS9S0}P$Kl6TjJ zpLYg96U7uY-p3A5THvRTy}oyIZ*9>MuaBz0w_zzJIvQ3#4MfCpuf-GulB2Cm|fA>{{_wfpb&Qi(HM*2*27 z4#PzLT+MJc8<@l2$k#ZZe=V73L2KB!ahd!unKioGv9EXN>%*xgNlNOik8Ck4QQtG4NBX^ zG7otGhslgxMsS#Dp$fuGHniueqeG0?Qy zarKZLU6BxG%usvmcb)N|dssM--dz4D#$*-;Pm}>r``02NI6_`E{1%_e5E+Z?n42dU zWfF4Se)Nd!riV36>hjO%)%2AUFeS2i@63Uk%KsRtH5*v#O6ZqL5i~=Kek=5{>8PUu zn$5z1?!0!>Z!|5Jg!B~>49>!9%fr7e9c}ZekRHmT#`(@Q?@7J(l4FZZc>d-GTBV%^ zo(&A=5Gg&NjvYwRe%OJK*sZQUo^DyeuTAOt~pN)R! z*`I8yXzxn;BVN?g&99nZa74otKV|wC^hNV|Q5zvdX^{oBtFw!br@?6Ys`;C&*KNmS z<_+gh-G_L<4~7cmHKu>kW-<@c#8T5mAf7`w9-@s%(UEspEhraOlUSUfo~RYjpL+qC z8n8AG1U1a!xH{2|!579c;PufKw5!Bg=bT&0n6#irQI(5)wEk1tpG2fGsAwIWPwpRu zWTZJ_gWh@EZc#ebAdRl+1x$(@&Z>`!Kn3f%w6=Ldu5s|=QSF4n5M1JX1Jw{$_p`D+ zDB_W8s?j3{0|uUcAMBTzF2^#nJLjHx5?+ksuMG)fnJ%LA|2qHiA8QLgKT|q-Wug*v zPmRWFch&?yX0%$f?c)sj5a}Q_cw`}74d{%ca>7Cw=^eOt!y?0}M$V)sOXoM$622?z z5HmCeR#hJW`j}8-$Bc0mY&AAd%I#Sh+qOov>tfJ7C`pce(4?(gR|#EQllo;w+m;-mSqa0d>Lp&tp;6C%<`d$6WcZUq+j3a z1=hYO(ZWase8^w^NQUjW754R4>uz2%0;-X@jB-`zYF@-FX{ zss~y$IxqQreCwY@dN!NB5xfS;jbc}RDk}Z-DkbJtzqybe`yJRd02IgXKl#qEU%H%e5MR-s z%q2<^YRA|(0|Ui!U-Fyp!D&CM0@UQ3n{r|?(K_*r5A@5va^9y z*6rZS$#=mr-d5e7-A?F(YFUm|G#bpf#kE=Q#IsXwSNm`Yqv*eb#&!Io2kd(AyK^At znc`LrsCi5;#_13lCC?!pOFOc%bMU!enlT4|VDjUvH66yGAUx~QgMn>QNsli6^}2FX z1b&N&V)K8sT=XuhhO7(jT(XuX*;Azu=YPzGgRP`1hAM+Pp3whU zY6+OxN=XR1u-rkN1IjUtmELv&II1^4C9-A_^r_V}C+&aQ%t4;ttO!6<=U(ww4+cDv zDdDg8)6bSeR0{j!xIVJ1Ql48vZ1toA+99WIGYE?Op#t`4t24jCk!*#{)3MYyQ(nom z9*SIUqu+bP^?r?f0}vVnXrKV>LU_Kz80m$Qz&`7K!?%yH+|6B!PF`85h4CMA-l~^^ zcm__fXn3wfM#l3n7H-=ErD zrR{9vs(UY7F)E6J+-FCQW;H)}&s(-=IdsR8a&Ko2QSdZ<$LBHgqA%6Tx%cNsF_c4d z?bil>nfXA5B0^*ajwR*Jx|H`qlmm z3&8az!*`dV_IHm}vk`=6Gh39hjOX6q*Zp6s547oi5OeqbJDOY#mVi8t03rvZET5@i zKff%etRH^i8*#AjD@c92dY8IZ=1u+WC)-c~Zf?~dPi^+(LFOi(A~QdH18S4t|D$u= z{B`Ri8Vl5-_`;B#1LU3OsFQi)!3Up!g;t`QiXJ3Nf^({cI|H!mkk*)pH*2i3#MHFc zq_I6{9(kpeka#4}vvS0O(S9|6Eg=(EM0t(R#FaU^K-qY4lj~_gjKS4_h1Myr4nS_7Xe)npz@(W*DSwvVx08v@Csev>X^xM zxm<=SyN&hD5ViQmk)5`Q!jgBvB3hzqut(A2$1gyeTj>-QgmI)968cPZcvb3eKlhY; z-xEIad}%Kg=t0w?)#Qg^0Q#U7p3bewdC}^>Q=nl(4m@Egn>d60&Q%WX138`>RNNX! z;2DBdji+pZa1A!u;lpz`8tNwoTEQqaUOsmsx0-^2qRwks9p~7yCiqd@x;5lYs>aMF z=#m(|eW0m*D_IO9`$_aOUU>NXFYT0z)5qJ5zFAfw&-Wg?MtUY34=Bo@^{EWKY={+JB>a}$ zg?RF)5N~6N(E8T2pu8yoy$q5*Eg+<+Ml=Tf7|>|mnz8Gx07M!CD3(@8JNhWH?!wD} z_QIzh;VLJaT(^^PO$-D;QK&FHRjY%SfjM+SHq$9F0a0RR#g$KSi%O`!f$lsn+QsQexF~}5gC)YnB3H;v zJbTfE4nsc=zzWS4DQ_&xe#a7zg>x+@I%0oTDkyIrTc+!GKo)+=_wa+%QfZ zSYsIsqX~w#5JMzJsjJr96)D^dqFQ+5B%sFFmbd~ zPPsvHs3djvU?22HU5%7<8y%I{Q@=x(Xpy53jC?o9CFy=#E+FN!-__%=pv- zYCm0#He#&o3s1SGc0YcsV$8WVx%5rr>Gh(A^`ZqUM+^{x2x>~2Mh`_bg~m@#G=iKUhAtS`vpWG}#U=ZcY%agy-8L|0!vHh|SB7%TRR)+UK8a^f z0?m?|GhrG9;zNn^SQ&=~u?8=Us=BL>%oJym1E>4TCTa!O_dWjn_7VWlUVz(d7%02& zUnFG*0-RV6DL^@2Ea7`&X0;w`0*}!z>Bx@h@k%oZ<+&=W7A-IjRs;gZ9xD9zKau-( zkx_^)nZI*xyxLOQ!+5N0Ef&i$RnxW8+Hhe1rfW)DHMj=0`Mv&)d;F;d;S@gIX$Iw^ zw{787iaNR1EV8V||G~Jj2T63^12B_|&$q&D#9o=(MeG5LHD(nGJ%vR>cQl1Ygr9d- zFa@$(0+AD%-0%vUw8xmgpx1}FW_Ytdh`oF;CfA(Im}H13_}4N&>7d@*y^5@f@2@Y?0`HbFKyiK1!qM?+e}jQ<725 z(x>qkghWJ%E;x7sdJ-Hpe!xmz3V$eC$-E{-3UV=c4OYp8#n%&^jy^+PDqC=%-hXD?@RoehPIy0Vj}$=mfjvEp^93bh0Xvuv zBuFWw#B1hisQdXAm{?gwK2je*gIFOSjP&E$Slrx~f&;d}-1v57rqF6P(68x!A?b;q z7Gtw*!)k;7-FumZ=_Pxs!0;`zdynYuC#~YU)mbGNM zEiMTS2QXNDFlBFc4()V1WJftaD$3hc0VH}BQTYrM489f7J`!muE;O;25JSgarAK29 zsKMQd42i*?f$1U#Y!Y}ocFzl&*TI_+?2Zm_2DFOxqIzt$&;u|pkpyOl$^}Bi)Ny>( zDC_Ep(AdZrIsEGc#F3O>Ljmo?rJneFCDWa1WQ2E5-zsfsKz-19LZB&2~&PL7k5a&y2R|H5cRC|hsmY-4Jn0qg^ zkg7ytL_d>$CYW2b4n4{|xmE(f^I3>Va#2pT$;pv)qHF;py5e)CixaL;U^*AcyeO9! z71dlX@Re+;68&AV+q?~Vv$H8?fPW1N=$!iiL&9}f-a@8R*9O#nigz9o!RV*Sy(6?0 z*ZdDjn|iXaN++z{V+EB!2)2&wlyyRK4KYJLL4F(B>z#z2zXY%Z9Ka8aKsv6gVVRTG zqlu}?0A|3Biq3^O)p%_OG|SZCq^?))ADFyzp5{dDj&#y@(y;v?3O508s(yu3+J2T) zhRVGyH}{GuyGM*ngD3Gf>R0mGb1Y2xY%y=CZOFw_k_?S1i8GF7dOmc?_sAp9uDe$| zEai0YHnr>kZWkkIj@hxaHdeYi`lH#RikQ6V^~L+2I#|TO>A(AURzRq|$j7kI_33$0 z#_pBQnPliI6fYbkz;epq)~h{XB^T?xcP88%g-)An1xU6f0^)2Uql?o`QJf=bbYu-L zDuChsIB0o6&J%==32tFRDgtUyZKyUxwBjEuuQ7@tb82<}M*^8lu;+S3IQF$OUL;!NSqx z21<)YsflW&ZRlhCdFP_oDcadoiRUP$KD-^fHFc%i*L3SStLxnkbMi<$@$sTtL^aj* zt?r;N8QI30R4DDbc=Uz8$qtV?Nh7l?bLo zY0s|9>MOBWHs@%ER8do5q_}4*ox5Whn=#LtLaLb~jJonR$WCZfaZ)WycT;LbLK?0) z#u|1s6u12TNF8t!_?JHxJLDEh_cT!5Ww%L5DV^d(O>7^~lh083MnKVzk)ptf*D~Pc z)bvcWBdi2OsYq_+p;wPo6#U836PuL33K7ilS4Y_0H3BbY;*o!E20v#C*Y%I4T}v@Z zX7akpt)mD0*ANiQ!4+mxm7#SiOFG&wQ!8Gd_DkOPYCtfU<6^bMZchkuh|$>nGR zWNhmng1*cWpPw$fI0Qq-!y2{)&rwDO+5D=RQDYW=5Ero5|LjH|FS=6#zg_{H8cXDd zsXYVV-0JO1(0q;-|JX+*>bE7{>5ZYQ6kaNA?FPD{?$l-9r|nf)ZoWs#6u8~R+AS|9 zLY+7Ex$2fb-&lY3T{@d%V6fRZ#pvzW&4BQX~dSgfU~AUgswNY+Kb)J-lj0}p1(`#Lm1 z8nWhDhhMc~x)6o6J;!?+@knxty~oVLl-o1>kjRA^R;4zNy;bc5(kW@A!I zb59+ghX4C+rNPJ4F4#xrQ`P2seSNe$TTa@nas8~z=dsXHP<$Gcgq(YZ0?Y>lS&g@J z7FSZ{YtGPg@KX#6<<=xpH&k3NWfBC^PBO-|$5E_x&)m$1^T9o8JgUOJ`F4Y?whH}wwcFy)3R zl-?#%avi6e*Vk0}(qW0;OMVn_Qp_?%bDuOL#Nj$RAt?TND~sBq6a4uzQPNKhtLsY4<4QzI| z#?2rE5WE&xXzUS1rNAIGt=gBcclc?sD7|RIf~|aw42j+PHFReX5p5%wAhn9GE8qR) z=3dMmC!4HRHlg@;$Xh4E^ZvE_vv~TXmDfLs)h{wck(n4hfRdDu`kI-)9&=#{eYtAD zTn!aSrKJzjp?@dxd{+X*Yy~_y_kf&^?#J0OtO0P&t+)j^mypi&!duEnjTMHSK41V<@$h!c3lAUhbN>1Z^l{B|N(`v30M z$8^Rn1jF4L&}ETyyn|+oYKdME1XmyQ>uYxM;EV7{+mMLId=6*V^SlSx!-Zt(3WfV} z!L89bnN!g`u}(XCO!{A^$3mWvtUOMRSr-Q43`+aSFlyQzQ|3;l>y(t!`CW}q@#$^#63w^k^NE<5_Lqs@u7k#&*0-uHWEtKQt%s^cWA~_{7<~(F- zB53OhLQ7P89^X4X_MB(JZb}5xIj0Sb2 z_aC1{7ZYwxcMSjacaJ?? zW%z8yZ&eemg;tVY=M#J5NnUw-OsY5^1j_;URFTR)n_peEc%1RI_rUoSzh>U|6kNxl zguUV=9_+%`>vRYmdl@i=1<>A{i-*E4uN&&D=pL>heQYc{2N^PGD}!1Q8}i1_5laLq z$I$M6y{*SVilJUt2P(Tiipr{<>CRxF|yv44QB63f??sJM%C{tRT`qg0-EVd1e_gPk66`2MhefGZ*W3e3ic z6sZu$yAn&zpzm`R`xtqzO&09fn4#%A05|m{a=27kxwn8rw!BUhrX<*Uon(ekJ%VDF z7xxj@h%@ZN_-;5-++G~@r~rrql{)EYej|wze4gkHRaPSI{2MAykqqdHFyD%U@0Ec| zt}O?Lw?5T7i`P;z$pTVzqqjB5nqv})-g_U-l9dm75rs4!*umomPtKNC>PU@Nq;r{Ce$njV87l@xG4RMsf;L zdYN}9g-$5s6XikZ^_C`gEJgo7XB|KTBT^revHjp{Cz3%M;-OVuc(e%V-^E%m%BEO8 z*i=~B8qm!TxcgG)GZn(2FAelNHGa4rJ?p(cx1oJ;V|FSE10IJ^cV3OI>U&CedFND% zlWlK?tSrTgH}R}P38IJm@B$j^_MGqF&5+`jRf;&{HM-{A)Xjt|Cw<+iM_q`&Rt0i* zQcTW;n-%_L=T2m0)rx=tMYP)qaYxmCfGcvNRIocUHJADXGp2p zImw8$gYGUkJ_`)Elov^h}kzCl(t%D}fh%B{<$5tFV%5)ocr=s!+|gOrR^}NIPUA}P$BN&4-nn9%)L;jK-B&0C=5r9-TKpbB_oN7 zQLBSrU!aisnwu%gd40@W~g_rFWm9 z%#%s~S&DX803q(o4?wkCY>0=D+id=z6jbtOnz4dK>&#*014ZrRK1)?_bK*^9>2#celNgs--;h`)k~Eu#n>uar$0Z z`O6b-Tul0B5w1fDl-v88ZcCuw5|pMv+a+$T&NkD>7yf*NV&rHzI}x+r{+ojcT^(e3 z31Dd+cqXj?)e|3H$RzVPomYI~bqsOFN9Ergb}q(qFL+4ZHSQUgJ9}4{WfNymBjgCc zF=%C|$80Nvi2Mp_^W))ndKmZr+Hat~VNx`I8^>8?HC!6|U+O)zx&SKY+dK%&r{n$u zCG+^O=ogwS00^Im4>877nMk9`^gQQrfMwpW0kJw%!B0XV;IGQ`(FI#(akc6+{!b0T zAOH1j-8n#zc}|LcdsS$WHXYBe=JkwukP7xCjdOdanxyBVUB$ins;-ho^Y1BfApnJe z{ck+hPkkBi`Z|76(Y?0x&7~d6zK0b+Kr5uE5ZRyl`?a=SiRG^Cjb@Lul5e4&QS!y% zBw_wJRyI)is0h#%5`p$gTtC~foKZ0N0n1s9y*4uX>)tG(E)Oc`ozc72qcVfUpb0NU zx=aZcJKO-j)yOYj1xxG1fy$(T9Ese+TCgZc2-a*rzd16xN-78qgt&?J@e#eq2N2p8XzFi+s1?+7LVThGTlmM42Pj({CM+C*5y8 z8$vdc*&+?A!v9NwnwU`cAznAYr7OQ<7?3bUREctw@0G~f;%_~Zyj zTqp+kvK*NuD*uk0vX{Yr&D5rm&v*ZH!wD_8q$*o(k^bx@Hf&7zQnr2OiuU%61n<_D zp|ybXd2xv)+iwLEeA~_ zoHE8_90ivu1MKr?i6+o^5@A4qb%5nuo((W8vc#@ zS4iYZhvbhZZvov2mnwmi*8T|g)CiZ|%`opZ%c1GVUzzAO9F#*6sRx0wR62%SNT;IR z%>&3`yWcAqI|binwk~0u<2AIx@LS1Y7}t9sfWXyX(aow5s;uzQHOo&9sB21=-D3D3 zHZ~6&lz1j#@sGFu+F(fKbm1l9GH1urB^=iqGek-)l9qurd~g zJv7n!A2-T1+)m;}736Y-@ZrNj+`LkzXuL8pK!Se8!H43}iZzJhbqc9G`stKxmN}JA zG_>}8&l{2{Boy6xSzQuH7=5o7cV>}S*lmnwyPyJBcwZ{{rx9l;rMUadzjtfBUF&F( z=~t)9k3rtBhwu>e>fqH1)Z`u)=!1K%WLAfwAmTIxrSkuLq5!WYdWCQ7y2Uxw7J&)b z%VhZG?M94A?eB$2(5JU_unB_D3JwrvW4kJ~wodc;#ym#+oZ1X{v^D16^NakCFdO^t zMjINSXnmp#i%?Zl40u@o8(WRy)g`puiFd9Cc}W%_#HjLt^bO~avkIq7$$s8rrv8g5S74Zs>X!UvFeFuj$KEL@i0;KErWGizx>#uZulqwWoG%!< z0sP=-jIjV(K>QX+M6oIR*@y{PHc4_giNN&!GeZRbkw-b&CbaVsM-ALN17e8fV3a2& zz3>b-zqbxQR|9KW0>)3HK%t-d!a^!=NkFW;qXL=1LQZD!{te*hX`RxqEs9J;4dF~l;#w&?WkF~Duq&tth zj3&EXQoIC74r#q@2*70XDFXSgL^J`^893L8?Xa8d{r9uZ+-`r|bo-Ytv%T0~=d~Nh zdkc(r<06FV6jNCnW*3Dh2YG}o8c?Q*WBgHt-`~4E*L$uWUkIkm?=`XTfgK!$0-%OS zmX3gL0%^8l!|q~}eBeF`O!Irp_*2Db)sbp8J5^_rZs{O+Nk4v8ulcFHF)rd;27MfE z5D!(33^|_l9|O`B@*2Iv~heb_6P z`_74yK7j7m0$>6Ku?K+r;^+)bVdt{HfFd5Cp{s*qL9rd8{YE;U{2QUIatAL~{YTe3h+6Xks7?i?3m@YhKku!!Cwlll}cd$W`3MQav z*+8PFHtAdQ8J>1-fLYLJJvC%J5)=u1!36y!-N-nbbHc#L9w%@)KbxaKSOc*t+cZd^ zd9$DJy<&?3u@H`bax&0H;hnuddk<86DQ(tcHllIeH|Ilcfq*~}O-OM92Oiw8;bR!GL7$|xGNA%AixhMnJJwt$~scAqz24=vf`8K7Hz(|}T#r{ekp9U>-o@))-yKt8p zqIH!;R&wGSf7_Kjxguh8D2N9YU<2#J#)w?eBhZQ@6)~;5OR1^C!JNGX5^1mMhgX`m z6*fjXZry_hLiqJ9Bi|Wb`=>U|RI+wrlxIIZf1n;uMIB4Yd$qW`sH68xT%=TkK{o`b z{(_fsofN>L`o;BUbPAQP*!JT9s4&VnT2d|>QgmoAHROh}J6Zq>FUtb=e1WP;yl{&e z^`{j9dGAHc9l3|U|7I;&1s`vZKL~v5PA~Km$h$F)<*G1@(j91gSMYk$;I)jJ_TexD z_g$s{Q|Lqn=ATT-S`En9yh*-Vp9uc-Zc=*~2B{6&dL#nS3SnH8;aUqXGdnOJ>FJZ} z&fzOD^fdAmwurGFGSAtVp#SB3Low8{%r2+Glt(6$zwq(W#{(~itx*l{Wn2!1fP(f9 zMrg7RWDIyo?H4WJ3b%Ae_g)&)zdHpu^;b`Nu0l7~YtIh*P9C}p10+PA_iXJ)f~t@> zS(5dW#cNvd^lqcIkc%enMXz%8Um1pVstrT+nM|L-2cj`osLI+IJ+_6xZU}&(@y}RC ze(YtJ`1hHxXPDP7!%y_FTh!O_(fYH;p1;>Y_^k4~@w4gQ+VSG&M@RCfT1HO{2Aaxe zJ|+ozvf-&#CsD<|j@;(@a^oor5YKU7Q+8JZpMAm&Av#tn(Omuk6?mh9ody?e;MD;n z_Wkxr&in1ZW+nQlsL_TQ+Z7enAGea5N$i94iy*IO%6lazjbEju&sC*$O8z>UjNyd* zaCy_hs~l8?Yx`7N>XO-Ezv_4%K2~ZlncUzy@liypXDb0+6=r zut`uL{D+jv>C0EKTC~4Zc@llSX=Fz&W0MG7rKAtj9nh%}BgQbS00BO%i5prDkJ z5()?i(%sS}of6U|-Kq55;I;L)hPMo5=ZB{lqrHgqY6}>uPt5Oq8JueI&9N zxH65Bs})D0c3hAh`klM}ilje)gw0n943x?Hzxq*z#LhOI2Wpw9Tay!INY1K~6rCm& zz_q0q!;C@yI4~uaU42g{wYwmc?meS@G-%e0f~oQLATpw&otXdM2O;m8X%v40Im@yH zQ|yc(?)006sAw<-_Pc76%ZpmFTGhiVMG9#!T3m3fN>3jK1aY?q5$sI>4z-Q92(5NN58`u5g_5o0TdaBrU{)V zMucryxLaTNEitR?Mx>Q$QxHUX!2F`Yb#vsa0xFGTX7UJvVRU%CyF*Wcvln9vnAN%~ zK-M$&+z=d~7JCG)I*q;_C`$xGk9^7~I?>qP1YTvJ{~ybiD(vsgy>B3-#U=yL= zh@ON7-kxSw*3=!dej^mOLIF6{#Gx7%Pu(9V#QrxLO`7feEWvP^X)>yjK->#&mI?4x zKQ(wYTgU%)Yv~R4D$juOclSf;xw`ANysclmD}EDNSoVqEH6S zpfzlg$HK}QY-ash2nQz%7^>UE!GrWrE>uMHL{{j|uNV=Gx+2OSCH*{K&Vd}Q12s4+ zr?m9JVKZ=iiI76FF1-?&#Hf)q`YKP;;RWLFd(J4uPk^OJ{Hk35gIn*r`)2S=1{bj1 zA+-kUCd>FHe-;R|X3+v#)y@8Z(b}Pl0(O+ZdLPs)#h_mZ(ifx>k%;SG;`evk(R(+R z1R_YL@x``Ny!n3Le)>PK_c;uM=Un-Eo_%&Tf4o|~f~IwdBmVCQ`UwhttscW&gH zgg(N9(w!EXjB9DyMm>NIiPcr`&Q#j3RFe-`ADl}k5tkiuDe(9j1+)ZoAB|NYp(bl?^TjMco%tJJuRl$XCE6PK`p#SaNeevv0BVnm zi<=43b3A5DX`L)Z;^$N4_W)y1srGb7Lw} zg;?<4h1rZsPHcb>vf#f$Aqa!MX&HINFj8y!n<7`(mHpe<7gRZ9k)#wBgdQYoPW%=F z46#0bsGkW>C{-pn3O;5)(ehWhs?(bGdRJ{99(B2eJACwi5kGXt*VmnHVThc>S~U%> z{rb%t4D;8kJtOD$KEjvuHvo-}QK57T8437K=&x!6uP?aiOIZfMCO7#Q1tpj`oDF!Y|!Je^At07Zw7aM=9NZo z6SC;Q*1&`vq3bdMC(%d-tj4V`m+L~)<%h+U-xq6N+O3c$(oN8C+Ee z#`rDclSP~pJ@~v=CY!{I0DnAdA z0HWL{7w`lZDO_3x#@JHDC6ngsGg`pOR+c44--K!K0B-VM+WW6uNOyjfeN{yCgJCiK zxqjNgz?93O_@x7N5K?;a??~6fvvs^tDZ!vv>EETu8%xP1OWtq&Edrf%Pfj(eJ7GgE=ZxZH#J9F6*Oc?@|#hVU|&w zlNzqD5PJ8Wtcu#V;IosaE%mar3rGXsiGO(x#19o}=1XkqO(N2>Y+nC`YzMoKs{piT#sOm|hY88!1@(fu z^garreE{%&quq*TQ_4eua9?+EwQk~jagTO?YRz9|lv`XiCK9#ibIAUz2WDj|xvsWo z^4@uy{)X}`OsGbTGw6XUEt+Lm-zWH+#L}z5_+90|tqol51+&XhFu2-oUPK8%f>;0& z)GGX$M5b*Bu&_gKmRrP*I&yt1wTVE{!j|c4-Jhfptn!vopYzT)7GGV)?xFUBXpDMICe(xfICSsFelV)Zyeu|NN;?5fU zkswQ^{z9F%IEEP85nV0STe>+?gb#zs@GSU5X;KAcw)6;Qqh%4w@V4PWXg+0Ll3r=d zyOf9fO_#;p+&^n<|Q^}PmFMr3%^ZXgR+xv?$bGH zq+gn@*a|>*699%MT z68dtvrHo!Q@1;DVME9dVO4PQZy)r17Jfky_Lz8-%8JvxEtr`jSn5HLt20m7#AG_4>5$otYGh@^vWmh}d?g;sz%2%9BjBSKoc^^_?qpX0D64~`~qa)Z%8oZ$j=`J^UlayWLbeUON zU+w!UwQKcAP)~{)oem?iB^kxctW(Xrb5}|wUx*tB^O)C3exu?oJITb>Jv%-l+ahD9 z)BG>n{OqVY7qs>C(ztXAIl4)C)LiC%#&*)LF7;wXkoA@T?@C^cwUm|F=khJw*eRk# zJL!ZU_tR}Pu*Y&hnZ>{Y7LF*FSuPxfYlU2zfM{t4*xeRE!1kU^Ea0Nj<>l+H4uE&b zEe`?;JQiJ0TQJq*HD-FzD@8Hka0QS3&)Dh&++{oD z(Ync)Z^nTEz&ACaC&dM)h9;BXA4_6}&4B@u`72^q=+KHyh8!F+Hp%kFHxfZ)qAMS; z{;%8lpR#fIb&Gxo{QanOdN~*RmRmL4zI9B88eBIp8bs0hR_%e6S*8QHx}0P|TTNj5 zV(SZrM9KwRoZ#res{Y%m*&IV&GR*$fYluKn0EXePF$o4Nq~dinpm=M&rx60ieloyn z(hBQ|Zd+TM1so+LS9y>S&v~od;tyO0r57m*NqOH?0+~%j>7C$leA+``!+rk7!3m5# zl4Ydo2IkPH9+AbdOhDk_ZEh|{8Ru(XbA26*RlslOo=yT1hy6Qr3v>6-`@i||h;tG! z2OX;=oe=H(wl9;=|7xd?g&m)O)=)bMyyYD4geGh|8P|dns!TjDM}K;X`VHEE3umix zy8&Im`i7Ku&|-E1*jIX;ksE@F4^0~WcB7A!KS=P-14V<5&g|Ns{|QcvzgN1no8!=- z52k@ft_}dTwL+ZSfiYOn+qtxM?hKH2bCxW;gC;Nj%#ZE`dZK41PaKZsbL$nE=-i^bs0Xm=)6a1E&QY{fOp5jw?2q7@IQRwA~4 zV4rV*S10^?^vf_babWSP?uy$gh3DC1hbP;3HtnDkey>Z^P5t1(Dtpq)Gzu3$^;b*t zS)&$CR0X@r!luRBTVSPAwVrG@QXSZg)m0{^+8+-d=``dt3ddX2);U5=x-z* zwJsQ-yg*0G_tC(gljy}h!wZi$FfwP|KHlCynvCGRbCWigX^Y{jaa~j6#p^cS{=|5F z9L;fGT2oh((w*y;904ta+v&tbr%hT!8HgH z?tY|e^fII_frl)c-sWR&U!CFG8eH|58pHXE9blfY=I^~BdDuZp35jnG%%mjz4GzWR zYwo8BJjole3+2`k`Vd3V77^?VK0g-V=ioAJYM@5fhoFFZ(i|r!+4NgC(;t&x*>+FQ zN*+r>gr*tNpL#>KL@_ufg%u^5JB+@#)8va-PJ8Mf{xqKvEcQO_9Bo()+1HwRYS8i# z=IsLpCpor_hmI{uDk##f*?h?X$vaiIIhzcJMCFm@Q$P&+^TGe z>kN7}DGtI8VzzGd`wTfC1^ry)s@EREUkyDf4D+man;f!~3Gxp7gO0($ee&Un5W@$T z_k2ta87XfM^}CKSIGSdG!AAUOwlWxKaOr!{*+s#lXJ%vS{oueFjYv_JPLc$-)mE;$!fY5nH6&nh$h zkp~Zcn?IqIh)&{d!x&r$qOwWd$PtDg9;^gGk%y44_lcCaOdfV*6Uf5kkg5PXu?7r% zrNN+;;ky@|I|PiKhayqH3}m9Rvpvpj#VCpQ=mY<^CpMpk9|B+GH72a92bNZKC)cNO zQir0u{4DaA3Jke)q`Fe(@e&Vx%$CVdP6b{XG5nT@^ocfr$eAV?V-6#S0RG)8G-*V- zSCn^ZsU0-Hfb<=K!pZq!`&fuh;>o6;iK$c5?j#{^3ixL7a_wm=6U&jPld zf-A5Fo5+$YD4C^zglzDE;J5VqEppfl0oEjmZq^c~KWJYhg;~FXbn+)+OUOoPBy)L z6x`yzG2H6kVZb^KiGF<==KySV74~9rxn0_v#%4JDf@Fv}f%><{!ZzJQ~6^yKpZ2r}4 zK*b|sh=p|WA;yEwU_R2HD!n9OM<5uotK}N(d-S2nEhM*eqW!21IwXB8D+F~7-sMm8 z12#zmzhrrv$U@h>qF9i)Ue|FRqP&cFRgy1|%jOW-2P+oTf;4cF`XMwi3B^KL;(oP7 zYv|hsNdyEkPhg{l3+3 zS6&=R86zXxX{N@=HMWu^>+?fHMZUt^6C_`%l~D#;2*$Jp$q$fT`(>>9K587ou}RZc zPH9qZS=QOKMwnhG zdk0S41#?7lOb`GqXSes9|MpUE4&8mb{h{X*h4zejpxdSNVZgS3QBR3l8p^>PNKHt! z6k;{NlodTiOW}AyLukR!R@+~6syPoKFZ|L3kvk%^nb!jp360n%E-#=6x_4Nt)WL<0 z0sVSVFL;3$3C<~eD&VK)(7ItwN<2TyO{HW;CIfOS7h+wYl@2?+B5?~CwJ%(rsqpF@ z?Yt43@l}2~MLxCK^JsnH%kvmr$PoU%FPk(J#>SDH+wu1gz}uYLMrLV93%uP!MOvvy zq4dE`T?NmI491acBO;L8sy$AVA%)B9FW|HHURGk_4xurGy6}=Xq#wXW?xay+F5s8y zqsd9|qB&bQQoF<*Mh%yW@71=$NtRwvzz>CudTP^N`ow*^Zu~Q^e3Wg)_D*2pM=RZ% zBRKp4GGDhOEhq?wlr4!*B2Vtn;qSl>%6h!fflqp%*k1t_wjrj$%o`_zkPmBL zlB?|umn#(lVT35jJ^|TLRiGnh=>QsnPXwB40hZfDpjzy%LSxu^q8?pON~ZC(&pA`_ z>GYi$sjQ)BwYSZP!Q6oY=y+9r+9M+EPb2}j4Xvd`eLfs#GwTugZ)rf+s& z2fcTLZ$~BcJf8aZN)eOCBRC0uI^-k3=RflZP0(;Km~_xqq~RX4tjN|fl{KE+B|2~> zMP%#;6O||6xh!Y&ed3jP4ku-7eUm6;w({e>V1#(Udol0zU%Ek#A;>6Uj+oikzOCgv zu*|||%rkeG{JhhV_tUxb9$yy)Sx`h$vr3yNWTH$=O;_vDK@U$X6K4=l8x%jwQpg_}NqI zNF9DQ1+&JBHamqH9C1M4hf-`3@56N10>gsC;`upyWd$5k58Yql3q50;_!ruCr4Ty5 zjB~dW{P{GqNHsgARf(dmp}6nxsV$6FZb%RvE28#av=U)z!q8_(-2Lxe$I3hng6f7< z3Ix|$p3aZ${YJeedbz8%epm43V9avEYsDaPf5kL-d&+Qlx(?>nIs>ArzHt%xfa{#I z#6#$CxmmUQSvW<~gh$Rh1%GOjh+nar8YXLb6Kspem-x}0rIMnuk}J9acjo~AOx_PO zz*8qz(PneBBFmWnS^|lBFMmdC83ont?wii-`)3?0JAl={1frjW*A z7Iw1SLhASM@D{Pi-LVabkI-j~A300ix;UwYI;4VBpAf=5o*`WWPFK$FtzBgOH0#vL zxrs4Jo%AhlpZ;=1s!krqUDE4picrplfva28MCElNNAQe3zUC{7@=m=7I{lxrv1qU0 z=wRkUi~!X^2Mw+8(yRL=VLDnE+CkdsE%yoG@7-aD5OAUzN+O*3d{DtY&CQ_b&*<^1 z9@Ec*Ikk6egy%jPGUV#E6ek7lX$Dy<-{}pyKD%_SbCgo5n)MNUI=&^!GM-ZX6<1dF z-xXkT#KR&Lxt1Eks2Mn0Lhmw0+?7yEAg)%RaXFQc`)%@5UOVVvc1v+81CqvZKZpzQ zAkKaJD`fR-Ep$CGX~s>h;B+mZT({{{^U&D|p5Fwy=Y&T{wOAb|M50m1A?I_x`MCXX zRjrvz_*rJ2NaMJTRvQO+d%gk^9p zo?KEG6r5#Sz9f3S;J_R6iQMJFIJDDGmpsr zp8kJtggjmII5ltm8*WD(%;)?!8=7NjBKcL#ktJ7jkGqu8a6PUyM1GA0@I>!;E&7m`BFqRHS)baYe41&rm}xMVaR3RsHu_r&)Xz&s`)NqHTVtg zWHCm3`8>gbdV?jPc-q7tgrFf!f~ANkKNY_`+$<%GJ3EhA-dsADN_rXZICEuJJ(J?cy%e(z{ga?j$wmdO`JJ3-3)2 zQe*LZr^j~Os4u|CwOiozl>iIiQrNP9HQ1dlbqBIVqF!Cv`iZx{tW*jEgd3YwTtDJs zy)x-v8*>7WTuZd`VyC`7h)tk5k&X(Koal>bw{^;{yUl>P9y`@J_1On?bGcmo65X8h zNOk(EU*Xdb%rdTQ#~n`*Dhl3B{d@?yTTPs&BG|$=v}Ase{2EF4>JdcS6jeT}fmMQh z1C|3CQdbA<$aPaNH_d0SZXG>6Jh-#MWPbg2t)IU9>YYq=tdR!rjPn1Ugxb{l7-Z?p z1v%9eez+aS71~g)vpx3R7!aec3R+s5hq~cdkdK18n47|~Ah5ki@%9f}dg%)gaJ~Z( zW8ckeS73$s`3r!>JFl)zCtq!ThvpJW{bh8F(w5j--JjQMI*D!&wS2eh6bEMePz!ud zF>;S_(%>a}g8EvHZ;#01P+vPn48I0+Mz)8q6Jrdy^~BJMPQ_GxNIy^ zQxEseqh&)@Wd~ZUmv`^|sNXp3x~uNbj+z_+Y)WgL(U242)q zyzs>kQ*Zl&z2asu4j!|y&;gM)VH!3DNFFvO+A$B7hr1%F zS)wG*^jJUrd;|g><45jAvh@3d$tD+sym7tbicLQwauTWF?tm{&fasI;6kQ~#{%vn8&O>3{f{pJegCeyz)#{5bganGhJ`OHC- zkyV3YGv>jt^`1ZJtf?(;Lf%tdWN7Dm z2XtQ>G@m6o&brK2oA5Z_Y|Lz$>e9t*mN-^OQi6m1RYnR(FbAs(DMBM*sP|sOFw7W3 zlNWxvzp&8pp44=(xn0O&tS5mi@2qV%YC;zTgjOE9yFJDE6CL^c?FuT zmLkDQnxi9_;|*9k(b9kXJ(%M8y!MVSX(WHT=2`Wm2tM4@B!<6Swp6&e2a4>{@x=u3 zqwWBYW7NSEqXCC<35Y%CE-OI7;@cfaapA;iFAo-4ZEOsl8cgVD=!LvNBTm8~JCQwn}ot(2Ee>dkhC{i9`u zN}yn66u{+x%gHM^-SZ?zAYHwkq@lV0z~Xx5b6H@w&gw96%&84LmG%xoDJkj!Fq{o| zZbt|0g!TFAI}+6BK0dENoqq%n*WOP;4n|cd!`T1e$(oOU*|y$!+jR22%Ax6hx8;+# zE>Tp43&89++ccjaA8oXcq@tpvE)*&FOnBPg61dH!=|nnVe4v+P5BOeGv1>KqNZGR( zgNvw^#^4|Vx9Gj;t>56-AaT%w_C$EVBOki$0v40IW_DPnIB#s$zUp{YlNw%R=%&IJ zHDgimL`+?Tos`^vVJy}Zyg0un^`fH7!7Yo9JtjjQIiuXJj!*z?sTma^QQ9ZKVo?Ho z+WQ-=AO=Nxarm9>xcA96wKEd;BRbvdh1PL2D)4m!6}Q>GuMaByM=j~^L<@fmp$mSzP@!f_ z{qS2QKmeni2Ng*GRgFdw&)@jGDn$S;>QAFO@^(T` zV$Pr`;Kcg8;zMf7wz2p#>8{zePY-?GX~wUoAyMqquI3=ABn2MFb^=8r@^|#_zxU z(#x$5A!BwJ95s_dX^^Mj^e%Bd@QpK>X11cNte|xs$b+I^G+iAj10&*(L2NIO%#|g8hy|0)QvY)E`fTF`XanK1zPv#~ zT`*YbrTq0w0?)18ezB){w;_oT>ja1vqIud0!jU^Bi9oW@gL1777KKTSwQg`Tq{5l} zD}R=n%aDrc0XDGPeLD;~U{5}QoCYNJ)toPZu{HC)++hr@9~A15tAJ5?&uy*$8?RXb z=si=~Sqp{Vhm3+IU4PYY-x04Ih#VDL7fkV7Lpc>UMPv!eG1On1YJCP7Ltt=-C&L06 z2qD%*L2VHOZjX(A3Lz-D{;pmH2;9asR+m%oq$qtq6Kj5)WpKm=K#PWp=i8hm{~|jP z4iZL36@Lp*&M<|}6*N}Ki;b_C0tcmfVa!1MVA;Tp46VV!^=>K9$Wfz5BL+~Vu`Ip| zM2I5VyCmoc>^evZ$%FQmCg5?!Rv8@#I7)rqfe?uvO-v2;qqe1djMj6*>)mRQqi^w{ zeIKyXKY0R+BsVm+;4Y{PggV2i34)@k(Lk7g4_pdfz*9lzhgGA_X5bT>)zH&kB55K=hcDFx{8=j^zN9U zT2nm zI_}vzS+#$|_)ZvK-$z%OU`t&xMYt9D#si^hNETC2QYS*~o1yB2cPD6;k~ab8 zZIc}inqI=k`+(O;i-ZPa3!_^!K*P^A*v)+lb29G)d^0`4IA9KbkB(Q!WHB&SY5z0U zhbU6{y64Gw^e(8wEF|;1oS8y6CSUXMZZYSU~VsS83!V4}0IevGn zeTaw_YVk25(BUsR8W7$INE{L+2))<5PtGRBJUF5}gjjVLkAd(5FCEE=xr#rigd3!Q zCvsp5{{g)6Rbp`(9s&Zv!|jg^8Au^bZ}-2dN4OKjG5)YnAS0<$8bdu zP)4ywyOnz32D-wE^ zAuo3!k^-EtzF?|y=YdNa*PGy*Igd$Ww&S}1x5WRsRFv0gp$*M}eB`$jqn46Kp;DF- zjUB(RqHTqTFw?$fSS_%Q5IWfUs1tp|SE(!W9_5ZdzP#wWKH8i;W*Tqhc@-2n+W6u% z3i(q(<;rA;g02Qs>7crf^BFg(W{md${kuhsGvC^Nq{t?t876DFe4i;STtt*$*M>`b zv48Vbmf^dqYTE=R88K>p9cZskE(BwF^ByKTR3q3JO#?OEfJL3Gw8KDuF)Uc->M_~h zqlEiuC%CI=zm6txsKvMV6rNoWa<3d-mMAC8S%%J9|D&ky-@ADRmiKqd}9^AT^quxl{IgU#m-Q3${2JFMBe zL{Ot(HZ9=8pTyxxdSjahN=4B|{v@^CDf=S!@Rq>Gxj&-F6C8>MHY@FU*QKEt9F&>% zA8R!k)4pWmB3^nF*xNn}s?3Ja`gwZ55NTltF?=^Q_MX$4oL3d^yAr9)WbOEt{1cX9 zV3B{J<$qgOtfpQ^UxUkoAt z;AQuug$zMB#Q>1cPn$~ktZEEZ_U%26Q&(yX%BmP^~8bSK-$ipI_62#A2mrefz zA?%QuJ(xV#_ah|Rm3V9d?Nu9se*tKE|E}meExxMl)Bl6}=3LyIeM$qN|7pqLBKF*W z8XSIwOFUL~f`<1Y*?V!{yZGrJh!DRjPGhgZeJ=To(TQs~uL!2?YI+VRb*U0QqdWU= zLE@rHLPEkS^A>;Y76`lqE_cgp(`pX?_@5u$bIxt_*w#6IS!U76e1D<*9@c7+zPQtj z(|@qozG|aF^1sh-q>}&in9Mkn-;dQ7`@>@oD)`#}*Z!~aQs?Luh2#>)R8>av=k zN>R+;Y%T&;S)hXQ=!#h;dVM* z7`X)#MlKP?&*hgn@WMtvX<`JFi>#`$5CBKA#vtr;(r`>j=YLOIh@u_)#q1hBD7A?G=Xq^Te;#$aZEslfrvNMunjNH3{f1k*I zpyLG7b<@U*jSa!n8`heI!1_X?I!#nNI}aqo5AB9q;FQUKeo^@!EVhmUhA2?^Mu;AQ zf*@&#v)*QaW)n=*)h;nHX7Kq%j2-_WXbmNEaDs9}o&#h+ z0MI$@rz`V-oNgNXCn&zBfJ6vd1oT{=5Kk6&K=~NGU0grdqAliO-J9d~r+d1W7*^ps!{02}A;yjry+PDfE3q<0mWp=78| zbnu){NzlbsL5a#YJvboD6bgc^n|dffTrCVnM9CtT?}9q-!BXJefRmUjVEz{e3h2)+ z%~AmuwUh$n<5$-;Uj;ly&t>Ua@h9j63?`pUZIW%R?^vI;TzgnA)hw{;Sd+9)&-6dt zk=w|(_E-O9`*YV*Fyl<4uK>2hbudUPzWPoM%jq-I47>hg%O0|Hi_PbDrYKY}3}*br za791Sg0H{3Y?UyJB0S$hRScpO|W*Ji;8Ah2lv!(2VT@;7n$bf7- z3pk>bJ5CA2xmbXKHfh%%)OuGFO`W&=fg~$zJ7abu>C;w<$jiYUtksVHK~`~@20wye zEh8&YkDbf!S^`t>IG|{~#9)C(?3x45akE~S9uO=GH;-!v>FJvF1{amD{rOrw=Xj)N zp6BnRHsdyOs;X9(geg= zOU*Ae_CMvp^fX)Pzd-E64CJKzVeD+GpRCHS%!efZr0DBn=;dAKwBWVh#yc6RPd{~4 zC~DBGSCc|TPZKC|W5V!-&_Qt5CUBN^N0*5&0nRK+-(V6;&z`BrH_jU{uacR?+Q5fS z)53T!|M>B!jdH-VXB)pN=zvxe7Owj-giOD6$FYNo5N@qa2#%~9sVE_&tJ@t3dx-&fqKeb z8O9y7=V)NvqMws(OthjQa(1E5Yq8P=xZri)_w1?DsqA8ZqyfQf3SN%Rjap`av zN8_{~v^tt)@8tFz!#~X$iOF2^`q4$m*&c`+|OxR{v0hVn+F+!t^gO; z2S96z=Yh$3KQ386>)b_rcwHPmE@Ixa!^Kyv8$E*eYHOU)&~Cy`uYQEUw<^sylZ|aHDK#y3W^Z z&r}d-Y&9WKreA6J^}a*qbM2CQlg8o`PJ#yNbr0(1MfWf z#@N>^GR{w;YijMp@Ur2FX=15(!BU$iij?&LdMstarMm30RTuyi&|dH&1_#I#H{bYH zR9pcLdPws-i>!+5Q~+i{cIUxQQNO%FAfA)i2M+3Pk>9KK*yf1QBpLT zk_8Vg(|6|Tbxr&a<#bQ4pIMBYl$y7zG;8)Y{aude*tVT~=OR?;VEQ;p62sS5lk>}o z)pnWSS!Pcr#F{-0sHxJ8zW93XodD2roaUhK`* zx%LrA{{HlIpxsGry8x!4%cwjNk!85A$)70EM#9Vm83zNwd;tA+HO+4i2QL%CAwgc4 zcT^3e-Mv6~cVFoF*a05GyBY9#t+rMNwqLPZ+C2Xy7<>&c1>%sTM?*fI2Ms8&)9>u| ztX2e&_cyx$#F_%rVX1u}COa@pK*&C;8eGaFfH3UzN1w-9KlRW>+)o=m z^1i^ULj^%)`R`_?6(8h2EryeGo0M`~7TOBlWu#QU`{ku1|NeH@t?1g;c)@_s;-{t7 zj<)d!TML1bX0l{vd^Dq}(^JtJ49Fu1mvLzT0dTnwWa_U_s(;H*{#mVX9L@Qv@F6Z@ zsrXg-{SE^4WzEL2sMdN`JWw-?s!j$9dmfq4vQbuM` zfiv_YAR{S*lq8n!b9oDX%P=_pe?DT=sdX{eFSiOi^aPFmZg9M)Cb60=4RsTB7LwIR zUecHyExsY_1;p)av)0!?l!ad1_@l|T4z5cXkN0KN0zN$dAOpg0iz zgMB!I(&6h^ZC z=W~mtR;Aqt9_AvT1gGD1u|N!_e~~2j7Fo(JYq&cxeu&9ohP8pRL^uah{2$+>tnI!G zP*MM~L;)iwyED3vUd;0V5VxMuf0#C%V+IWKr~|DI0&Fwy4D9uGafL4jo)|~b&x(zv zg7=mNj;g2yZ4`k=fXehw;0n_XoZ%za@rquj;_Z@gX?LfPUbp>K#r_SBOUE0+lzXhK zNvVTDm0++%ulnO_))$SZI05y#P0p-w#RiFtgCA58$3_%bf5!4c_x@$7vU%A5l{& zS6JblELr@>Q6B#?Z_?*qGV|boVbW*(=eOXcROg#)Y&cUSZ78l#a1;@L+t5U@v4bD) z4g?p@j6Ot}j@^hh41tp9R=!x6FFbCA2u)A`oI8+3dZG34D;Cxd(V`7{!L%g#YYyH6 zs1jiX=h!w_q5S4K02!5}&ae;@b%6mKpQO1QUN3?iw6R+l3nLL^Tlf6IlG!(Uo*M%WUFB*!A3y!0cK>F6U1F=&JG_%7GXQOkMz$hHycWrUf7MN|j8T89QamKO+4~a3XN3P=_+Z@)0ARLS&!?F(UHcsQpe2uJDM6#6WV1A>@j~V%C zpNF^@fWx%0VOSUfXaG%flYJvBwoaGAEocZ)vUEp}BUHqGIEBoqvDvLG8cSDV$}?!c zP9S}A;TYXai2nBB{+-e=_s#MtiTxz}SQcB%I*vi%jJf99%AEy2v9KRL5@x;X&np|T z5|rbw{Dft=y*Tsc_(c&?Hq!^vBjaNiaG^mDW$}5(7BFhZHgTQ64Iu3>-*o^MbWmIIpv1NM$FdyvICb=OJ8d%{@w5b57a18{gz7f@COrrv&aPlkFvFZu=QS?hOO-yB@T6yfX@usCG*a`y-VEAm=Q3$!p#Eo9Mm<{ z#zi`Ui4xPb|9EJRaijT@N&^{jS3hrwRdJ!^nIJ*mbPO?X9C7w0Qn=l zIMh*Y?VaVn|AM=f#2S0Xl9_4tQrt#v)Jz;AG%+iV~1EmO5L3n-Hk6)+jc=Y`W z%dfVsz|*Aea^9b*Mm;3?XH6g9V}|OYncs)ZnsvRM|Hm87hyNxEhH1CS*8*iw88^*=+kQX}^GMqE z2RpfEj3H!HvtD1Z($!`d8YhwHJ#oTc0Td9e9fX8nhpKufarq2R@KgZLD);Mfl&Yo(8hKk&0gvS%&2&F7UEX zDr%}cdp1}$QZaK5fw8O^R13{Tsr&+nc#!37_YYX19PDZ-(V%YdeMG84)8_W~z&!5z z1OQorV5*QYz?TVrZxQxYT~V=ik@+aSeMLrp3mxS<_x~{VmH|<%UH|YfbV;{#NJ&c9 zNJxqUf`UPZw16lG$RHr0G!jxHAt5RyqNH@Ah_s}DB8`;Pe~sSv^FGhF_lrl*ISzYY zv-fqawSJZE1-MxvXukMVsYiX2iS0qnOZ(D&r1U(FHv)s0cqAd&jZb_>Bb82x`=TC` zTTV`Oc*1CZrF*ONJI*;-@Ah59#ny~h?mW9PdnP1p0c_*{2GeveJ>f zwU^~;zJ@s}x;$SR&g5==oJA~Mz(|R?AGjoIzTZtVE+*q)t|@Hs3Slo|LP5`U9VQ#R z`DSG{31T-DrFU?={Ys!z*wh{T$c@F#7o}WBN+aGDQ5M0&YD_pwgjKw^>M9%FouyJk zFLuKWH~7W19xza1^PsIz!1OF;2e~`=)KXSEeZmGKtV5gLx{eMUa$<5OapSPYe67cf z4jPIS9+!aJPx&1o2;s)SA_UO)HgsxU&6ILF;9o%*`nljg-OP6=3ZOB1mYQ;;mL-u2 zU0M3=T4-0&r*78VeF_fbX|mx|oMM2h%V68|Gar=3XUU=b-rd{$u8R*Q>R@?7p*fr8 z51Pt{RRxwvQ?5yR=CPpM<#87(w93I5oyh>;f_Vo57mQJyUHoxMG?TF`@%XvnpZkgN zv4fxfhe{-KS&S7S6WArY?~|{BQkg6+ykRP&t#dTo^!ch490mgr3bqeZ&7K-IUeRpb z8zT9ts8-+SuA}8`8uy|7lqP+A)90%x(Fai+m@UEmnD`~&1`GX_G7dednwXrMBAsOR z8xsSDu%tQS7!D%_9ec7CO4M_uPu?qUhsh+uW{hUJ|8@i?&{uZ=L*hGRzg_Z5!w4=R zuctCJ0RW&l_ujnb?yDP~tqg5Sq3x03!`Sm7J(8C)7413DCl9ZJi&!pDW(q_@~D-{EZ)cPH*C z^x2d(KYFQxdVA}pN!L}VTYn;7psWStS+pfo*JT7WS+1|IIz04+2|Q8xYW;z`^f4kY zLJF_mYoSL7I;G@0%OgfnjAxw%Y|}E10MVWY3Pci4L0pFaSrHr$l0X}tMp?iT`>757 zZErPCM&Ku7l&VUeFts;Xiwp|h^Xad{V^-z3k${bOZ4O-|nsItPnBjWAC~d|0dc=I7GKUqr;vI>`E5P>Hl4x1K z0)NuM>4QbCl-wpj58B?1nnPX>9&0|x2mW{>vYE&=sm^~{1^w+nK<5~?fA`mJ%6KN% zHc%HcD@`AqtQ4-y2h0}Pc+H!R;wZyOrU;bH=I|7&dyq>xo2;uK<||Zx7emRZHCrfc zvj9L$_r&KAErBFQhg!^H)TY7mW=L8{Q?2jmrJs_xrg%k!_YYfA!EKD{5z-0a=EBvQ6v$( z^y17Y8s`yk&7>31DJQM9s%I1o?VrMD!tl_?->yfkwC*1jf4dYyM!{Kaw6hFlP`)za zsSu?w{jpFxKFS=lSg*G-$cJqLDF^?7`f!RZrP+AxKZPmDBvt3l6X<}Fa9ENk<@#IH zns=8u-AjX6sYyJH#r>cBKSwJH%pctbK8JN+<;jG~a|k0Gxpmqi zxEd)~tM40{dhuhTdIWogphxY-BRenFe>-Ao07 ztRED$>zZX^q_LlXsp&ezxBo(!__i714~~Q@iOa`2&i944=J{5qnwB@6Od#m9Dp4!%8%!%-2**j`)}a$-$*xqBuPzA*q0xYl3#t*$ z;$a7_G1jNFd&I*8wxTZM-yFzAK<@f;<7&iuv%Edz@5CUAf%)RVQ;E~&bxAu|88pea zK751vJ8zSI0f{q8Vf>$)Um`tUp>E*OAIReA_Ebx-X&?z_G3HwY-%AC3p=xjQ|Lzjd z7ZOp<=Zd%=@FmqBpPqxSOY}zdCj?X4Z(pnZ*WAe4Ft<@n>glSs-+nxLi%2;%*L+-9 z{-<=JsKF=TwoU4KYTM5eVoM}c4ZEIs-n;YABFttya@(4FZXd0OTbFSA8LHDK}L1=tLh;K5$Oy za}1vv#Wmfk%QMBl>;~808y52mRZLZsZEbR%8Xcl^o#@1=M`d-V z)tRI1^Vr)eS|mrGA@H#yZ5WnY9&^me!#$QNFGbsDe0#*MBNz(T>H{llu7NcArI$R) zYM8hn8~;2)ibc;)u1xnsVp7koHpOeNWr{NQg9O*-zct+YA+x&QySXUfPc>M%x?j!jo_T))I8kB%mRcN`bTRbpgS;Xb64Ku>guagsyb+ROu!65$h7qQ#|4 zKFK>|ij-N_CXNG5eD`Vcj+s3KojvJ$JS@q3>ukm+4~$aQ*c4?$*;PqC%my_-{=C4--|L032Aw%~JZsbJ1kgQ-9JrsBr_}La$ z4t9x1jNzk|P9sbTIQa8z3r>{LjdX|*8Mm>)LxiC3Sqy(dbEV60pKIwisErbThf!ql zJ+{7089;rdot1!g1Yx=S2K4%hzGWSVu0_5gYj%A9*0RZXsUcNSU+hS3SnIdOM27ps zjq!zTNWmeu{g>JchD%<`J!CWZONZOvYGMuFSU|tzWSNl+@(>-XVX4<=hx^9zPJfzJ zaTtYN-1W|#Uw-9d&~w{f0Q$b%IKIij>lD2WTKLx1*v=OG<)-k_!~Mm}W2&D2{i8e? z&gNybzIPIzfdUJmZ!Q83X7VeP=+O0zt+w_%XIT0M8F!f_m^=-C8V=IV+| zryI3=*@dDOd($18e({i^8;<|HEKMN~#J=Hs-pK1<4_K@JHGIKe?k&7w?U9@?eeK-5 z5cbP+vu-XI57x|I;8Ya7DJE+5vTiJn+2eE*2r88ih=1(AL`uF!2}*Bd){Ni-OQOy| z=k!dcZx+dl?KfId&UNETz@L-FZaO#p(iv!E6_nbg8CWDgbMwO;$%9v=$+%DhhqH7C zWty~Dl6~6{_aUSBk{l5H3uH>hO}|-lf>Dbfn>IdHT{p|PFal|E?o9)SK9CJe0HmuU z#)0=Q7?|fs?Z=k*k~%9G^UOkr9+s|=O>$tPC|~WLz~I5=TET5aCQ2@nf}pZ1a;*2m z%1_Kcs;AjpAN@<*ME^QSxG&*csuw^m5$Y8!t%QbH0Q3hNkWa`rXGzu z!?=`j6umE(TwcQ4CM6;1x=xN=r+gG?c*Kl?_Gw)fb0kDBrRT)o#J63Y#kUS zxUq?!FW%yAQjhEj`#fG@!}1pewZ&1Y0!`K_s_@{W;3G$*McO|J?TH(iD}#X$#wO8L ziXvet^W$^zQiO^ftKAnHmru-H4$bvxRN20>q(!DLlRJ}rS0@xIjHFhee9XTP1@NlD zhkI-8L#d{P=#P9$jUAH_#J_-cZ#c;KYIlvCLf-Gw7{9V-9QTndk=xXg&Ko?^rm{hs z>v)!!Yn8X+N1G6DsHuMgNcK!+fs9d95MXDSPA8p*O)5QEj;@2fb(fSt-u3>yV<&wV zhbMw@igU?I(i6t=_LwQQWu*>9f$M4vp-MFHcxv&SCBnY1=>FP0k~O2n4=I_t^JkR zym~k3EG3CFlAE|4Vte||#DBQ}ql=G^OF2gXR&=YycTF}>BU$A8tOG4#kCiN|>|eJE zlofN;pHKc)PMw*H>Tfj8{V8ZCxrm2yfB+MsfKwdGx5@(;A!fwFZ=L~vd6|57t6z=$ z;3M9u>mlW*xoGC{=!_RRb&c#E4^rZM%`YlN54+yMjO z9y@iwLMW!^j9L+u@MTZSo{k3aX8A@^Fa0}kExHCnlu(5hkt~)zDW_)LQG=>AC zr=yDDSy2c{)M4h9>5u&pV@~cDNpb zYi5?mwSLgx>jnwU;=9RGpv= z{@`m_t2)eG3}lAn%P!(iF%TBg*H7QyGkJq52D1>%L*@b80IM( z&7jcoJ{n&ek%H=r0QT4*9<{DD_Td_`XZUy46=`I(NjP23Gs|;;ZlC+dN_bG2RN@s|?rvwHMF0zwD0%W72nK~I=exDd zfski2w}KlHMArs$%FaIkPn#$ABKWa{Se7kI*{|FZgDn1(8Hbu=*q!i>UB5x_!oA_>?p)-x_g?v<$}G0jY;KsP zFK(HqmG1XlnGELb(%D(YDeCfX<4y6=LLU3MU9MdgQXWJObfKhaT>43Z)@_ub8D4Ag zEC~bxt|K==dPVo>8F)-fKmSx_t6i}`N+rA!4T_zBg|#Dtvvis zHX%e$3aOh@_u$_Gc7TJCKwI-8DdiiYSnC(WyxnwdO%5C;%AX`S+(mzX&+jLvf2Tw) zk=8o1mb@{WFY0#PtxWKlsQmAcRDXwLiKRD|4U;Xq(l0O8oMz%p3S+t2oyx$kCB)`u zJf(CbMs;m$CUtP`q(w{mrQP<&-6|`@=7e`YKXMS5nKn@b`eCd;q^F3_{V7+))Lk*& z%2HUZcfQvB3OU#Ai*OEsKI-BDZ30@r56X8Xwx2t{hQ%Z0ntT1;0_;oikmoB$APMoq zrlMa9hi_x*kaZtQ66x9jj$JP-=O8&0fAu!7`x;Lc{W>Vs3+i1g0k9}R6%z`Iu~xL= z^afdsj91I$Q-f7rZ-~699q9QK*ZIi&g3D)#^#Ra7&CwD^37%{ZTc)p{au70KHmD|U z)TjgN6gI9Vx4swAJ?qmJj={g(-hFeBQp9sPjDBMHR~XlaBAOIBj!ox4A@g*s|A2x~ z7boYl&A^k9TKOb5M1X!o&GZL1+p4gl_vNvR_a-V%&A0Kadb1(nAj|&4CIrAYGvPt2 zY(IQ@Zrz>oOBci?stbxVmneHxK#^e|`V13vZa> z2Vf89AQ=3wN=T^~W1^nKj714kI}sYopzzc~xABZn6h$szOYLGs7)wpwL`sL3Kr&i- ze5$Zit`iDV55`+a{rX8p<5*A_{UVBsu!K09GCYEQIPD zM(F;rcDuqN%252}ToDDy=ST@Qb+3+F_A9Swf2Lm<*S}6CPgv-fCBmKG-4;%)hbQdz zt=YY=O1u4S*1Z9>D%0>fbYI2sd_y7OIN3FvbRAc;xZ>`QhN_2EkSjt$Rhf4vN4I>W zlqlH8*tJB*T-mUZ_#k@vJh#RvM3eeOcq5-7B)aewmxUfWl1uQs9wEc4#1|B~(XVHz zcSZ1zzO3-{lbs(-GoH(YAE!u9HV7rvH-1oYNIvm(Q9oHnkK&n>niLv-$txH8{<>yW zDoe(u1WyHf=^_Kk)nHw5_xSB!F{83ucWP#hU)+GFJ)S5_=Uuq`Oz2G5<>8&CTX#0xK#5Ja4cgJT`NR58PgN2dWc;D1)=0lQF6H zAyi~`F!wdC5AbW?4xPDS`zl?QA@z3xF?MJS&qT54no7a<-jvUDOTvpmVtTt%4JtvU z#}AA!4gM*+^pe~ai{mtDhno^<=lslChD6N~U#qQ|oW`nMG{r<&-R_iLS$;j{H^1t? zb@R3XVPV9LhY_*^`QKRky{GCs=EI>`AeUk;s*{}!nl&-c)z8bd4OAE=^D zk%tqdRyagSD3=?0o+|mQ&RHvuBiGLa`udvnZ}&*|?`@;NTv6GykvZ-4Iy0eu*wVR+ z!jp=B+%mYoARi32Bp)8kltN;q<`l`Z&_5sC7U8im%y-wFO! zEr|$UKUb^iPHDz{{j^*BPI3zchCGs1LRssQ%61Z1ba(JQhaB90`GcyysjmLoR0A?|?6oIf@7A2H<6{bK_91nsNwH*kNWzDY)@jBr z`45EoJ)&Hp;i<_a=`t!YZ1mE#bc*;yEh446LqF_t!`3Nw=Tr`#x_YxjmJBk>Sw&* z6%Lekwti|3zNSIdlf$jsr=I)Zied&KAaggh-a#a=HZeWS5;7Uw5>9&EKZ>8^IB9gD zE)VfkJw=!5C49b<86Ma%u}wd!91D%78FS44aW{7y!k{kZ;15xlw`FPDl{f1ssM>W< z%UyI~hqqiOjLaad)O6_Z!MVLO1w{DmJR_<>zwL9O;u$p^e&3`EPU79!``$EAIJM8n-W>gX3%7n}311kF2{)X)K+NlS5DIEbwcdZ8c&1Oczwei8O`5`% zmelY4^8bZ3+cDfGM7wi!F=YYa*AeyPJB?k2Z6?DBka6HAHeScmMm+6Fa3Y!V3zi8L zPxGh@N>>A}`0rI4DztgI1)s)+pZ?N{SG8hua9?0!5Lq>mAbHLRXV`I^ zU<=+q`}bB!iW^@hzUx}D4DrXFi{5XbePK;S9~pC{#LRkUwqiN?y1V4K@C~m{e$VZ* zO80-4{9dV=tm+>vx8w`C025NOeB3pGD@7^N8jn8E8vlmuLrMZ8oIN=(7kMOIFU|Zs zWHu1fS5+F^KVa093LIn16(RF$9-_Ws0g`{gS!Tx;6%9H6ES^n2Di<3l5x=P zin@Qi2ytQL>$^dHM_2nUe{WE<2s&AxWBQ(Z$s%Ip!iQzab`H;Iw!85!qHD&o(i4%t zRMIIuaK;|DoUh{(N<8O3SSBq0;PdO*(GH^6x0E-$U$21!Y@2ws@3TnM8YpKZ>=@y` z$zYRw)ChusA_r3dQg0r~W{YcK1t3mJKYFE8Z_UE?3#J60ek4xpfS2F1=*J+~K8r)? zK)1`c%l$Ucl*@+9z~3+Zu&Rp#d#zGTcwXdt#E54do6U*RkcUO!_Sa2lI>Q4gh60YE zEZY1?%t;b*pfj?Vz4}yqVWq;YixU<2vq*00@>Jhy;nSRztB8+7x)kR zfZ>K!y#Xh3a%G+0A$UzVlw4M6N9K?G{ol88djY>SkO#%kd$-tk5038R`RYFv&bMSX z<2;7F5Ns00lYd}39)TSCDs(U?T|`o*Pe_Xq98 z?1u+WD3a#i-gtD1-I++W zk$i@C*=_y}Cs0n~i6~h+U}!?o@s~m{9qc@i7g_Yq>LcI`X&y!(Xw91X=C<)*WE z^*|-))}T`1gIEx%1A#9@Cc(GqQ;(yFu~$a$2W>ep#c>7*t&_9jCi06mUkS6A2e*!5 zlfv-PYGNvU9z7%0<4;Ap98;1nF;$#R$sqLYn*wl%D9d*TFVD~_ViP0Y@7i?&Dkg8r zkDY!GDg&)fSXd^8Ik5k%!i+I`tG}MFzO4T?;MyZj6_%K*f5FJvWawp-H z52npO0BM;{<~)-JrfKdsShDCr6+RCIoiPL)N!N_MzNrKcc|~wGJke|_3YTKY>1kp% zVNTZtafxm*vEg0gYz{YFyK+eJQKBOoKhB`{!RY1JJ1v|f#9|9EC`^p_T5E4qd&7WJlx6bT;yeDUzb>=1K41(Tiw)5^S$7Q8K4;>cngX?-MW0H{s%ARM{HwciH zjdcBf;BRE}d1p=Oti~U=%9U$bwJiBS?g6RMe+%t$GEI6cFham#0YT45={qZbI+jJZ zR*vFO34#qF001n;;Dsu~i+y$dgza3Oeo9~H{LERh-=ny(tk2S*uE8+^?HNFO?lv?S zG!)pS;Wy$6Uf(4+O5wa@`?h{<;~~UFU7*D@@wPaIVyA4QMu$Me9@sO!;P3Dt?$M!-$e}o|H`16)!)># z$XE*uGq|T79J~rRIoef!NUDx9YiPNaP01#kDx>F%)oG*tQ2~jLsoUpsFGMG3bNIDX zRzL_!MV1g+4eW9p)X9QzDRs=367P{rfp*pXPbYb1d4Mcb_C-KGLSrB?EMWrl`7bUG z8RN{&25RoDVG*t2A7emNopGx1ZhRSfkJy)NXB5P7nLkDW*{qd-($PC0=%`WTD@bAe zVEv*~BB(XMsyWrS7WA`r7E64O8p*>5w@?t@s(~|tz<=DxNXlDsKKPw+Jvb=i9;1YN zFRBrO!S{D$|GTi0LKM;-r3Ra|1+8GeuoSknyN}}XGXge5RgV9tq<5|;w$p8(!b|n zG1$l6z)!mD8)YAHALPb8_XUhYq}e(j=+EYT)NhiE0r%X?V|lzJ2hiv9cW*6syk4hU*J`(O5Ekx1+zi z-*Jb+AzB6DbCH!;P$@bipdM^7H$g9N7%mxrR`Cjm*7r;)*0~A9>2V^%i(~7!%Ihdm zv>Ui>wfhU{#voThKl_4{%*PSY`+6^&GCvN1OI}Bqe5J;Vs+${Oi+M$a;M$2e0JE%l zpf$>Nm8U@N;IT4m z+T7esu zg^A5P<9e4;6H<cl1OKobisuGTAc%xT-KXLwsb4Hz9;uYz<+LGzl(seVdK-B5N|*nwT7(t zC0%^%-8EJIntNFZAwk*kcbI5Cq#6!vDb+2$TWqK6ymfr{_~hg`UFzg`JZa-lNyXma zebD&cyVZSJu@e0>V9#D;6gzA~ouOvvdi25IPuX$u`o-EX#`PC@$;UBu?kkdMhY5i^ zvETvO`DG=l{o3wZwZGMU-hDzOlKvCA#7uIl$>9CGgwY@ zgXg<%&+>&X0H2+QcyV#Z>PT|Awq?nX(*v2(r5~k-VXR%@VWP*DP7hJ5W?%%tc=syN zsnfTJqQK|x)`*)EelaSK-pYe_z0}gYC`Y*mQ%vKHX zehAd2E>6L5dj_Y8&*BpajYX?0A<@d+sq{@QntL4cIBO~C4FWqntx)lbA{qN&u1065 z9U}_}3(H1#!I7{>n)(Nl_SUeHD>c<1AYk~fzm$Iwnp7ix=>C45F<1Ie03a`s9L-jX zM4~w1+QJwz7iVD;^}^jynmf5hIla>TC4!;XCAS@LTRc+)Zg^BS)#lx=a>h5EgdMNA z6#On85%NtA_-oKx2O~_NV7qZM6IXJWTN6owvw@E59OMne!w)}G)h$U9{@eh3@WkaA z@%f0ZViFjx%#ovOtAU)F%3NOou}bEUN3(HT(!00ht^M&B`5iCc19{@cw-ChoBK6Jg z;|4qfpdO^)gtwqO1cbemmx!PT<9&*6$?Bpu1cn=2uWW!EcIl9qGJ)vxa~I{l@qrpZ zYIqDONF>PYHZ5t_=olvQ{)7BAT^I#7{`2Ms0wYBZG58N;;X}~0WEMXninYFDE|aok5h6yn*moLaeVf!|6%Fys5~+uihi)lDJQr*Hqu1^Biw6?#G1 z^P{v4X!+V6_g93kDPUW7@YM=^RJ`cc%f7~m6^Fd+nZ+$d1YI@OZo zQWy?pjXb(43#{E3(3G02;RLa18blB)pF=T(a3bP+0~Ojq@W~=hMXkvO(7+BR0iw%C z2DuHWfAI8(qy3YKI~1=n7N>|@HN=&j>JJfusDx4tOxMB4B*HZJ0wsRP4Nege<^}~d zk^IealTPa%$0T`ht=h;@M>pQnc>jP^@!%Z#|<`gw5c`PS!GLHIV zm%wlNLMS3?k$5T>^sR%wNs#Jln?PWuqC)e}m{ zA=((}_!+#_DB?b=PMsA{!ewFM{9$T=7_`8mDSSr?JKJfK8H+TTRysuj&~-@!der(`v&!zo9NcXz(>c!_XcKlw4mM;Rp0gOxytZX| z+~pKr{v4Sk#|I&F9g^hwsTP8W*5$L0hc}i$rQbn8^6GYDgr%k%nt=q{m?3wsD`9N~ zZ)xTcVLlkI6~=hJK%|lQ4M_zj$Hm?LI3}jbi+t!P^hhp_^h{IbS$u~yie`Uzzsy^) z<&Ccy#wZ$%Q`T?2R2iFouTSFe!XV8c(U+7ozio@D#G_Ma>y1cE$M37~02QBjtQkz- z5uQ2;8nxFQ?KaknjF27Wy&mEEoZ7sb*3HiCl8)WuoP;RND$T0;u!OgQUYZz z>)ssqA98$Ame#1~Uwlw^zqe;A!#n$+w%=B9ZU4q@pnI1c)!Rf=>%v>~6b~1Yhngms zPXx^e`GCJSq*6NwqQWFo9YawYYF2ua$gNmpBag@oxgi#b)NlR;dp6(%oQVxNzS;9!{SEMd>PhY;NgU^SR?z51!9lyl2 zJ%Xs=YDw-e+fAj&)9OV6NKaj>%}}-Fbbcc>spr;--bhnJBty!ufIv*^8oOTTz6O=l zWypDE1F3(?xlN8vOD+XgIroed!!0F+(qiET{QbF8hnKlTi>|Z5|fK(Z-Qw34N{ezh`-Y*^fjVtwYN5@dm!Yt;@!jA?olX`y9XbVQiA0X-Pus zNy)yx!yUc7YT4f(RS%hp8`krqh^aV!@0*K~oG&8vYPpC%r=4u(^AfzBEEY8C)SRgG zyQHTg*~q@f8uK!KpAM{=7g6JlPojBcUeTtHzyDG=k}2cQ?%JAut+aL4g;}jX7j$`x z=)9W+N9Xo>KHGeZY$VuaK0aNnL_-2=*W!(QEAJfa{7F9LsgQWEo-1d*mF{e;dERP| z+WhVf-{9GM!duNCNuO)5b;1i)PuPPMZR7M$ebctK;kUIg3I^Z{_I*jVJf=x%6-eIv zfe1DczIu2^uRyq85UB}jx+(UotU}PcrMT!ICFfw_2l@{FrC-&^PZdRCNFp#KE$qR3 z{OA(BuqSb<8=_3f*9r*9>4SpvltcW8MPt@sSayH^baWEY{vMRb0j;9PXC7s`-MS>KxDycaIE`2e(mMNvQAS~(b&H-Zi z6_Ev_XB}Q$6IKLz7S*~|F@$nVc{h57g^lwfcEb$fGmMthNh!9ZQ`NLa$rI;CQ*>9I zd8kteUIhFO53JLHN2_OU(`<%6a(>EOJxlErVMF_x8{@up85?b0<1EeL171_duNnhh z3y+P+aKq&k?gtH92am&okkMKyf^&2C#2Q5Q@#%w5Ho<&dSs`?V(*&Zvu4SFKw!1WT zj+KnaO(L&LU*8JnLznrT9ejsdA5Beorc9<4F2>(nTIau~yJAc8Aipl}?gPc_@7i1+ zkfserVj|bilbYBZgoZK+3gesIOyX1BQ=t_Y$bq4jpE12)rv{dq-`FQP!jYSUF&X;@ zGi)yORE=U|dKSg|b}?3`gAv(!_pmHNjU&D>oVIYftW{v~X*bna_}uZ@$L;4aKHY99 z*S;j5B;9@a-|q~M)MTP8CXrVcJ4;Y;iZ+};HSKr!*?r+90zD){I1@h+6@vwqn)iET zi_$tre{t?%3?YBvIWEsxm9Dq=$S{idw#4GZLE@GOf~bLcAl-#Ve*kl*EaQ!9R!Ub` z^Oq1tywAycg2@cA58D|-NRGdLjq;!x3*9gxA9Ujve|x~0^nW)XZuEyrcmZ`&8V6(V zjn9hjYWqF(VOyx&WDt+_su>xCWa848k+bH)7+#)LPtuAWSBtOv?OfiUFiw4Ii?r+WyEo@Q+k}B%GPor zYI9ksid^R!3X6Nj{kEoWWd$*s*3_9Wx(tZkGfDp$NJEWBLmkX4z!e;2ja@P%(ByI; zHAt@=XFL+OCZliq4Q@5z-$u0WEc9GmhRnMhKO&~XYzj%_^s#VXk5IwO}&a3t7~ z{n^lnKD#7cvv=0aC*kXvT<82Qhq8m!6qHu0?%HE`_fQ6Oaj4dFM%!m#6VD~c5kc?D zb~rHzw_ZeQ&Eg2gTeN4clzdOq2oCqXjJz;LbBoUy3JFslhTQC!ywTB&OS@yW3PBhA zyq>;vta$deLi>HDEbfN!Xd=v!w1uhjMU$E+0-MyvWpBB}tz1uO^H&#SUs{OlcnHw& zUU8ihSxIxu+Gyv|G${_hMad%70#-K)Q;hFt&7A{G>4MP&0=2v8ugvaB=%F^ut61l! zD}!-w1fMm0IA4ILk26mkJxRcJ>w|Pv`+G?R#^rYyT_(=m?M|cSSy=y|% zr;()VDZa-%oTfF)n}b7(VJ^92%W8F(SUCDQ)wAMq$j!Ns`|umVKj(7>HMvq7H*6hU zm%gh#G2P4g@?l%d>b+M%bl;1y3k0wTnp{MrI#?U(7awS0b^MUy@7UjQ;~#s78QE#y z8kT&+Aey@~K3xwc?+of-bH0f6CX8kI3Ee%hqR#ZZ#fJMDbYckj(48B_zcRGA;%R>D zw+JYC@+>91`2F)TE8Va7LeZ4`DuXoR{j}qwU2~*`z+(~IW#C)|)IJ0i{?xm>5@o(K@WuJ%eIXrRHu#B{`rk_*cC-m_) zX2g-sJhwEC;JjtEd$^7hkban>Yps$uV(W?0Qs38z1Q|+ zXd%A#^S2x0|JlkUU@N1*H@yd+E_n$Ha&#WP`w0I=S6AIZkuL_1(;hsY{ z7I|@16-=2bf#x^fm#K8tv^Yp0pL`LE-64VAdOR}R@MAzNGF^1uRc^dl79PeBD_) zP%^wEAVC}zi$(}gEbB+Px{6fcsw3~lO(Tv5JUuIQ0j@^_aoLFyV~vw*U$ zpDz2V;M~E^N(CU7U-lD2LE`!uD(6`>wapIR9pLMS2V}bW-V%r;lRJj7VJYCM&E*cg z&bSD}te=};E41g|6O9_g8Lq3X0$f}IXn@!$zKgC0fSq3l^;w{@GDbQM6+ba?=gY+* z@Q>oWx;9z2GE!zcseQX<^}{dMjyUc~4=f--!?E2V3SaVcZQElxsq`nr-hZdPG$UY` zaSKu`cZU6Z7hz!LVcYhk40?S?#y)a|hI=vDLv_w}X|k>cxC!0w{P$+U-5|T6Ui{5R zS|WhYsz47g4*377E-ghh_xHVEf~Y*w&*^GG_22v!GwW}{-*SV#)Ee0;DHM4h(mW6E zz@)|jm+)JF`}~Tl7JjX9;-yE%;a$f2-;Fl{<^OJQhyqJvJ{3U*g@1?LWT4{Hp_kI z2e*IOd#e-PRD8-mo0@u+K7gyPC2$8dbQQ=lNi6%pQ5;GqfPX|Ft@?~8 zNh=_*H?VttFlW!YQ^wx@Owp%UpRry28MR=QcFX2MuM0Sy-|e(I!Js5LS{_P%fhTZ5 zjhEu%D>$lTVKksY4O58~A(PvW4X|-^E%8TjFTe$CG=X1V6s~9o;QTR&E1PWx&${{LHp&khx3R} zLBVD8LCEfMAn!fs`bxoXV3gtNjhW5Ra{#?h%u^NxEEN@$K%G&45uh;IRmaP`kA08@ zeJaC0=OP=}UT(FTrXQPYANX}6RfvUOW6sRYegDAo>ZYY&?$eqrukI|~i&6s5FEil^ zicnpyJXGdn&2Wst37>V$pNT?ahi><@rJ8*ow_N?wOO;kumkhYQVj z1yK9l6?ow70qHFRxl1T1{ai=lnb^l1KgnpD%aC*A8GK$5)dBEOK5*SRKMRTuYayto z&jqt8-nqq0x4<*sllBESW&{P&c3J!d?vRXOFr-kPSAQrU4xtYIHd^C#{_mu@xN|Qh z6vL=%-=EJZ&xGVF*zN_~s6->fdbpE1*eQ#0ztiq(=Imr0P)HHgI|kIA(oayPm5`ws zd4Ckzn^`yfz|32I$eN&QPo8?duxGyJQ`_1zSnz%KBg-s@Ru9FKSU#q)?5azL1jcy= z)+sj<2Bfn5e|WzzC^~zo{!X%_?P2fu#Y0UWqQCq20YO3r$_El+^_!&Z{}fy~ocLKl|iNpAZ@v`LO5Ad`J9-m=Vx=%9l`F{>Ho<3R4q&i~E`_%$*X zl|U<)@D#_*{*nf6g!^Bo9Wpe-zlWYcKF1HB8GDs2fgt@&0@LSTNmcR8gbV1;YDJWC zyE+@zY4}SHeVnVsRCw|6HA`?EMetjv3O&Q}B z%V_?bt9>LQ=LAFH^`~ppg0$QVil0--iSQSgxLsh{dWSp5F2+k?A$CIr89Lp#1X7>P zKn2w&y4W9yNBepjSU}z8?Sp_Gwq35oSRWz!mmovV6iutKAj@5t!L-?aQDAr)j1S)HUInUzp<0FFg$4-ellhEVhmMd%VL5ulaXa;O&G{ zz+@NDPgH-lu0Qqjl-y#TQJ!G%V0gmY2n$EgN8|1_yUl1)D%W8u7ge0 z2#e$5Ai>i@G!oDNR#r16|c~?n{Tqkui5#hxdP0|IeclPe&qzcI-P_ ze%rP4DrvH1m%ey~e%DhqkANPJEOJZzn1B$H(tE*)@Cc5%57cN`reg(k&+v!27^OoZk(|3fK5<(X;-!)H zXyu%vHhL}@6zT1FOdrI5<}C1c($Bz~eZZ?5L2dtJ_=b1%T6@6R1sU#hb=}Y|JoN^2QzzyD_9RP z90yYmXCo<{0WC$Se0k!0N#b+}jnCAdu_wzj(#xRj_ZigDg1!XTnkYiW`uRD1_c(e; zLx{sW%`D6i`k#OQ@jw5IHS|UNJ->2{{#DKYYE<}@4+Xs#fv?dyzP}gzxB2_q^M>l8 zkWVGvJifsHw<`WyBI3U3I0UYMpVs;u_kVr>+?4Lkj6pdfoDNasr&!`CUmc0L3nrzY;LOH(7%SW z)20lB4bl~~o$)&-;2u2&eSjmJh3|m$Gz1nPc3`;~8!yQ&f4ea&U{m)1*k3qV<`oEk z@(20QGQ`-#R-QnnUn7oMHP;>+8}AvqWCuw$I781dsI*Vs>Vhkw>qNCD)IO}ZdIt{3 zUNA4Be>rEg_rhsZ-Iz0PL`b!Zh>_=}hpCp9F?c~~3(@Db1eh>V~5 zChVajI6XP8%1D!TyatLuH>j+ufGy+?EZr3-)T$uxU=&w=!#}f#RwnDlSL=3xb_<`f zL$K77CcOj1{Sl?L+POKLX0Pt|fPPl$d`Eoq{hei}H+4JD@`1X>;lC^inL2X~s^r5r z?`7h~4mwT9R~}RKV$lpDSI!ZvL5pN9G{K*8W43bwV&dGuN&i^U=HAiv@HhZ~%UvRG zDuKB=4vJJKl0HBNY$HAaB3a96=LoLkeKI@(INJ>9x;*+Qf?tAWCKdIu0aCwY0rFx~ z_ALdtud|i(BZVQB8B*%&4w-Q+2b@y}vS6B10hg%wu3N082L!dN?uSHMh2B3m5O8w zh3rw5#=dVU+H9#*$`Z0gWJ|@N1rcRWWX+nq&!^ux=a+u(`|o>Q?|UwPbai6Pe81o4 zb3gZeKlh{i;+oo0MQw$RnE80aJkt!=`{qqV_!R~+6ziEW?k^-C z2j1F7(nals9VInZ#44!v)!f3jFArczMfof6=|77OI$^+1JB5c6_-5S4*f+v(@OrU| z(nDE&MC9bL!M<^8vEdU&5QxDM?*XBZX%Q!2lyC&pu0l}VjA1U5 zxGxnPy8N}Dp*Hxg!%hMik?RX~ChWx-oiI`M7#T0ulr{GQ2(!I8X>|~Zcp}t5*rE4H zD0Ud!q~E4UY&YXN3DscSnKFv?vNBn!N^{YC1&kE$W>Qb(m$&_Rmm?}Z8;k_O%-lq{$Yj5mJlTdnQCgIA%`vD6ZSIuX zBJ(@hCm20NB!#j)cZKB873QAaXQ-*UFqG&lO!pRbPcwy#EY;+34A1vZC!~0LP_DA{ zKC0Wl!L6r8aHp^X<4`)4nNO!nNan{Twhzkn~on! zl4A)ca|TUpIhnb85|ms|>Q7>#Mu!qDPXa8~!*MDas`pJaDvwQDLNB@FKOY*LX;tC| zF0vsvUN}DK=_yAV?$&YAMiL*6%aFn_}Pr4g) zQVS+z!)%y>Uew?;5T#Fl0BAgR^-Z|%mTlD34eYLW(eLG2;8^lmO#e;nRf?s%vhI6Y z6bBg_3Ke1DzQIYHI5s?Q*8q1jctl6@YB7R>*NV+@falf-gvERxG1xl#-lUG2#p!*P zYDsJQB8(!Pn>9q~f~(upogGaM0B(K&Tm4G*%63Q2>%VmGT}(RqmR`ZN;v^N8!{}^f zq)$f`@B75ZbwYMEufhqtDOMhGLlZLHRvgj3&B4wqP036u4M>vF4oSJv`wWAZ$fxvF z;XmR25cQVW(I}YyQ3*P~&^4^`8GEm^yq1cds27Q$>ZKOtByh@QN8EZ5C)pQkSO_(_ z3kD7P!gvg8y|z2X--6nZUMDE3`;i=_>4DLx`Fo|%v?Sr$}!-R3j9e4vZhmVNhhY{c4lj?N-xn}_T4!pt%$1VxFf9LD`@ z%$B~~tb6}7mb&p~jk84~m#B4TpX=U+Hseq1^Ta5kq=og9j|b;4O&aOhpddfh8>3xJ zzeve#f)HkJYqdZT8M9EiJ-msjvii8qFPk*NlWyiabnN6)#|AE~jyPRbo`#|S?FE3+ zUoJ;rBlOxd^xI~f&e)%Yr?J9rWHwRyNF9cYw1;L%#b~+Eju1@1v~rmfB}< z63wPO9EZi2qcBaBTN$vn18w|)sKk88LO2T6o-F8_W`IuafMS>hv{50@I<`Xhf#H0Y zXRTKX8|H34X@W;`1OngkAqL}R9eoA&j+iMY=I1QLiZy(y9 z%0Ic9UFD;~5<0Dzye}d#y)1zUGDR)^z0b{C6^4y&G0eD@C;Qz ziMYJe;pi8blt6w7lz$Ghm^s@X7FD5mx^2cesUN`kT?ke!n9R8Zk>G9J9VPDfBlLwC zXz^PXfX+e)X540!8T$l0Y_P)F1BmB#a7ZGe*8B`&T>&af-G?Paw{$MnHpN&rE2>4K zS&I%d3$(o0y*z^OB><)lKq7W%lLQb`O(;#rrty02UwF*%$+>TNz+qR-ikM!57=Vg3 zF@V1a~)&YqDFZ(&uEQaw{kCEIrh z+EEj+T471ygWE+t%`0-&Dm=aCYdOiEG=o|Ar0ju;-5xjsf`UwjEiAs0BK^a)>Q^f6 z{EKPT4h(DAQzX2r-Wtr~qfKhd`S79EFx>*^jy9eqrhvPDK9^%dir(t&f!*ucn7XD6cld_|XUgXX%CIcu342-;TA3OXOo&(NxxSOrs9hXK z9&T)ncHlVl8@$O9<{E4tF4`MMAy~r-qM33`Hi3#DNp&c! z!MnW|wQg!_`U5hKMKTl8)TqOdWy#EydErkO_fQ(S&JhbsH;w|u-lVLryLRIja9-KM zqK7<~i70wnVB1_Z=&L10hxw`km^?i1M-#Pcht$!LXyvcttxBrQZK&yldut+rmHB<1 zb;Q(#)lu@2z^22bk}NW{h#z!Hi|eZVBTYkWBOkPG1V+8!b1nynjq&Xw9jCnkZ0)lS zjDemvri8UbRvlQ0iHnG_B&Mygl%SCMeQI}^?s8hYcP3pNz+zdN31Ml@kM`^E_n*no zVNB+Q>ds;6=7HDQ&G~6tR5n9Plvbb!HSmSR;fE$&3UxRV9fw*od$8P@KNAa67X%#1 zF*(ISYo~e8;$tt3V=k%soKuF4&WSw)Pl)my|6v@7F@Vn@gY`HhADSe@Zj6^9Mdc}1 z$KorXmpi|e*EUDJoc!AI8Z0iDJBH)N1gc1xm3J;{j%YA*Kj$6Xe@@F%x$)W;hAm<> z^)~5DwOyW$N{B8t`VBKAV&gAS2)goku6Pm02Iy-yiqUCoE}E=3?tUBG8X`49H`V{} zCstl^%hrWG+(~DtNBAXy`lgTQJwKibP2tPIIblx59+y2z!u=k!!>Qjszpm38<>12( z!?TaZCjAN^SJ;z3gc6r&z^i>TBhJg5_S&C`)gD&MNKqrB-b8zuddK`B`xW)$UFW_? zb|?0l!DwAFB~C7G-NdC}iu1#SC4&}P`=-0cteCk6qn(<`&V6Oj-ke4xW0@rW5#+`+ zAJAyLF>!AN2SYKz(+**3TzWe%Y&Q$h;>1L7(z~>o%>ZKHpMAd$`TcLQU(~R0FCxFY z>R6bP%1f?vrY1%StTiz)F*dU3sSYcTU9Ax`!la92Hb85| zniwq7bcPlQW*omh7hvWQjpBU8Ab!v+KcRW07fS!=I(w5JrZrUO4P^<|hX*Ez=ba)x zCs?B0B^uvtPwo8`u<0aG;*Zs;5X`m(hD?VU%CD?t9E@sh3Wi6!HL_TPL>ARe7-te# z`9T4DCIsx6py~F@%y;@Cneb$tl}|cY0lWd=g^gXctN!(c;m;FP_}Tg&-HlhYhyQxb zOLGse?2%ZB6gl+a2Tfp)hk(Alf{y#Y^a7ylAYKR!*??5#lmVw@0eJn4OFc%V_?@kM zq>kiOBZU+icBSlP==D3;6gp3!;~)5UH0o*(71q4tJ3RWAA*5#!4&bOU>CzE7Wv7o9 zz;aykxIM=_;9!<h>&C}WT^>2iTQshBqeav^-1bcr66RXjj zDreFVj=N559-TX*@p@9sTj+?OMVJe%^8+n1a{Q6=Ur?{Awr)@TJJYit`tKfup+N*k zFK<1#wL7J^dF*c%ZvoMwP_3UtfpH2}5ZV95xV;C$QHAXb=<9cE>}LTMo+r+L>Te95 z0S7=khcfAe^Sx0DBd|uh5WfAxMN*gcM@3&@2GLGqdY~y~%&^`Ah}p-G<^(RXy2zmt zZ`j<`%zxPX$Lf~czi9TKl`Wo+0nFXx8=Xa;?qsZ<$Y}Ak>CVK#@2zwG)*+KcAbuW! z&~^vd$%Ry|4v-vo=c*HF0p!v9P>d2649>l)@bgOyhJK{PxbNkaDHR4w_aJOKN(b5r zn-}aUyp2x^tc|&(k^nzQmt0~2*Em( z){EH8zd1C(gf@>(zH*swFF2jZwrSTJJJI%5J%En$?XKwPx-x5E64mP; zb8&G}aEAU?jFaUJbYhY~LaVll!2_6_;r6w&{nn~08bsUrXJDxqg*nl}`wvSR289t* z-K}W_1IK41j}^KxSl-&W4N?enhz6WYU`?Z(t7>9cG#WwNo&#J*1pTVE zF1{#DdwpB7VLn8^hm2|%`YRI5I4_Me?JVP9SRCiQbvB_H*0#Rx7xm-}`}8cNLnKC} zr}rjz%(VDDURw1AY3uNMT~CH?it3ZN1cNFCuwPH{J~}}N0kq`#K7pXQ#%$ec1Gqt3 z$fY71MRxdqfYS9WnCic_0v?>Xozq+#2^iZ>^cz9q?UJ0~+3+I8{pW^8?zD|&15FbDf z9eL+)WmF!;P-1}HU||m?h?0xnR5ku|HT&G~1y!fe-+b(Uz*p}BDRDU2*05S*a-5~}+W1?D3|l98ysWJz4jVbfnF`54yOe8rZL=hB zYF##d3+r**b^@ZoAm0`M5MI>S94AWWG)*ja9efYqjJWA=N}5Vb@EBg`LyDX2{DptjcZ&b>sMEKA=1aG>i}O2uuz(jWKggB^WzI z7pqlbYIdkAZaw@EZEwVw%J-qsqPD^g3>KC<2H zo5Gh~6j_%=%Xa&h*ECnH{= z`U?c*mzM}G#qVvdV&%@36=$8meG??t>dp(gUO<^5Q;kzFYg4W(^m(0i2u)xzIP=e> zia`0%&f?d4tFHW{&Js5_#w^8=NRjhqS%x&t;UU&E^Ub{77csdS1h!QyS@m@$lDtmB zTP%6X>6jW{u)O7gkTmDVDM{oUF2UjpGi!M?g|@hhtbiv5c$&+LByF!+PWubC31E}E zrT&|gNx^Dx#!_yGh%j2zIDWam;YMc+alrq9?vON~CCjJ#AK#cj`Mmw_j{KibcRB6? zK64pIFa4dG6GGIS>jU@O|NPT`{QMuF>l>PL$0 z!)}GaMUa##j(|{rbVgVXzXr)7=p=hs)DylO`lZKX86s5eA(w8a@Ao-B_dnsJ=zt6l zwW3)Ie)igab)+vAEWdB95j6XaU2;K<;3*KgJnA-`opbMm^%-xoj=V-;ksvfiC7u?z zLV=U;Ef{C*KsEcI6*UUMFL?-&Y;RJrTrS9t;J@IpLZD?>QH7tga<{sGEkF5>iL5&y zfp%EN>^=#{$0MYlp(7laxLG|gpLQQZM=hLuFg2ia9-`kPsi;*RZ*6^W%qED1Cl6Wj z9LwiVE&>$rL5rE+!*t(huz+KZO^VOEg_%*5fAr+{c#2hg|379@6*%#_ubg_HVBuie*Dy%Omjrhi0>*p47v*JB)o=JzggqdkuNCX z5KuBc*WjHZk0Cn2q%5XQVQWaoBb%5C_!PEZld}$g3lba1QZ|qzng<@Ag$aMAdoUrZ zT8R9Wi^|?}ITbP2C82S?cn=cg&Uy7@8r+mnWUyS$4cUCIUTvNIBM7}2I}IdA7dwWh z**SLS@9oE*2v#Dw)$sP-&+*PZ)yyDk-g|XJxpfQ7D0UEBwn6eHOT4)&MOMfxUF2)2NWrbgS=%$s|!T`xGEvXRdYZxL)8geTVq zlL4i-E)K(l*d064L7Wk_?m2h~e>_-0qk`*318k~BRq)2AO9e)q`{ZwJX)a`j1I-6SV=%0aRU^0c3oBR+k=r~?XP$t1c zo;L{qO$hwtmrPsJdugfDr5e99z!WS|JhqQ1{Poy}(#PToUOrTrK*ig6pLXbwKcXI~qAL z3snUQ!lcdTb{Oaz-ohXM?38bc&&-hwX70N@g9+mLo5vz1wE`m()-l3|6!GMxl7B-1 zgbA1vnzmcSt*s6)+}SWJSZfrsX-P|`B_Oy&E>1UX~Y>%gWZuMVLDNO_!R&f-BnSj_Ij8K)44u3oG(`vkq zH^g(gS^LIj*8~m;!!5j*gc%m=-@qjv0s7d3Q$3HC=Lkze&4Q#$FhN)#r|M_bExLRc z9ZCnTo(5!{u*pTR;j1y#->@1oxT@+ZSez&cJ;lUSG z|2q4%qh8LPC>2xFvs(U}Qr4_wh>5Y&I9h6a*N(RvSI|T7T6mxaBp3%Duc%QlD+5kE zF7lz`^KidAfC2Z1OuO>eqG!SELSLKo!?Ow;I^KbK(i+B3qTk+sIqZv)(|Dwhe=qCJ zu{s9;2ztMB$RGlM`l@vH!{2`$CtM(R=0!h;{BFL-`idfK3i0_z9zO2u-;**s4DFV; zozrDeKd_}rXa=wQ@yF5RG>EU4(*v@(eku$c;fNPBrDAFn!K7;kv)6?%%TB{wI-%yp zpY+RBr?`m4W?Yulrl5WmN`0brewS&GxRdwNG zu|Da@v^z}hmy%#PaBI^wn`6kA{bOAwAS^W{ze4wwSb(D!8=pW?zXLnNr1(KR$wk0N z(`vN?)M(JzVF%H(atQK!0WucrD-twah1$80xM0wef`jpd~SHlm=NV*h)jMmoRtXy{gpB?3v46-7FsB;tA^oDoYW@8q0 zUV!RrQKdt!eKo7I5MQ(tHWzO*NDnlHRdYXNJ_46tGq-lnrhhDrsMew@(uW-eQ4NCW zWHf_eW0QlZ-sODc1cJtLu!r*s#roCNw_3r4um$b@fjiN9FTf~m4iRPrsLak@Q#j*6 zusGkcm9+K~(JLLK_-`?2*<%#!t@6BKYPQSxNaNovb(LD_DD^#Di9pL>G`I59*m)@}P2~4waSh|6(wjnnM`OvWDUR%I5Pu@N_ z)(ZWt1+;;=a87u>56I~R&;$C{4!s+@3fKe{|AhHR2lzD60_25qkS=n;bkqSz%rupD z8?p6b^p{r}VCmc)19%HJ#A#`G@_-9k0d^TqgYF90dywpp5@8gj`=Z{#cJk8x%g0Po zK_<0>G9KQpAAW8{i@%C92L?MOhUghIPP3=0u9)+vl5WG(Se`YHM0K7rIQ5L91(mYE z8&zIUrKFy0y#Cqs2F-(mNDWic3FV5XwrGAf1=KSNTueULeB1b(PfXw%)?5jx*7Yk3 zuFjd$q=K1%>dzzs0fJxqFQj3^b05iITWX46N7zvBmZSNZ(qwi2AnZ;+VL%&Mfs~R9 z1ua-IJ*}zemokFx+n1a)Lk!bpP*4}VZ$c+|ZhjL!&^G7eGuV>%0PfD?!+X4ho^x1x zz!4|LyIBZ+h^R1XHX4HNcBgjSyNV-z2Xyp~r9YvA4@%p}BRcTm_o zqBth(+zMp9it_@h+@^@Ky>)qq}?REb*UI2!6;FCOqN5J)Xxa zz}z<7i_DU~OWzKcf{>8~o~#+Ls303vy5KYT1g+8Y4{Qdt*uhHRB~(mgDq94S3YI>W z9lC!=^89>%A|jaxXFDLV2u)PMJ@g|bdehs+>%FL`f?Z&_^|sSb^I?=ims2p}UEz9Z zzrFPvGhz%N4GU1-U?)U+AqW`p6!wP?g0EQiCm2=@bEnHfcEJK=v`TU%RfAj zr^s0uyrfVdMO3Ym>)`Uu%rAEW4=i|a|2u#?9F=v+^W1AsBdP9v>Y zD1!!%aXyIH{bphKsw!2eR7(2qXcspKED@bSx^=IF!#?_facisB*%eRjdkEdq%%7U<8#SJ*!tm20pwC~Z&i7rE`ROkBzTmVO zYW!qiICq5A1QN3MUXj~4sU-w!R;Q+3fOPE8o>v3=8XH*WY|{4q`e{>zEkK1Ho_qef z&}TLc`$E$6s?lY&!vj;bz}C5;odwRWn~!WmRG9$k0ugk5O>l&)rZ#gZl&`~IsWXDtaGICXA!f;z3(1;b{PEjicQsjyZf+_ z-wkl9?PVZi2_H{9cgsCJsp9Hx;w#~$v<#FJ%&fP)XaZ&4KQJwQx7mAKJ@Ep}A1$aN zTFr+R<9|#z?33ar0fq0l)rORNHTz(i7);H~kntpj zvzg$oSs2S*Eh@**&{!|1_%UpK)Y6=^ezHok(GoOQq?OPv5kP^?(E4D_u)Ns{;uC2q6qlyE7-A zB&!`QXyD7cLAn22J$oMhcq+{aj_V$4Pn=jfpKXS^U_f?4Rb2FrGhLmpg{kqyCjnhf zS+OuIq7P7lIiyIza@8Lvi1ds@PAIy;9(JxAD)n$sv!Y^xVhI7T1!J86UkAa_wbPxFf@&+mBUe055+PbC|-L+}tu+yPTlle%D^o*Dp z@Zc_Z3pa|Ri%gOm5h%}9W8RljFRKu0#LPgo)I^?R3%|85Nejo#%h%l@Ki!GgjacsDjJSu$_nO?v6qEQPj#va})H1<9Zmon7hKW^o2E%4x z&EUW*F7&)-8lV_7>gbRx)A|foa*Y=LXh`ZpPSGm6wF>VNUK4gDG|CQ^dLNkhs-gwuiwfOS-*L-HuRokJ@ta7& zZ^BwMz!XC;?eW$Lq4jg@Zw!;u#Kz%^VGD^@OTbX&dM}MU4l1uFg`)FaQF`EV5Zv5y z16d}tu+1%X?#w&ZhU0bdd#oJ5X`tG)F3o~RGg?!wwh3CM85cdQ3!Hee6Ju1Ll}rs9 z1}dEudK5OLKr-H|Bvn!UHQSI}H9=TYk5L$x$atzESWa9;&N_HRJ?}nAk8i5S^zaJ2f{LDkUh;iNUckI z9ofgSyFMCd&gMa27t6Auscg^5FcO+rVsI~1zVZ)sjfb?D4dQSO4aG7pT0Xr4#(v*( z5}AzH3GZ!%A>v!urwVPS^@M0#&)M(t=h46hURE~~DuM~z8*D};Iz1V>_C?zQw~PY! z7H_LefQIF9QofP%IUC8Ic_tu>@oU#k+KSORwQ-87bv)pAmE8@=_+R7gp8i@1;Lj6U zAmf|$xtsh3@u*-OxHKenWzFw0z8RA7wT{>QUw(lfLeDf!&#!m5-}(pqr=zu>`cUJ9 G|9=6A!yt43 delta 89775 zcmaHT1z42Z_C7NX!T=IOiAZ;MNe$f&9nwgMQX*X+DJ40Sf`|%ANs4p|QYwfNN`r_3 zBA|5r_vktI-gEEo`TNY_8Tj_MW9_w9yzAZPXB$auDRC+-sXF0VJUl#;MaMIEu<(az z96Hp(0>Wa(A~GCeO2Q)GhnR?S>Dj5}IdP?$qhltWb5$;aM>LrfL?qM{|@VCy0! z;_Tq#VrYItMpMU4?@t>)C+}OXo*sXiiHZse3kqXJVOS~|Dd|5A90UJ6lMolK{RCTt zVIQEVv%5%O>{xvPOtv=$b{-p;u*oJSEh8u)CG+opf8PGdwmQ2WzT;#8H{yj*#w^sJB z3oHT~$8(3mUPeZgLrmGp!^b<&94tX3yZAax18YY)gq7l_z*6x!Qj5vR{BxFKQi7sb zGG`;KupS{cJlz_!q=XBKVc$`_&Bmy}sA~lT2+vc9NdDt3aVaeGa6|2hnit~Fe6yK@ zU}RVY%>^29X|N5ro%m~p;^II^(pZs~_S7POiT9sRI(7T7_DbIW>Vf5ddccWUTKM1p zurLFA7K#7v?SEQD$?TgxF#GIh!7y^{DzYNG z9DF$g{)}5ac$cR3ywPn$_Dv_)zoti-ZEFD|$+iuGUC55HfW5<>a+05w5R(v;62_7{ z(_s}~8o|L=UoJDOok>LP12YFAY=;9mTtXZxoc;`a)xgdX-NVi6pMs;r3W<=-di3@^ldyExU zpvD@U`$Z)s_Rj{QqGJDQQ2QxgjU5a1kzu{0ucr6oq#}i_maW0Qai*@l(d`HW&&Xji z*|12ML-s2h7&%-<_+QWIdppmN|5FVlM6)RbV1LFg9Q^rTPsOFNX;}?e+c*F7KpZ@H zj1yP=NC*rL`_l<-XXuCUk^iHXV!ze$+eGad3h{pk2OLK1d^HEwZa2NwarzY*h1fqH zONdMU^%z^b^L-u$_U2D(Y;EesGgx-9Ba9R-E`o*4Hq_$(tbzTP=*1;u{?$HwyYx@{ zGZMlQxWCrg?K;A!gs*wJd%AhL1cH-s%gNsR>NTfZ*(_`@hU|`57+-cOH;g!&Ar6*{ zt=hMxk`NdCFHMyYeav!j6=xr6$;g6X*tet@6m{4h?OxL`!Jh=?$bS(BSr({rR~*G5J2aGh={AeuecY1d%+IQEXZVbU3N5_oqP^)rS5uBh1ZwYcb`~&gO2upIvpQurwdq0e*08cR)~2p zli_F1gG|Ny{RxMV(fa6pf4qrH$kgyggrId3JHsWP)Y-5@y($~y56MzQr7XqDOvA>K3hVsIOKV==TkaOx! zSHmwxE2k!m+XU+=9DHSy3hr9_OJ-=eA2S-}6l>tX1XixgF;@}iO#@*FH4R#Wd|juT z>bhUO(vCWNuF$bR{mD*tCid^C?ct_&I#KHc1}3KPGEwEA#*iuO6&xL`R?;+~{9^tk z0bQ^w&Ld9KV8j=!1zeA~^W7x!*2?jbM_@Bi)YKeKgITT5{Ckrbu(d8j+A&1bP|M0T zSlMV>cj0#Pm2Zu=mJy-dwH^EZW`kl($W*i!c|kZ)^35+M>3GNnO=b=Xr;LDv(A{ z($1U`7&ub8%q%E)IcBH#c)H={Rkg*1o<#56X@6o>Xud0EX=Su*B{=di-p1)$F#?KW5|&(4)^R%}5z874lo{&IEfph{V+`_NJ6IZM34bH8u5T6NzaBi(5X{ zI2l1t!-=-&GqN#oK?qe4?oB~en}&x((8BaomAJ1x9H0>Hnv}=f3y*BDsJotdukT^f zEOcs(m|{ zmW%S|)sciVVQF@EBdPhJlcRKztFO8vT7ABy58N7(-zBUt(Sf|4DbuSZ(wLn3eH&4` zAnC^a=AMMnu>&j2pRffiqvqD~<-791FL)*W*37qMcTj>BHLCfcI~Gn7J;#$ZPC9go zVJ!o@K)(j_WNu{??eud!Fos_rDb!kFydC2UI@{_bVB(I5XhfL!no ziRCjbFIK=nSk9!5?_^yiHmyNa`s*a@Khx3VltAYeI+u&U2`IdbC#b%;xyg&ASU_&0 z(Z1h5nZ(G(PNJ+1o?O1qJ3l`!>ho1!D_ijXft;wQA&sCFcV+uY6Urp#=$U!7GL66$ z_Ir~k)7fkR%O{+n=Feh>J5ODP^1*H?6+E+UHic|LHr1B0qK_HgG9>DkUBrQD$UCPzdiftrs~dbSi8*O6v_&$@fFB}QD^r@HKO zq_C+dr-2NEiqZzpo2T1f->HN$dHo>0Ag*~#8WGj1OKxI5-0^VSt1@_{0hHLTxCaD+iykm=BdX!d)xIjowZl zTZ8Kh^>R%iJwX;X<_^1vE;3c4YuAe1KGvuqMF>T8r6=n>h04q-O(}Js!&WJ$9;MEN z?piN7<3BWctbT53B|0-mz;|uJK+nrQ251Y0Q0{2s(_fsvp_*rM>u{eP{XOQp7&2NWx)__78fD+)S+IkWF|Pe2IxT19DWtJ zl4dxZmyqCowl17K^%BfwGkTVV(e7i!#)hOToQQ_gEhknwlA3_BsSr-&p_{WN!{q{t zk!Q8_|#!Y2*5L1E*Dy%WQN=i;lRfl(x-(Zz!@Yn``wGE3REVQH0Z2iod) zpz{yQb;&Jre7fq$l?6(?3d6T{1nmONgS{R&Fi}kzS67k}(==$<-;E4*9r;{=Pn2c2 zR@v2h|CcPnOu_ut=CnE#iJ+(dEKPELo4EmhrG2%c`9nUXIf0$P%LMry zv`AVw7ow_?oETwNVPz0}TE&G#?p~sy_L^PHkMX67WXSpirZJ^i%{^RZCIr<<4Sb9t zgCi3ufiW?jrEDNwIC&Muw^kOJ$iJiBdU|qLS90pw562VuubcANdJ%k|&bHy`>++dxM+dNWODLRqj;I`hDZP>ok zqwp@F_fPbUg4LN5<~(ADplbg?3$%TD0+QWohujCQOiZa>-loED4}B;$cKKYwfFR@A+c1a$8?j{1BK9XnVhxEB_uO9Vx7ZWt1db&knmQU z_mcL~+k|6-zNlac8CVROzlzyD0+tAS3n%jJDor@N0OOgCj6r?+FyHNAE182E#M!+M(NQw#CHl9re~Phzu;P>u1Y;@kmYgtD6*)|P)3GqF zX4t8g4@iKMLb|!SJ#YG}^f+(lyN+jbIVDf5G7q&X29CFDHgDm4H}Iu39n4k4p!EsI zTxr0gxGm_1bhR(4iZVhKw&d`Xr@#u)QR{ zbfXjaKr|ePyXP<*0nf(6H=autsJw&FWTt~ptcJgoUR+rCU|;iTz9v8KXSBT1gk^Nn zG9g8BC_OIze1$@@w@S?fs7QLwl!wXbL4=zr!$*Xrk+&Ycz6^o~O6OztIYlu1$7;+JwI^n@mtgsqe=the_bB-w+W) zHmDXrrza~NuF(w8198`M9Vf#<1Pn+-j(X-I%h4O%zoMu)Tbv-4hQ@G$K;!WZ3AG>9 zsA_2F0E3db(Ro+oO3?PGnbYeW5yoAF3)g)`NTE6e#Fw8}E8D*))iP2#DWMwF(3JWU z_0hAfdg|lZS0L_!76ySVF4!t?{59(nP?ZZ z(TLHypAK5gGVq)4mX|)Isd|cf0~|D;};%pFK(yza&~O#MC6vzp3Q8SWKX@Bv8`s3$oc ztq)EJQR)1bo^JN;Ii$n>67H=^AuZdty~fwkz6ONSifgwXCk9y&B^+h`prymID2 z{ovWC+;G%odV)*Q@7XmnHlg#AT|g1P4_7a;lMmw}>!5(Zn?RTsU=Xr6c)iKe;$eCoadNJf(-ZyuB^xfo~b0A*}K!YQA<{G~;7#?RhcId@+CVb)26FXB` z54}Amy?p!J3)qQ@(R5li1V{u85Msv#j15TA&_oDss3715f;Xc0{f*BPHKy=-c`>7a zFRZBi#++z1$F$+N@)a(AxkPDFW3gu+A1<;wP3%$MW-6R))?wJ6)$!F^mEvs+V;$M7 z?OKuyNUazw&PhWYT?P~Sq0&JLCbt&}QYRtq+HWyqEaK?P13V-`u-zJ=T6q#7cH7adQ73%1sL- z7NT_5eJ=rR^?^s`#gRp0Sjb&_y{@i|MoqqctKz$bLq7}bw}Xt*I14WMd}YRR>TNyY zHF*JI9Z_I!=pYe?=wDyoy<9yoECbUvZYj@09BG4Zm3tqh!D+68U4Qi+7W-g4X!K;)QThjuf5r%ETjmC11aP#;kTWP3UTbdirvw0V=I1y4Sa=;1Ikq zefD2McV63r+)W-#hn>TzY$X)!V-Je2R2Yzq$U_*pZCD!gxwOQ%#w|4^~j zd@&*vY@8Kyhr`Q3O{3AB<2tg8g{KIl`ZV$qoF%lu#~DBK3~;Q}2yRc~z!D}f?rAmIP8Q~5Mo)14 zXvCf5o3}|?bfE8dq??`L|J$kx;G}s}MyJV)U z%*G#JvWCJ9fMCR#33?+N3a~rbU^*>`z6B4;4P!Ycf&G@xWyf`opuAY+68(|<3Rzs zT?(`-bQBuDe;z0oXnKS*YGlJ7*+IiPuvZqrA@Cyq{Dh~?rKDMT)H6g>BrT*khf&1< zRzG=;P8LKre_P)ZqZ>1+#FNvCyRi%LyQH}F>^;CiIKAQWxL3~OHtL*axNqzN57cq? zL*bxIz(zry^i5>|lnLVeGw_ttS#R&la-AOiOfi0TWjnd0B$S4hmaEM4ncC4&pH^jQjlh?#~xzuR1t5{+Ma?{{BgV$F{Ad zboM&3>)TzI{4@qxKh1~AF`xG~7tcq=8J&y2)cWdsBC%wBE3jdK zAUFP8(Y$V`%yna|yvBV}vgaJD>v)B=wkE(A)EPrIl)!KIf31!SrLBH@|C9rhFNcd@ z_Ca=wI&A|$V~Nqz5^a@uZL`$KCdm znAYX;wa+^}__+b_f+pWp?ONwIpOd6!^r05v>evmX5fIg?wgzo~h-OlLDRYid&K!qs zJiUBB{c3NrmTLF2Gq>kI-9D`}Da$%I1v2JM9NbzOoMxo0CF(dp5R-Rfp6K39cmMK6 zYSx%lzj+;CAXh-R$kBeZm59>9$ZbZW@Pd5Pc^2?`g?GArCgM8pdlU8foE|53;gk+eJ}2WsDTZ& z{d}g@8q;G0avXNErIQFGM`l@MK;-H)P!P{sgYP-_tJIaO>^+D5XG8rbX9_?_<|IHT z<&oENxZXTK6w?nu-mV9~zKOa`2uCzOFhP}jf77#opdNK>}e6Lhg)FB2$5**Eyy?w_8gnc0SpicXyDEt0uTzIfJwia}6FgJh=>0B_z#-6x##uqBY(KjPBK z;)e!MF`))Lp{5O9^Hlg$2QZP_3ic1E5RCjbAX7@i!N}jWuChkhu>rK%jqU9!2Mjd~ zwxK97j?Ri1Z`0?Ry?lIo;O5&YWKJfht(I^iR}vw=mRBXcLP*2W{_+C=h&ppT5>C^I zw4BKixmu``Em*JQHq}sFquUB1amagr%V}wWfLPm*PwVvL*y-Cy(SSVpqu0FOz|k4l zCJnt^{0)3;U`MkEZb`QZuiS9& z6B0M08Z^LZSanw^`7u=<_9-8;1T3$`)C`>K4%@o}PXCa8EekN8&EyaWJ_j`|>K}-h zk@5~2)hzI2dKk$~BMBbX*#<6Af19X&jMBd`UXfW`%5X9L{(xi#Lv( zy`kdx2Xu;MQMt`-D(BEmxW6?hWMX#~QCf8+Xg-G45va6w9&?CHxjVq|um{_#d@%9A zEk`3*TDImixR89=pKC=jtoMDhSgf$A22vM z*Kd7nR6l@la9*M%s3CMf zyk95CP#bi>Y(04BtO5dmKY(C4WvA8HD9Te{ytEVhY-|#qPhD>%-cm_9OLD%ToS$5t zoNVGTVzWSRt4{dK-s$cvWE&p~q{1D6pWO~1*Qv1^(Oa+BnF!t~>&kFNBQ;cIe|u`L zSg`f7&vUW?SiucKF8m&0DAzkp19E3EBYWdpwyqm^C)xP(n zBd3hOXw4$vNtZWfaS7xGRXssN>#5;aiOXIdB|a<10S4f63Nq@2XLHC1O1?||MXlH= z-g}Z7_&guj4VBqy8cmWx(DxPr!g(PVDbnHAXuESumBhCpy8lcU%#*+bN7cIzj%Bu8 z$QrAzW>_*51$YzFiV={5fCTTtYvxue5VWdN9z0RIM+4s4XS%F9~qG%l(4 z^8F0Mujz%s6Sy;X)Q&QDL8RZ;za{A&iPL)QJ26nkwu0A^z}C5`&DD_w4|@L0>YH6P zOinO;vJxtK6!maq<6WeiI=_fw|VInW#sF@gR9wsD&*wdV1L-LsNz?8um7# z>#ku$Ixhh0yNe7;0OVE^vhV%4RObaF5x3E1jW-i6_JTP|DuFXV1~>0lBXrE;dt3zS zYe-J+y?3SoWDJ)w_GCEeZ+&?!652Dyh=_WcP5UAx-RMS8Lueb8)P z9c70MrkCLzv@> zzJUxrKAkXYR{!NUo*!2odt2DTs|J_rQ>fy&{f(fVnrRM9QrbEV@xjqObVuabuw6+d zsyi1Xw7>rpd8FWZxac1Ec>lb$t@gMEZ0UPnZA&{YcL+vID1g{h1K}}sCu?yT)4%Cj{+aL1nE{d9-o0vs z3lOmbG3&|+LFqt~;_24cAoIR^hziDgTxO(iGXbt+B>zjSZpKC6ozm9mUB@@X&jTv& zw5|xo_??uUnhIHedpuOh!J)2IAyRGIFPzvIr1@yb^OkwCU{S!(9JA#Sd$)#; z)d#Iip`8OQt+?j-(NLQFx(r8wu~%i6A{#e`Z>EX59TC0cSHFk;I$e}}^rNT=dZBT{ zOhLqqY@kRU>%0Gx+1gt5^&BpV2!jtfgJi&p3vY1bcBrM(n~J1?1xLq&>mAo*Hg**% zxvqWcE^oU^|G}m5{702BCo?$mu7vj6zqtfR)`z6lD|@m1D)0x#>|6UGn+Jy_5|yDC z+usltqu+pX&)NVNq|$bYv}~QoI-N?tDK4iI0}RIzBL5Ap+^Rmew)9gpbP40Ve?UD> z)R0EJEO##G#NA|}D~8ykah=2cY(>b&S{F0OMpm0$=d9GFu5y?bo}_#N*N@v{74XjH zGQZa6*83pkRe1H>}+>0@x%ayVLk6(d8jqh@DgN6&2&XL&hBi7OR?it;7w9 zeZ92nMJ>;gmca7Yc*#Q~y#1C`U`u6SqjUaaqxs-2g7>lU?U&7z7+-4-It!*d)zWBz zz}EP)yeF0hzW05Mm4R{dg=AJe8?50whsVpBMSSz}AwO!uwytYt3gEalsX!CS7GTM* zSqFo_XmXt4aY431CvWFp_>V4Lvf?R5az9qSK=A2=$n3o6P&;C->}k(DGu2DN2@~e? zXVFhto3|c@Z;di^`Mt)W1HcaGfWTWMqKBT%X|c_nG26e6SwO!c-J$IRp|E*$;BKrn zeq^AkN9D(g5RkX~#zIvLwZIeIRm<82A`ehyUios4?)beF>W6fD|61Yqh_0EUGX$go z(-(Kqfl0$Q)JLloy=aOr=W>5{1p4U7`ksouj`IW%E;CN(NN|S+m}JKf3o&=PmMLFH zjP|ux1Wha>1}U+Ppqb$^P`KylyN9jz-@j0T_($}Ch$To1y`nlM9z5-fDdS-Zxc9vD zQ6RC_lV5ZkMolE?ocu}AF}G#Hltn`%GbxA_jXl>4AX>jswDL>XR3}>YQ-rP6o($N(>b-O#F4%Gn!3Q9>p)%8mWz( zp6+H8AZUNOZgX6FN?fFGJL31FSg~j8bX%r6DP!wGsDR)L>_Xkskq0?^1ndV#obk

TSU&(^v9 z`-t*v{Lc|xXZKn-;<@|W=eEwZOHv_YH#Q3;jO#tDEvg-2?8qopjbFHZv|2eMI*&K4 zoyl8rt8nE*2^+c5?{WwPK+lwSX`)G`Lpf!?T4>{a_)XW=`$OyHEO)i0Ln@rZ2_xe^)i@=h41kln*^k1n2Dm*JM?S!@ z3c9hf3xlI4B1}jGAM7!{atfooBPea%ws?LdFL>br1PB9tz~#2AC;Meh$5F5$%Ks0~ zg3j*v0{Qhl+HU{$h0k`+#?|mygolE^sM}17#U;PBVpS>>Cnw10F{b27UiPTla+>6r zoR)xvljX}yce4?ZNqB8$=7k;p#AEj0Wb#lZvXc}0VeTCY(?A{oEi56$-}Rv&h>fqm z_D-3_qAjHHim@mkK=#0?I7k4RQ>j_Y0K#=^cwAd7cSz8;&XHZ6`&$yfY8sDaqA55{ zyq4diIIVJ8<_TDd6|1COksiT00Rfj>Z#HB30WPnKH(ST5mQc+1%AaBugK599m%nhB zD{;egTL@Yqj*fQ0zSy|hy}j2(eHT6um*&85!Va7&=D;l%p*bfBFY_zU9F;v8DKe~e zlPEEH(d40k5wkO)Eo%hMSf`q({8v{`&ufz_n8P7&c5+RGI9~?EIASyUhh+Jii1ezB zC>ILVhf>O=`VW=03%Y97OEi5mdGIsJ3Uh`^A#ftgFNvLaMv?7T z4If%Oew}ZJnBVxBuEneduTpiLRl|sME=1#}psVEq2RezCz^lSUYf)05kCunMH6X^j zjjFmAF8XW1oDPTjq=o~d)I;C74jQZ}M6M|!5r4u+T^l0Y^Fnku1W%~HJq5YBRDT^Q-rCV+!>bHIYMB69uLg=aA=dz>RcPc|7nMI=IP= zmYEH{Bva(a!5$!=mkB#+$lx|m4%=U%?N8@O>}@@qiz363wK9X(8#NXd7BX_K&$hLO zZ1&KWzSUq1RQ>eal})H^&%uE{ZchRV*szNrRBpdMTFSu$9aXm+?@Y!AZtU-TpL7Dp zG*Z&<`=_EXjV>Afz5V0e*&@5nNXFg!_8?>!s3t94J@6%`z^y<+kaYP|KdMs!~6&6Nz} zsXT?yD>&Ye;>q4)xnIL7g2dc$4LGJ9NX;I7($zvAgFJDCq%@PyYyc$BtM~>qz$yUS z4L-v6z&UxRY5_=7;k}4EH2Pl&qu+zHcxh0m-MFjN;ITUVAtH^6l9DZvRwyf-LuUX) zvc}h-ML0f@rkh(i%5;z9VJMb}`vdX(FF@?l0hy~cN}t@%vvoTw1mH0>Hy73WQ(07M z;6$P5`-wCH=|EplGVlWa z&B`tgig@8MT}ry|(-VWKy(16u@Bt1_0JZrO7XN5JwXx`EJ6I{3oPx=t`&s{_WZ)+! zKz&7#{AEB;UJt@4;6&>|{?oqM-t8&{u`UlnA=99pwUSF`5-$V5vB;#%bmbIKbvW-| z@-Fe zA90)=M$WZ}5d<*7Pcwq&01}mH?mcs6wBFN}S{x4J(NA{KYlO(E5x-9;;Q!_50S$NS zDt`bp+zGBu$h|DDuSKs;B^Q?ml&>Tu?BE;M(n7uawt5{ zTOB2@cLngKa>_01F9R^>W^Q=bGDzT!yUv+b-&%glaW*Zyj+o*&KVSh#o#qLS2#SJY zUR2SU{kPgcIMu;{!~1tW!hz}}B3{$z0}^2R1L(wpIhk3vYMe0mF2J51;l>8lBg02$ zS&7qrs;4kM<14Texc=#RiEh4}1%6yO^Glm2)mM8<+hr{a#HqZPRH+or`_k3taWu^9 z^zRVuhwHLPB=Tk&GmdSBGf!ac&_5RMOw4WWJNK4g_foVB;L0`LxhP|ZLz@v)R_mfJ zZ}|b2P^_Md&W66Hk1BLkzRrY=%0Q|acQeRUBg zRBBdP(N%PnjF$4RF;l}OfJ?QwdU1jn#SFm+V4nCHQsmh6h^;4rS^qjugjOJ2jgWoO zBKi-E{kPFl0yc^F>+L*RUHBNs90ulO&Aax;;Nom@>+`V!H!l8$Y2vWjZv2(HNW zTL$3gE8Nja=D2|?!ldKje-H?NM=z!M>&dw$@%N^stgd`!@BTl{sqizh>&c)fS%!UdK^vJOla_X$a8dK(j9!z7WesA z5e(?nSBk2H@JPE0Kt(|IF4ZF_wvj+`${43wJmApd4EY51aUAY zRx2Dcd5#eV5pfefzI^`@Pek;GN?XNmz!lVV5h~vRATqA1n@PKH9Z=C$Y}ruw85C3& z1jy7s{NEfL|M+4>c_vuP8Tbh>yQ(H1Mr*n+f-xi zbJzFk8x{Z|cLepI--lQHTIwAJSyNy^1e^w-j5J)C1dw(e0rG2DsF^m0=R9m;h^9lN zkAejTV9*$JW)c$HRsGPASs|(LW_bf$n(6l$0lnLyrKP}^V|B8g7Jg^_ir_cb)Ny%} zaZk+}qUgz&4(oO5wryM>k2OngY)PFwd3CGcPQw{O35u)%SgAHpMXBTF4N|$8-b@rI98vvV0VE6fPQin`!trt0FbL zh%f|rwNgdTwtZ)Z_9{>Ry^Z++l!WgJApI95p!B_#%B-7TxJ|y$%MUwn-TncnFQX3j zplCCCy8C_73@OUF3k+dnZ(<`bet@3JUr@syt@-oDeEP@DN(-%9voB|!MiD?{LDRbP zT``%KJ&`l=MOjyUht~k0CTTzRTd&4gQ}e<{f~4I)iBq%q0aAfB8poOi7X}I=HG!6C zTpueh*+IJshTy^!Fa-_Z&Rw4+K3DSm4@0Ymj*;QVU~I{meyYQmqRpnZ z(52g3h~)O%Rdi(--BD_teh$UAy#IxKe+Zn&B?rmhmNknt)foLIPsXTmo+TzWc5wLt zR57)az8;}{89na_i0@gTCf5&YDJ0Er_Y}Eg9$6gu&mA#PFZ1(1B=Zq2mxqi2nU8G6 z4bTIaE#Tz(T#}X9x23v5mE&l*AJ{pw{W0K5%6JPr9*{QIblJHZiU!Izd_KH!86~08 zH)WvPNkl}Z;vtON+9kC-u-3cc&p$D8uUU7oM9**nKbgfm_{VBb+(l1Co}k*j!q+*X zCD*y^^$u3Q)LuVJ#9fw5a;OGqw07TbXBr0m6ntPU^LS&QM1ua0MwaSYlIys)z~yV%9Z8RyU>4S*B%)krzzn%)p!fo zw`%#i=bz@AcKeK;8ADkh)Wq3(-$~3p*wg0mGBeC^;J5O-hP9d$*{hmZp;?AN&-IT2 zB8NNdM!nxzrs-6>oL;}y#92L-g%4O?~f}Yx4n(Rkp@^0pffDV&M4e^;4n@9sBmh?jptO;<@43{;&XL~@2;qbnh1D`1z-KK@ZS zQSoFFEslsq0MF(M;TT}Ta5;I5D=CjW86EtxwM?!=ZV-?*hYw<@i`o1J;JGd zme28+GMd74Ztm?(&_lgqk+A18?f;s)_`{UyYN#N?l+>ciuhZEg7ykmx| z5KzK5m){PIHY%>{NQ%1%_J4qm0D>~(7|u5XKz*l~mOviU_vIF1odXP3r*{A$Rh`l) z6gN{lCAisJ&8~NE(&zTfD(+J@hIP*OFh01hK@h^k6BTA|pVOmBssE$?8Q<|m#hE92 z@@b&qI~_R!?G$oFE?g8xOW1v1MqoVbURkc#TE$d)b;)knrzYr}jA#dIk;VND*>c@` z5(AovMk8&&Y>Hlb4Ym;tmt^<4Uf=y}PkF&#kXmNHev(%BS}S4L9q-2X2;Q%o+TQ`x+^sxLHIv5=uuKt+gnmSGm40 z3jqA9mM$0bdGcPRj02WRk{ij?xX>KEg;=tgemLUhBdPg_tgSnKgHl2s`qhnl&16<`{2Dd>n0T)~o+z)dC{}~o( zQ>U!zJ(~k}bc%rQ8ENvpk-d`gvDxyG=?QHgYmxG|8{=sks3NOUERLHBE(yVJ-n_{x zeHl;^*i&8(3^`tm*i&QDn!<#Vl%>K4gfnbRfXt{BXaI{E5>7y`y<_tZ-DI6!8KL6l zrenNOm@8C7f~$n`TQ%q(+c8nz*#IR?Bl|v(gK^p?3xA`RbW6p73+qj*f+lYOPbRak zYc`d+%kRgG00=jYgCBLhd`X!**$%Z^wFa3CWbiybNxnouYE7wMarS_j3W^U-VbD&EKAJxv!!ri?U2#RGJ4f@OBZ#aSuxJGKnu zxGH%z067en z@8EXQA-G{vmilFH4KxH{CPg6Nab!wI;SDZ(DO%M!C-sK909P3S*yXP8Iel4xArt)` zi6a94L@M%dG>DFun&`;C^#1a?q((`#6SWN%#J%gx&JSGxoE}{P>G}QG<%{`6UJ=Pj zNhV`_8Q_Xcu!B+6`c;t<%{10{=zV@7;BeGCf_S!iT_u*ZWQ*MM%BA3)a-z>?FfbR| z8y^ybyt<<^jk-1&P_hlBEFQsc$AK+KGI$zZ^GdpkmYw_;!hwu2b#*XobSd@K=E!UP zQ-kRcsNbn$P!WiD;8ApT&8;r>1#med3*2OU!doC{+g8p=@$TKbbQYDf5_i#!AhIuM zvcK6?cy+cd)UYMs=BO&T;$?jQ50=}hhezYVKZl;+zw%sDp8k$5pIL=*qRc&=8E_eL zRE)b3F}?%<2At>iIFtrg)a2P%%Au-;=u6B68+Q)(7E_8`-#@{TiTcynV&yiwDGeW~ zA_1eR`PbSZZr`O{RlxR%rF%e{L;A`5LDRnBK)4B}rMNn%G=@{d#ymf0 zcU=IZuIZVN(UM%+TlK(nCD9NOA>fP(RQQ68lT59q*d@G>P zJQV>?=_t~HvJ1g^$JKOwZ&o#&=Xu0@n^|+iO=TcgsX5oERH<0lyXXoJs-V9G^A{PcP406U}+^Ac>g6< z!toA38|9FUap6|z9RkpoBkn#Bz;TLQxS@mSgALsn=0gB^ErFY28GxB(*?2{9tB+08 zv5)-yPDRkIU5<;>tkq1Wzn&ye4T1E1ZFLCXf5rZuW@PQ;*kBXTB#v z=b(vf)k$abCm=a#1G6!jcwwv%uun=y4a#e}S}VW>Cqr=GQbHsETo5pphs-H@W~&ww zXk~3y0vc6RJ7b)N3zT>jQT!MH{5`{k;5fc(U8FW~M(@j>NOtm61hXnRJ{2<3aqDyc zz`gH!pPKzE#jg#c=SX9tUC2((-oSu+l^NIgY}=09^*X4@9)ar;ZH1}@2Db2#62Rtr z$pC&={%^Pz9vfa8C%T|-+|b^+@R*~a(O!fgSh^aDy#}fwyzr7+k;6*N7tI5ib?5R) zXjR@kBVP}k;KGMRyIe%^Ytl=37!Qf_vv*$jfWz_3pMqpzK|E%GN!E3yeGP;Udlc+8n;)Nd;qsHM;YPe6AtOabvMRUX(8J33I|l=sG&O^-%U{#1CwaPr& z>r0B_NkfnVtNkphnhpFri{FVDp+xo zBu0&KoLEX-PS$=i*8HV1P2^g~8!}=`^XRTFN`>A?N`4#iNPHr25izuwUJ&ArS{8`T zDH^G}h5Gpg>J&CywY)a^0yXxj`jr1FX|@pT5jDvj^66_Iz8);lNVka50L~4djDYI; z?2U;AUy-`1`b{)_z0-1?Rh=VRq?h;r7*LZj&fiPk;D8ZQo15E(Zi*|l zT&w8TS`A;NN8D*frfJ#V?LA?$%NK@FYP}a&`uW1} z0-yPbU4PVS5e@0={hNCCXlWGIa&yI8lD#{BN#fUqgIX%p5yCQ3sNyFbqSNoV#Rj#( z6#?Z=)(<4r9k{ToW$+1oD;?i(G5ZL(MR>?6W%P@`$pksd9h4NA#&WK~DdY()zztq^ zaFdaeGWu%qcx10{h~4AhM$>)dv+eZgI)<`bA=k6fPsiVmMeap>?R>bA&mMPmO(?4G zRl9oLGwKg;Z85SS(er+SQS>g;O^`2@fghvl&h!h!qk505ufEW?9mRzcE1#;3EBpD({GEJQ4YJ?ooFyAX)Bgo<3K_KPdirP_7O!2mMr!pQ*B` zw{+z*O@D53@oBTz?-)27<)x--AZKxFBb0;LE z+jr+92qM02ud)XXa!x14Y$Hv%e=8}W4;)<36YHkzN_zPyVq<}jc3n&PybpWcyU%%E zcEU7;H$SKd7ur#fQ|U`f2xw`)g;V!MaEQ|?CQ0`eJ%4+&pm_9jXt&PLLOpAoA2Q=oVZ-Y`)cgdW-?#Et|CYE+`d7 z{#>8Zlir{~E-T1zU0hGByS4_k8b8NU=vAp8`sS;fTl>JXK@xzis%iP?6>l3|O=# zKNfJe*l~PguZ03j1P!|Df~+!R#fE6Ab@~RFncjOF_SEMtCvH4Na$h96v-`yPrgGpB z1hR-ohlWUNw0$beBh*_4-J2}{i+fQQ&=u1tP!K6nrK+kL{i#rc!fC3O>^D^$7a95< z>;nVJ{ryv&5kX^Pt{3Xk+p-Ro$++KlpFz#<^P60=WF|#Hh)w-YDqywI8Az;Z#&jep zLfN^{YGu%qvdvW03?7}V-y#=BD4gtFr4x0GLrH0>e-G?ZY6P{hvP<1k3JNxJ1F;AL zDtdZ8;4e}9><<8sA9ty|^QI8s1!F;oL`1sW)3r?R*}{SY3(@H9^=V67Ip`w@9H5S^ z=spjC2OI21v2fE(7@+Zc$4A^@P9L;-OnY<1{~uFt9TipgzW+Of3?SXoNDQr@bPQbr z11JngBOub<4Ba3|3(_De(nvR=fRxfbAV^5Jfc*A6@9$@=-ye@_Jxi9%Idjh5_kCa2 z_1aZ9{5pM!nVk77y1-BW?(rqi!wfC&`M9h?;-h>c2pZ<=m-RtvDQK-I;YExQ7*bAm z+#b|PxZ5y{->uGAu15NDIGueQ!qqG8L2 zGwDSiLf=vOI=jVdOIPj8Tr&n{qAbcR43?rKI16d8Pq%#T=daMCl(Z3W?kZ`GnO!?;>0QFLxHLU}>SwGkwQ{N?|C*e`ceAlR2 zjtg(n87owBK3ZvVU5&geoH`*DglQ9Y-zFs`!O#89ckTjj4Lr}-96uJex~msW5fD@D z7O@cp^xn7Ok&z#619g(O-VIj=o2D5wVuN5XUTiyp%EiDPK1NC7&=}v{zcmVy(b!c4 zw;?z>A{b@5qtRfRHVPaq|&^!Id!cYh#QiC|A5Ck=lrqIk!%yX9_Nw$H1VsIP1si*1=3dT-vT%ky^CdORD>_Zx-v2;=y*)!X)chadBz9_Ia4liTFDVzv`DO!fN-%XK8D?U zw)mu-J&+Y#GaqCYi&lMj6rjQ6@awviC{Qw&_p(?#!lh=1@@N+C3-g_K2drS-Z zKc*^p(HykXN;H(12XOig+-*fol6@-8z0kPH=Z|%_Bt={%iPU~D57r==;$6<^jULcj zr9H*o)8>=veoC^pgXvW8d2phG>j%W6ad1=rL7KqbI-ToLBDadi8OdM}cY{H2Kck?B zkT{mtwZYwcnXm2zf_MP{dfdGL-F6A+i$1ht9OK@Pkz^M2p-)LorAX14gqm4GN(Aw=-djSXMm7oO5MxPO04d105M@8 z%4GD$zNSC`=7>IteZa*5iZqz{90X)jw{NwN-gjOtC{ViS8pQqh;=KEwh4pW+TB>`X zS=adT_m?-v->0~pxN8A-NZRjQf2}qaDWdHyy0^=GR8$w)y?>V8`5$a<%kY6FX#4+m zINp%)-V`rQLogjNX+onH;;oFUs-Sy>ioZd1f=aA`TMT3KrhJ~l-`zuA_5R~V z!eRMN3Gwgz^I4Frua~uEz-!j)h)s&1rT=d7M%1~Y)fy-*7*9T?*##MyiCnk%ohpl- zCv8TdRD>sT<`!t$P~WTbgMkPN_j`PgdtOI~ldc;?BR=imgS>t-2;O|JFymFYbc_y0 z5%hJE)8fwiQ0n9^yz)_9jU#>#uKA&Azw;9@kdkrH0JGpA|AhempBV+B3XZl2hRK0| zxdBOFc_!MpQrdsF?nS4^+b8g@dJxa|MlMoCa!NYU}53g^+Ox{Yq$ThL#vxdgELGoZgBM`VV<&&|mUC=gJahzlrn zARf`()Q#@59~y$(2VqS*;3`tLnZunQKaMxbkL--txB!fyG$joRMCMUv8+qJ_DwH7f zf=nu46z@+~_wN@Jxson4;g1N?c{oW5_r>v&Rd-v<;E``0R-tf|QhyiiFhr27TR9ki zQ78ZowSEFv;~o)%7-GjYv>u%z?#YsNS;=>iJ4A4v@W|TD|T(P75cBj z$|rMCK}nUc7cLh-Xgq7FZ@>cDvE53BFdCYcxb^BTGPKGYKdv`33=DnrcH%-+f1JWC z7I88p9pXCKgnQGdjhIeYo7s#ns$S9i_vn!$kkgUT^5z85)wa35${_kB{fH)1gJoX2 zS`#a79E9qf7O?wq)8V+8=RE{pe*vrqBm$)itRwAibe-x@{l`FK=~lqz<(W>3rJ^j4 zW#J(^Ef!ZnD?pz@*`r~N+Zd)wgrQOvP-ILmJl}y$v5sZ(HD|QoknH@)$LCt!P-C;h zxLzjl@@MP{VUe%;#^ ze?k|z!SApLCb^J-%kp@841Al&(t%gbWM~9M2*?a$vm#Z@tZq)s;!!nr(vcMuu4 zoH-}oS`dhAc41~!g;Lj(IUYnY2dT#6WX^7&Sxm%Hur)q0)w2-Ed(3m2iT9tAJgIt1 zgbRsa4nOvw*72cb3lPzB0o6jRQIq_c<B%Gn^A=NQbxg znNsni#a4KQYz{l$Wk)e3Dju)?*?QjRwb>l(Se^x_EdefG%jlMZ3V%6Cb zctNQ6rnDhzeWDM=AxdOxw>xhQsza>*1e<2dYOiSJ{rL3ID#eDp^eA-dQ#c*Cc9d$T zHV!EVb%n0j5~;sN-4Rj!F3Ubyw^B=XVNuZp`gmG6000d#lqd^pS;1pE|6#loVLFe7G%`wE+aqQxN}JnFVJjM|lB@`a3Bf78}QLm)kXX_W~}! zs@)$`1_qF@BdL@N92YOFVHMOU@oR-RpIcTG{Kl*p=V{;THWBB<-nY_4&+Ph2`Psjt zM}`zJU9%T^FkBgv&dWYRf5Adyz)+st)0Kx~D$MV1spfJr$dim=*{7zvc#Q%->uwo& zR4Lmm9~Kh+omEcJuA}EoggG|UF`=E+4zGH}8~M{5-JMIPhzeB*UBidqyCjl!+ZF`? z55feUUjN$ERtWALUzd$G&iv&!(H&NM#D&{0H-Wp+?a%ZBDwi&*!Sr{0(%ip!ArOJg zOCJh89BOqOxtGN&(A4#oiG_(S`z29@f&iJ05P*_VYOIcm zGmt{%tUXYL!9wX96HT=oiw7%{8d+~-xeLVi;#12E67w(`YcLhmeB}m&%060&l5Rlo zx)n%S(C0)@a%UwuU9yPlndx_LAF?W*fT#7z@mw+J2o*m^NHXFOVL_jOit7OsBJbF9 zH{y%BVvNF4eRNYZ1tbxncip$JNVR|@ zp);9{@v9p8SM9%i9y-Xd&dSa4IGp@l$G7?=>`fIO`*7O$_0 z8J((t!uVx}sX{gG5l4E0hD(PZSxMgotMOgN0 zeRXjj7a88?1uaTsFAOdWZe9)YP=RxB74GkYf4#BC(CU?j{CCY5uso16a8m0ZlCbO!oe;F}v}=eI zs{+)cTel2b8}j^j#ZSDEWB13t%GFXLuG4H2bV9$rkW+>zLDuSn)p-i5UWej6t%@nj zZ1ZPdM~%wGNTXeFUGyH>g;l*gbFvfqqj%D=|I_o!VXDV1v{62zN2|{~^8ISb&GAK1 z2FGevTeGuZE_O+rs&(UG&l5Eli4OJ>-K;!2F4jaEf&Wa7(*PqupLW!rhiyIZXzhQt zkdl}Bev*&|oZcR3cJwrL@-__zdc!6Ig&;nc(Y-AZ1RTSGdGGIC&rW*2CwXrjyF|fV z;~n87{r@^+%(y1RAG6(g^COHVhC?`(f5s&gVKg~<{yAY@;;e+l_93gWUjY*dQA8ON zEoVO0ouhqKu3rcauRkudV)-Kp2Bh5-8;CJ8Se~A=y;9Zg1o~!Fmay!csXPOH#edpu zXCe+u?n)^wh1}5FZIb`lQas*4kNn`BPazzJhq1ec`l#iEXPVC6EaVaG>`3jdqPXT-h4?norvJu=jCq$AyZ)rP z(4|esA8f;X8w+|b#fKaA9|8xTy1<272A4Lf^L>1CnvLBavWbY~KZFRT(o({*(XP^K zwjqEm%@f(Y5(E`UpH3we0Zdw6E6<-dAInrxT=8d`cr82nLlDa?(W=3(vO^hNviyj5 zDet^qRDASgV_RzVJ^Hmesjbm)`^1FbOGSnN_G*md+5TZ~rP9i(C-2aY16xIV;oYb*@?arc(YHi3f2e)tC8fu%-AaWiD)cRFk5UaWOw zcO~Nk|4`t?nnF|6e$4==uGNqATA(}FU z9x=my^WHr6+N16{o>9&S&#iGCTZgQ-b$dj!7_w7M&=|S3i07m;4kXARKL5py#Q$#G z8Pfy*0Pg%N$#y_bTXsH1TmNfjrxLT=b@=(lmQv>D=w03FqrmxUD-{K-z~aejwG__x zC~QVyapIqCThq=4RpxRtjjs2UQdx`o)xWcYE6DQL(QBBYZ~))NBH@3}0Sa>!SXbCe za3yDtoU>mm|ESDKhLRQJg{{-8aw_)Bd_UL<8kz^!C;o=hp;jfAzLqg^6<bkD~^c{{jTQz9x7Di@06e?*MbyNYlaA$p1Jv&>r^yi7%ejiIC{9sJ+KrG6&({syYuxvC5H zB;1Au^6KMikIYnuyQLHjGKwG2%~K;bYt&V{V##$oA5``yX*N(`x4%l2lcjCAR`3@% z2`o)Kwtr^O5ZHF-pwX7`tilvHp$fx1PMmIg%aTw+JTI~j1Wf#zw;r(<8vAA(a%SLP z+E>}$p>U5z{r{X-yVlbmtyIUwX!xDGVELr9Uov(BEE22Mft%#DOrYNkuygzx%otg{ zjK@{$~3$HTfU0cm@egT3_O`LD^n{v8}q|P?qpl z4h_m)T>#-o^#R$TP;%|v=*m=yoe$fVRRl1XqbNx;kR$m85(ktzrh#+F={s25|9zx< z#~TUm{6FH6)hUwd-EaDv-)ciOc*zvBn>lh3@5kJ~GD-V=19ZbXgfEm)kTgYF4pDPu zlEU2uYI8cQki0<%JHxKO;jH(u9}aM{l=IP(>db}}`=9Oh6R<%Aq$oS-@K^95P$}0N zjfz!Z5-;1(B)CQotB|do7$i_-!h@HnDDrQ+qyZKe!~AsRoN9OE=HIs?I;!2WQe}C9 zYQ*2Is$lU=cUoOrHBm;-E&0^JubH3QGG4@1CZyy?FfBQRu;d#{-*EFe%k)oAA1TYY=o;L;8h)V*I+}?N4}mZLCV`X? zL3V*@LuCN0xw;H#fP8|#Yb!j=RTx}T8Vsl$Z@=eCKPxOO#PKg8eICj-(1}O$@h~%Z z0*4X>UIda7Z35sU#;HXoH1g#$gvfULEBlSN+e93exxAJKNjiSE#5zm0*wPAT?f9>_;7aB(1?WweDtE?W#V%28FXjP{5@Ytb-#RbE`l*B*M}EF#wPht&?tLWa_D~Bo;Lqw~ z6;1F{Ahg7&lq%a!=OE01rB8*psz4s)mIT+VH*TE7VtwdXUYyW?kSSta<`_%UL}97# zrMq0a6Tqen4IkKkUtA!s*S)D~Jg#R98s<|H0nC39+ZB~~f8A9e`~gW19fFZ_=7>#i^AT}OV)hTUv*zwu?Ya~6Fe}&( zR7VSaH%vd>3o|vxH`p@IeTIW22ghxpy3NM(Z(xGbd2{r87M76|N877|8?j4#LjK{7 zOUqYzr-7jwWcX-pAfEA3o$FUGomUMi*?J>7U(DPh;u4K^4CO0RxLEU!#T&bPFd z+LZh{P9%>Aebm6qCkVN1=n5{6SPVB+`s>hwX_Q#6HkN*Tn;Q7;#=ut{Pp@GZ7i_{A zL8*h9Y<3Q<7QL##SuOJq8UjC1dt|Fd07LyaGnoSkX{$gL4Uk0-A1NieXZ}vj!i^1t z=>m1j5Uf2uIO_Hk8*8m7O;s}BMy{WFfSnzV`GlGc9gbRlkKn&)tt~*tNR*A#L zJcs~>ZBfJ1K~^0-Hj0PoJ_Z0}HDOiP-$sw2CP{klKG_R@jpEQy2k9oKk2j#I z|GlnEr^eJ@!xHuh*t&kXP6htxH07j(=So^F&pan3onxClb08&ya zaFw9Gfe@OPN_z-49uMmz#@eXDBv=9LO8vkygyDFmYs+3g!CeR*?IOf@N)7?{rqMVr z{Xt1kVcYjC;lV6nTXq22b(*rzz`ClKwk|Fo(+1AK_^_5Xs0ZW(Sem|pRBx~NEfbLA zv<9}X-9q*6dB+REpX5N#m4?EEL;%n$bIY1Y*Lc?meFbbyEcDTNQp5YCd10-vZToZ<(dUr&6vn@dQluS%-c;=QMz zhyEk;sd88_@96i)N2*3pcM#4~fWOFv&qWGHVH-2FUlBjp^&b1;6<57|uB`!|=8*>FXypHd*^Gs3d83LaM z&~K{bW2n9XrfT=tT+G+=9AxJ&U?n@^hB(hdU=fJ$u(GzARV3!PmP6 zvIZbO!aG%IX0S(rEYFeV%UL4A@UD!CpEY~D_d~IOd#MFr@`V7#-f{hTIS@qJr24aj z-~9~$?#7n|2Vmz`N5R(sVCvt$OXmpzMXCW-65yC^Xn4fiiQ@!9Lk;h3>`#w7Om*Of zf#Kix^4@YqwT|Bt-b27~EKTm^tRi`@yf5b})^%b3)=uJzF z&sq#J#2;Jvfi9U(J(z<_*~(FcI0do>mn4qQZ%iDbjU9`#XhxdCJ|65of4no4tJhF;oW~l3ud2W`9eM!^TRc-~9Y9@Q8ew!Q|`}?5^erst4zd$Hzz%or& z_7MaaA9=0O^332FD;;2#4Ki*zvA`wAN3qfH#wqWf3Eez%`-mH45TGP+f3_gF=C|1G zu;3qy__NTi2?0Q88aTQY616W7D%#!Xy7~(U5c-pB$ejIM9{>POqrs)_sfxgja|KUz zoZI=$uleChla|uivBfyGJyK0Lnh~NM^DZyQh*LYa9Eaib$a3tJK)~shY125$zH!DJ zmC6VLX^hJK^D?))i%G+C_;voB6LWf^K?dq6iC<$^SD;ocLaw)y2uqOsyx z)Gr>_{>1dF1Go(R{Z zL34?HxL6(mm*X|@AW1JJ^%C$f7vyduS$x~!hs>tcz=J_|8&N7j~2`4 z;fo5teRWkexLoHo@OK^p@_FoP>obm=QIG`G-g`-(7KXY^WIR<+`V^lE#xWlKr5{Zd z*ap%tjri$3QdmlNF&jb9BqkU>a-02quzXL|3a}3t8)mS)FjvEJj)pCC@fk(=jB(WXCzw z=aOk~LmC}BYaUwl#VZoBgj0y%DU83o7zNAth_mL7Bq3`&v-azw(6Gu5-rVD12kv-N7S1c_wDIqLn+kW7GFFTNT-U)qIiJWs0+G6Pd;#_4q9XAFHZZ@tgskqkW2exe-T$ z+ZxXd@O78r%NpMy4TfBj2PU)RFW20nfwGs6gC<>*zu4Ri0s5RY)b=8Hm<53G`Kci8 zFBPVtYc$|Z>j?6-cGlgmsQL$YKu6p@>mA?f>RZ5L@_|ph=kUg*RAe+3X}|gBe?qRv zTy6UjoYTU6sz;%WOi=XTaJZEVG+qK3$eFYvdH5x6xs%UDg;|H`#mF2tSd89{)hL^B z=a)Yq-aTcb#X*e&<(%S7JcL9(T7iYtB7{W87|k7W)^RHHU+!+Tr*#?Fw?oe+fMMa| z>34~ieG3^K5J{l&B%v+>oJ={Wn~&A|Ek1FSPe>^T)$h86Se( zqTgkn#WsqJZ`P{$5-vfqoy{twF&XH+>W?aex&LRRZfa#U#6-8h?VPKt2paM#*r3NHIWK7R@f z)}Q8#Yb*jo?V03btx%NIu!9P?`rST;^m}}E_S-W2OpZyw=SEcl4BRgoHFqc4J2 zMXGEkDIioG4#mAgehTbeW2j8RgP88zyK<1VEt>AB3jnmy%-LzuxJB}C2W^83!DV*1 zqD@;?jti9Zx)AX5UjXO(ZgwudO)5iUGwleMNK^RR1 zVm(3MC6FRwlwafCmhXH_fDGyKIj>&5nVP~CeVn}`nvIZ_3ELYVV$ zMlL%i`o%cW7MkZy)!ncb2^$cJqvd}qGj09jJ$_jnal_jkmBfLqg&W+|GQ$tWsk}&d zjrggf3rVWIDC0cZWrKm!#Hf4SH|J|ngs0ahPIq^^mF_hCKW>70@CAamPEnU5X_v!1 zzF?&rP(@8wGQSCKrsc1hAp3&)a&S{okZ7fG5SS>eSX<)xN7nUm_X9fHn%fXZp$bta z&?C_b9*1HLNLZmgwf_ih8Xq6197f#JoetG(xO$SJp~>?!l_^p4;^f&P-GeXo2#78O z=O<|OpWxJ|{LJb`l)q3 ze3>5750Bp|ibNobLP>Rm81$;vTM-gr=Fi<8UmoA(uWS{9Q}bt$*3wPzi<)L4sU2uP#NYiiU#gR=6n$=4nSMAw zSUHzC*VeEE`|ZW_`rUFjZ-jr=Yt0C{Z?dts3^Y#IQ3?suBLap9dWMb#(;Fjs&TY z#UHe;^>rg0PA5MuS>OMv|CRq=qi-9vUz4^z7S(rWftMfKa%fF^upY2Itf7Jje|DqM zd}!GlDvu-Z@Phr@#Bx^{wr^D2$~iNZsbILN)wXpY=wJ#kj($`9&A8@cNv)Cl8>oiY zOgXh{{C`buyR&=Fue_p9e>aeJ7K!f+l7+hd@T&C>#IFs!OL;;F*1c(9Sh%cgKDQ`n z<0u4dwFK(eWu0`8#ouHBLs!Z%hCm|G*G})-MJ;mdc3AT-oWOs$_}TmZalVDpRifa>;9t=l6N|Oj7 zH2g-XPxT&c90{t16)dSIDf8AkPV5QQp0@N)Ts%sr)E=g&M9~bzlOE#j5bVjjw+gGAKC%&?{lxH z(Oi$hUo9W+;_=I=;_Y`qrOo_%*Nuu!FVKvMMgAx0M_Ll-rNO0cr(i;EsN3rc+g)4# zo4utD!&lvd`+QgOO;xreZ74C-zilVV9|{hYuGa?iQwzduFAMoiAFi}CnW=si$nZU7 zY_K~?952(KiYv9YW~}W%nVJUf#FYPe_WT`B8b5a&Ui}b7;7N0e-XtUJcOBH^*Jg6S zb->0?H~sg4g5qb*3%+gbQld`TH|f%Y9hc7bI)9_f^vtzG0*>JmqdGX#?Q$&=g)sAwxs^J%D5AtRtk8?KbV6b9SFWRGm8}MkSTFIwOX?! z_IbKKoV7-mNR#Ii+KQUl+O6}dpUNaDIm@iJ3$!gY)i3nHZL0s$`?ziK=q!H zy_$pfDg!O`>;8*cpYDrdNjMGLD88yx010i#%I#}kjIMwDY=%aSNvjwi7kMmXXL>{r+Fk8LBlX* z&lPYcS(>DZgOt<7!B=MY3GYq_UZ&r-4nS{4-=7y<3P>T>wQZYr#)_dL}0O^kOsJI<)>Dwt#PntOklYFYAu~MUY2VooCy(&LoU)WCxhrNtgqp6!%WgeN z)ub*f8MGbvMq&gm*(`7dAzt2XYFWxHiuIliu{M|?8VjS022M{QLF>TJsd`s8AP1Nb zFL63;H6&`s@vl#~GQ}u=o^=4X>G!J{$sp5#se=TiMi&FUm!zYv?uj4CVZV{$!<(sa zivu_UQ$n{BH6?#CZbIoM%KUv{2LMfEv<9ZWC9UAVpY^teML+uu-ERi*X;4Cn-#JlYv-Y7KA#eTQy4^-}}B zzu($!IZb>>TW#ljJ?4xvFngaONNCZzQtMJxQKo9oa|iXv?d;;W1V8n8z<;;wi>P`P zsfgpII&f|aS6WKrYa{Ec1yCUDbU0(1*Uyn*%mxgk+GU0_iut7@3@krwiSAoGZ@Wqy zshBbZ3r=6S%Mfo?8z+K8J@vlatHyVKfP73EnzM{0`(l*E7x+8Zk`##hb=XPE@Wdm? zEim&B@uVVl#h49GU92@?9eel96Ym+pVrQqo))o;vZhst)`9BY8++ZPYiD|1j)FBuM z@^fIoyG->5y&qA?4C>}Op)s~YeOw09}s`4|L_kD;Y1rlJU?A|@xz zZla7sFar}o+5S{Dr<=S@juZR0u`Fw71#Gm^EL?wVs~`#^H7N-WvGGr*P)9{YkK{532Z8R+A2zLW?GL_xSj7)qzV2dMf#N3lLQ1cW~UyeVPwP29tt6s7#C zd~-E5CG}WjZ}i<7ZOOOmgjCrNF1x@E+$}C1e zDQlVs+Il4XWM^&)`N(x*d!ePFT0V*bW>3-oH{>u(1H_OCWKkE{Pr+Oue>Oe=*1I{i= zOIaKm@)Art*)$BtTh;UD<;v3UcsE)%rzU0}ksahkXmLw_~LCjIesU8zG#u*q;SODv6iVHRG-4F~h*L9O#!lP^) zbX9=I0L$r$ceppXaS+^Y*Gr9060~oZq+Zk!;t6q!%Jod|+uqY1fj7~UyesjT&ta<< z#mZAgyC-H8JMbbzep9l)H&WXlp=St~??=@4Ivch7r;Jd)w<3K2NVo(!aY%nir2JrOu1qd`Ym7+Hr z8T+A93qScY^Rtm$1`zg2QEw;x+2*3R%|DaD{tEleOCJutj>#R8ZNuYfH$Q#@e=wD- z)-dlDr^8-|`~&Rs-8chp_DH#Z8T0KWD)zRMmnn`8ygmCcqcV_Lx$b{zvL%h!{mPd7 z8dL+L1L{~*ZjkAe4{)fKxu<--OMedsMArG?4;+`OAGkEeEO|kHQST2fmGhHa78K$b znLMMD)^l8a021CUpzV|2dwN#+>!e3PwM`1yiPZGy-oEbyD2KPN)Ys_BDjJqJNfuqB zD4=KkweDY-Yqm!MM*@^=_+0i`YnE7tt#Y&oFZ%PmCnieo;s_*9l4~MK#J(!O+22)t zPeUO5cIt<}tUxlo5y~fOwIwV{EyaDywGk}~8&|Ocys|4EUim^nBgg8_JWp}j4q@gXCVh30bTahM(HoNe$ z@i23z;qYX^4CmUfl|K*t>+@12oRBaic z3Y4h&5DBN#Rhv&?{eK*LmR{?bU5Xs4=Czp$_7|I2F()rDdwLs3up(F!c9`t<88eRW z7wKU|)JNB<=4b+_WTbe9V%c+5A}8Ob&)hGDjKYKxcHXsF8Arkzp&bHrNh3FcR^UR z8O+yjC|J#yw8YJxPTSf?{c0wY12iPb+>p?}pPt`34sO@&7&Ir&NzA@(EQyCl zoSuxob#T+nFW3C!>8)Nv)PH)o31v&ZqoG}b8(aMrgki;1ZpW}a@SUr(qevsIu98As z{;3tGU^rBCA?7OPo%^kmtf#uyjfNZ`E+5D6XNvbL@Xkm-(yI z%;ItIYJ<9AwU#e$+yIyHk^kWH(*g9Jw`P7>xF$cN9=WwF;PJsjNPDAMAdN?c_;HlI z(ah8oLEVBZeU7h&6iq)%DGgWTue?BUn>60aaHpNgsY%29zFfoX49e7*t53zYpZ^4E zAlBq19vrT=c+6YAiuL@ru#LC|{!}4To&M5dD5an+puUo}b8CIpnL(Q3kE4nPR6;}` z-hIdz(uo&@t(m`O#PT-uNL`=Li)Uo+a0@yRUR+CqMhfzL3pSkSNr$zUx(+#xjzDA+&uYyH>xrQ;$sjANp$L6HxIy-_0CN4Z=xbeJ{pFQ41!8gbcNQ4s7?V?N9DQVJ1i&h;c#^B}!fw{j zz>zaebMaM|px$v?u;7zR6$DnaZ3%MzZyXbXZO6KWU$9d0J`%MyJ|7=pJdGng5>_^? z)9j!UM0_8XXf#MZN;l3ZR3p0;GAjLP-R+#`>4)e1EJ3Jb=|ueT+)+fhz#PBJ#83ad zkXVMSeV-kiw9!uT&AgWR%J9!WG!K77yz+18d&MjWLZ$%`36I7z|#w;cEv~ zfI`iE=s5etsqJJkXV<)71C*;EZ+6d#M)vZLu)E3((2zA{f*2=YimngWn4?|*7wlICLAfLk-Rmf`=qyyO9$b>y?hKTU=O<@pw_wazY& zpB*r{hMFFHIvkRhNfL%#g{C5M8U6oM*SMhNDZbl-TDJ585WR%2F()cr+<>#FvI+mF zfqq-GrVyZPay_Y(SiJ4QQO(E?2B=B_V0a&TDr#$@bF1*l?**;2z>e$F>?T`bUHlgN zDS<2r&ky4jCUeJW)l$2St8U4vizh(+RtLX@@g&7tPO6)NKxto-^TAh3Tx9KV0GA%B zuIAej=2R3ACOFfXbprMQ48ql&izm?PyXtq6b+x)*^>f)Y*Cz>o3{giP|7_^b7-kS; zq=qi6PgB6QLwSh!(02?f-KuXj#vr;c+gK2jmad?5)Bxw>Bv_Y#YPHmV61R`90v$&l zw+)zPL-OBNG%Q_vU*m^JoQ&(a@1oeQPB0S#fWL6NI`~R&L7Mt29KJK(^pnr&m;!87 zvAx~O9Zk_$9;m;Y@3MtOw53l{yFwoVw8lNYeX>5bhNFNn) z*y5{BUcJ)qGDnz{n=(@GGoUe5pc_OY;`ab6WUZO0gR1Ut5zbae@|ZB9e@GYeUI`l6 zi6lLGIex!5MwsdwE2WXw;;qGo-oONDzhlyGglDK(7Be9EnooKrBKXYO=f9a#&H$)K ziu7#UBDghq0lk66 z-!RK$R0=}9kQ6gjSg`-@zIXyc!y0P?H;$2@yf*e-iQ!^MBN=&U^PM&e$V)RFEU09B z_WeTH$-wk)E|qQfrzOGhW&?_!;^y5)nf~I3SWB@ieC-*1>6dWk;`6k_bCynGEDq>uSnKE&Wb9;3?$gajp(%ZRTxFMF+okQ_N z58@O7k@!2-0nnWX0fyRH4Q!hfj*S6Aivu4f4aoV(Oc*NPP?tvZ8K z15!Z?IRy;^Cx#O!VnfbBUhhrBLuh{@yjs=7V@xSB6)uEgeQf@9)@=Od_x2fLS2uro8+JMXL{_;8+PMUPFB!6HFhO6 zfL=|nqV-VS_hKz00inFAve3v{Lu%&M&rwl8(ttc-`a8hDSx~*Y6f_?2W$XT|;_M)2 zIj|I>a!tm6-(7x4d7MV7p@0%TA-L9WRbjg5vYD*u4L+bT3$T#lQvG%hioeaCeY{^* zmCi=;<7uspX3RWLokgI6&lrbIbdjtIE=Mb#J*>7VXnExnEvJ42f`^wKFM!V z8muPfPL~V079M+aR-O2?yP5q@$Jc zeG=o%nM3Xro6h$7C!yFyad_?;7FGCI5s1XIt{MO^fFbaPuv$`wVBy*Cf~+2jB$O@X zTFb{$CIrJ~9?BHbPC8=W{|sQS%-T);?o!@&MV%Z@)zSWmm9OQ%bx2w; z%JWH}!5Mfd&>e%@^^)W+N>D}}2MeDFnoeE!++osH$qG^&&oN3Yb8cRHzqW}BxJiad zD}Zp(b%g@7`vhuSH|p@`D)}3{P9A(U#b9KoWQw}>RLBL-LN5hhIj%^}9`waLT_+`Z zIH|RqG&n$Ts5@qzWSfy5C!9CVUo0fFvtUMXneERgwfqN{A`~3XLP!WXfRb)iM%11zM z?|y79IK?0{VHa4YzWcpsN=n`$Qk^J&5$2>}Iv!n#%6he!Y0$ke#t5zME z?_G^9wM(!QTm03MWHGN}b_;$@a-r3c#6L!UjTY~HsQ*UHwa||Kkjeh8Q0$q-b+k%; z_ZjWgCC2PE=hULfLhy~R68fa|0V9Cm0sufC4COkr3O7n#0^5mjFBb0o0kgpmbO~^D z$rsv{Sdt6$4L~-?3!ky@e`ox5$G;Uhph@^F^(QRnVklQKK+A=bW}98VY#-+|$(QoI zek?m8**(MetCTgs4$RpFcS#;x4j#@O?^)fDGT`^A`-S)!wrs zP?<4f`qdDX4W!IW<~5mmKyc<25JXZDm(=Vzf=YM`{wlb*a|8Lh=0Eb))4@eCU zR&7R58VWou&ngc~g6>Db(2OC6JjTQYw~86E1GnTJMAS2dLb95{G4$X2WNPN0Ik#D> z-f`0V<}$wwI%y{N+_WJvfhS22^`K^M-;h*WM!&nzI|cdx_InRp-^KdFcCk{iYcA%6%HT&?D$OK zb<)vfPSvo9g4=ugQ0m-$tuCyAOe5~saHZh|kb{5Y&(M`w-M12HnZC!pC@k2qBBhis zMv@=G(V8RUzipWkpUy=$m@@I<2{}w;5krSFTz(*Tbb|^Ac)^Fv8KuT>_t!;dDPA*% zSQh+RRZ*Y}$Jy+frKnK)GuKdQ=%XA@OO^&d7+U^Jqm@63Ra12|=awD)=jJOj$|3wg z92cRakuskFU)5N2Mt!uq7m+BzR-y-8b)P0b;=neWye#lQ2PMkf^ZjG933IUx=~1halXxpH#0THkFNfz-#;Iu`d5p50}la1b!`zBg#` zW#!^%MArEYo^P-RS_DdG-t(9ghu6X|7+D439C$F&Gg1oAPvSs2mEnadp;%ybacx*~ zWMZQ6gd)8pG)}sFxLk+gZeg$Pykl%kukDJ;uSVSj^SLD6#^ttktybL+`rE@4p=kIw z@{XKc^gA4Sm;KLdooqk5QWAU$!H45pRc&Dvi6Jlh%H+?d;;QPeqt7P$o@+MKG9+{Z zosP%Znv!}aq-OqvrG$NvWCk}5yspF9a*GFx7Q;wnlm4a%b|-{F=jsRtxk z*eLy)dx+x{>9`5y#UUxHe_#C9*9>xgLsQp=yDdZ4*h-;$3DpJD#_pkdrU zO;;`doW)h8aV`DX_ph;Dv0w5~Mc3w#LbDgDqt99AS@0VlRMFOEsoW~)vUt!(l_V?& zP1`3~;^ouTA;H3Cjkie>evAxdQRVx5CH)44%X2rR(D9h@HMR_tX#Qc9O$eEhv?}+k zg2=5GVq-?3bhG+by}ryr)9HdijY*2%&=n#3MNH9xJ8<4oYS?8~0nSUa_*D4JNVbo7 zDx4$AtnNp$KmghoX(0YCj4d|6fxu9Dmx^$7?0{#T;6ch4Q(xY{+4JDDTP#ZqEbNV08IlhO*CqMwbhATa;zb;7&lc;$RP zYca<{=f(@en13#&bXZ?5p7Bv7ac<93UL@W8k|UNs?;JUl3K>TSMaw_Z!NLuJ3$X1? zGg~}WV~6Gj~Te{B7EJXP-(K8)vda8PDShR!h)DvD(2m@$EAFhZBc!stAGY@1#~-01lG#DUvQ z=`Lk(M*MH;U#k4vD$2Q-m|*FeDnRnOdT(qhplDW$?Jiual|t#RHpkyJe6Vh(W*{i@ z)f3K&=|1XUV&vD5#@tie;|YE*Kb*zZ#gL{+KD|wfVZ+e{yA%f}#vI()QkhknHdYZP zOyJ3oK-#W=b^7^ZD-$;&YK4}#G9oxb;*5^T+BYh?>bdd*mx}w#mQ-2}=*S9&4XDg3soKE>jxai$`Rrmse56V}efq%kTU4~XRIVX{Z%KQ` z)#|5NB*696@3GA!=F3qoJ*Wl#hkluw(5Ue6WNO6eUZG|}n4+0T5ts3cz^FrJzv@|%ZZ-+< zka^EE1ksHb)L!tA0uVe*=GlD;+t4k{n9a_&y%J&7X@2~|V61xfoSSlb)QHs{B7M?475hL7m0yzoqQPBi z5lZ8AVRGx}xwbrPt9G-eSn|==D$XuN27mU3Y1hr3toF(3Wh3u-tbSJ``7kji8!Kwm z&EKpnrlB!c!JOIIRW1quPc9R#cpd`LY(G=>T>(-H8Da4HDrr~jB%s!wdze>W{p)o1 zz7{^lHnkR7`IsHfu1hMpH65#KwOdJ$h$e4zS<0@NF7@4ZaAc)W!~e!Xnk`J$`s>r; zWs@uIaMyEeg?72ROYh}REp>$=vshS%y}lcZxMD{4k+~1?JvRhOHs=P|;73>f{b(yP zx&x2HwYi)1uw~g)35$@mD8CMW=762WP5Zq-^KIPfc)^YHSXPR^1V-C_mIPwp$fE6# zai;!eb&?;@&>`&Vj|}F5ZjyKng*HyPe7`RGkU7UNebzjscZYsf!L_qje(}@NB^@Pp z>p#j}=8mhTuE9z123S@Y>dT1i{zf%Ub^+3`PVaR*>AED%SgQi2sLL+F{7Su-a!&O* zt@LWu?(J6@Rl;6><9qrJ)=X#paZ2=3i3&8j!nm?t@T=f~s+xEi=^4Gs#9vf_ycAMYY7tWGE9n`_{gJN7#`fg>KA9DBywiwvqq?$Ly z8X{!3lQ*v(VRhoxT%CF=HYoO1tn@9hNNbJPpx{oYYMy6on%CUPEzi~Q6IW=98MXP& ziFn5y7}07ylt66s8Z*`0L;o=s9zz;kTvY*5=PypDk~C3F)TJwCxjzTC23{6>)!UaJ zV?VXYDB`h{RldcGND=Q>7DkNlpI8MTH$v+^-o)%!vz_-p0oTVx+Zg5?Tc%ZHYw7UI zXgx&}r-4r^obmBN39&^VvpF*O7;Op9Y|y_M=Euu;4(l)G9ZR%xY0nBHE|@F@Uq2+% zTVubycvRrU8w>u`fvf(Epkc0!5)#v`{2@zFnS77LF@&>jRcXo@YsWGRXFp{h^qDwA z!<19Bj-9P$Bwx$KwyqJWjYFz_l(HMb9#A7PH49^u;uUCEgkRS(b8v_Hy^+6wz1GpR zzj!RAg6m10K#5xW=>dHU$+3*%FlksN6|WW=g>ODENyDtB5S3#Ko5S`K-PwK`0%om< z1;O0996~hDWf!h^dI+{$d54KEV3z3aE8Qpy2(<813Nc&X0b{fpHi-qOxD zd**$px$u-)2{gZfDgL{lXRqDXUXdsd?y?@XUSau+mWY^9X6WLZBld=h_4=Go?Y@>- z2hQ&^9+)?HVj`sdBu2Nq?O6xn8&Vsfr-GeT@uuskB4BPsGYj|eq1WoGZWGdfPx2If zl2d{t-M!kjH~BH^%ZAK3;^+Se$+1iPWQ~@(qj>bp&gs5nFRf{d!+9eKspghJC+Vsf zpPhoUbi7t1|E_fiF~VIrUE-96#5FVNK=GCzx%b2xioI=15l{GV={Vb?=$byDF)SU^ z7NBSz{=RJ4VGx=&;ZC9_|M6bAB3r%YE4YEqcW@gD_n~~@0~3!Wkv|20yWOlhdsOvG zhT&2g!*3rOn3?EnReyS^zkEJDN9eHXT=gxlZ`u6tJ2foO2uWD|y z!HpwQEmANjLcZoQtmM{W2hpAWY*gT*0Yyhy*dXxv$RBv&u9%=9(=v!f_dmh)4m=$y zn6hK2E>&JTwDM$?N9rOvgKVPM;vM?Kl*5q1n1i^(w|6TeR&l@i|5zd+ z5g@}Wvh6*7&F|meu9y4p0EIvgCN8(L^Ueg}=i5*M=49(I=EC1(gRadmH`-^~Z~z<3 zJdDFd(DMkN^&lDFmYRt(407%36Y}%xmW#2za5aTIKHmj4(9E~pSCRZXRH9dLE2@++ zXDo$mhApoSToNFaGd6yNlO1o?QC14S22MBe8RNgm@!U9uyx}HzZLVUZl2+`A{-)$H zt6y4)l2=pi-|a{xrhHD1JyB&KPIAsZ_y6%lo?=By+=cjV1Zy}k3^A46b3pFMmCHH5 zqx>DzeCdo{pLaSJ-!qOFwOSGK>6Li{xBd(8ZO6I)KMp2@!ui3L%>mDs#(du8H)4i3 zWtxBN=bR{?**%I9|Bhlji=O=Pv2}On7tp2v#pK3toRsL{Xcz8_Wb-2zo7))?j#BMI zs`d1Qzlv$U-ricSd%eys41qD?GTiQN3VmJw+p)_3cI>k&I>Zs}!f@~rrn7&9*9T30 z`b4S=*}1@^?1O?qKuz#DftcjfIRi}GzSczXxme=oJWDN!wPH~tJEfH+|F@)9 zAQ&TPH14lrsjN`($LphBi3Au`7vW|b>Mu5Qvs))fKE5v3Kx6xUv|TjR+omKTe#hZ2GwgQ9eikqYN zGR5^S`f8;YBg?V^r0g`WbTh94kk_?`oTDuQZIE%h{K~nB-q$dZ4;@Nq2OxKlhB;uZ zXSdv1MY1PnO`uC}D(uLMsMEJ*H2_Dk1Kh?Sok$@Ih$HGR?tJ;ve^(KLxvGkfDHiH? z7A0%3*U4M(T&hOh(_m@@#i0A(Vq**>8^5D6h5P2f&j0+a9rrrPGJ zUFB|+{@Xt-_$bZ(m~*d1^3WA6t5(_*9&2}Nd?fisalU>_>a8%!BIca=G@tJ5-{Olk z3x32DCSFqC_p@IX?!yNS=O>}!(p3az2%%1KV9gZm1K712Kt`=VsZ7DTCIHW4@mSjF zn#KXg569ng@hNz($3jb{1?>Te>-a|i^?B(IqgI~28$&{%hsiYqHO?}aM<>enxDGf~ zB)&TN-#wIhc=zLb|9e!QCPoq_?)bOv_3`hBWdN?n6oP%JlveBKKJMgl54(YINs$IT z$#c*(n4nPk69x`^?E#3#Ou`2~V+>g1a<+j880WCJ$e)OsL85M*&_9^~dg??1+HIbX zGxyxn`3spPYWhzN-DVaLEwsATkKzdqlRmm(IvBx-K$@D@#NQ)KHP^*N!9e>QXx`y5U%GW zM5&2nO#=J(`UAr>a+pRhA>P#sFD=Yr-YytY+}&AH=Q!n(eto0{%#^qevSBFgY%#yo~Dn*>%W${e!o9x zr;&lQ9NWpH^<&C15Z8mF#^&k{J#<^xe~s)G?jaaxB#8|pO@nkbZ&%B!_!dMSnV?o} z3VaETy9Na&Cs}TfgDCKr{vQpqs@f{ikhty>h-1?o;|dS^5A#`C2V%cw$Ge!MOz*A? zkqJjYcli7si1+X$FCh$xs8~&}t}VB_c+(j!bZcpKUqM5^DqPj2iot2*FrT|^;uX!H zp1YoJpvye@ec|%Dd(_r@`JT*w8S3ahh=)7^r1qv$bMCHb-Kh$Qx7%6?94>==qvIXF z+Aljx)QTAwXUaCDoI(Eh&Q7PxctI*&~K)bTd;rc4}1hC)Q{Zg zT#L)PsR5iBlrTOb|K*?PiV>sB1Ci&If1K|-d8R=+?Mlc4BCJR5d&45lll3NV;k!;= z!1`C$h>~0LN~AMJ7=_z;i59#p!7mezT1p^nqK3`4zQGiWA;ExSwm3(oP&)CdQwK0= zs$%uj;DdN%g%7A1QaSR z$z)t!TKw~SyA3wiqkA|zdWBVz`|&G4DP+NcN2RB-bNMBhfXC7&KjVB9q*T*`zM7`% z!h_DmA&}EONXHfZ{k@1R&zE$xf}t{F$;$+Dk3$c97V^r=lwCw1Fk13; z!V?12LAO9Q&~tl1AVuQLaKu7%kd95!=~2JEAJh^UMRdaa0&8WUK6%E1xge*S?rca0 zyKTag5R@?u(WAr8b-dVIY;Xtq-{0@-04FNTNK6MsWn8IR>*AM6Zg=tQe>ocyZuwU( zRg6bG}Zk1K8xTe8wyRYk+CQ~zkRHBTXy#?p7;JlVb zNj#j#cCIsY$3r@GWQJ^U%-^B^2puPLR^V*@(GK~Q`!^Sdhb9Bf_VdVp+3`r)LGDKk zC|>=SweR>5qxvBGy!%K=(Ln|5*)DW06!x4idangCjSt2m&5s1tYL%_l`~Ef;66kR~Esa_I zI>DkY5==ky81FGY$sr^15MFL|&~WgcdExV~m59GDUihu`gj-J8qus4pqqMs7S^F(!35oEsvdEQLJgq~6)b5VePipd=hB2u$aL${6q5aV z%#K~b*E^T-(eIB`9MYQ%1G9Sjm~y7I(yzOQ#nru3-X|mf!pKivDU6whkz&hVUMVAo znt4iai;Yvpw^MRA>)NCt<6pkiO5ZsVBl#uXNQoaxI6>3}op53>O?eNE2-1{l`ady& z24m+2NId9QI1}E0e&b*pbEwKN^y|g}2Yz&>a08_t@^y`#vf7njTj0kz_qlJ^Be)1mMK@T(-7D>b#XnKP}G({@oH0i|t*_-rO5-pswG?c3DQHI}Kvs0NWq`k#_?Ow}=wcA_s;pU%mO3iM&a%|s! ztn#lYuEnQgwS@e6qwtu3;|N@w`;`10t{Px72*+jy7|DlhZ;($DRMSnMZ$KgRJ|kyE z6O+En?}u2r|8?{vd-49t9zax$^9L>*Njst=aYu3K8X))&661I4|5^%+RyvnortxmF zputLBRK{M*fdhF24x~PvVaBG(7IdMMJ8!E~)x7F6DFqk%0J9C=-dt@_T-$)Sw z9hb&FL8$n~vZ?^|A~Lyz_%RCy6R%3|R;kQNqpy*EIOdPvP|j|$gttXhKoW`__*s3I zkEET@5gUED^q74bOq#h=G1MyJ*}$a!sD$)la1)*~;_bzMb=e6JVtudvhgeYJKBScB zTOhZpoIbu!aq%my)=?Q9fYjyQ2(D)}aj)&Kd|+^w=No) z)8geaY0NndKf4CYag24+mWtHWbJrca_Jt7(K?Kh>ztr7}Ch~ern}1$MNKUeOV=llw z>I^L#cUXS|c&G5!mclyaUasluxSoP;(c{VRj6n(0qL%25(?t{#O)lDQBAif zuOM=J4^p_TaHAPRAUB;BXao#$lF z00F>bN4aVZvx|Rg_cA=HJ%W}sXkv`oJsX244TdZXgu4)cRcm%PlXTX1d{fp~-4 z+5CN$V-*D>JE-3>2}jf?%rf=;k_sh8^QzfGxPL~2&VP!B;-FcHnmg7o22^TthX_|Q z(HuL7;jqMF&Gw5@r-6%TCMRNCZ2Q6!#*Kc&ULAAW{;BF+xe;9ojZB#>2Ns*MqTDSj zC3xJL37Tsc=lWLKl?lNmFv^EbX3nv>#)GdId=4;|%8U#~?abN8*XnnTCH0X- zvYT9vNqZlTRon!^ZND|HN2)`f5A-nFxo={Fj~R#l8i83-X!{ZI3*G?kx3;whXbpz; zEkR4fgJh^rr~9Fe%SxhhX8BsH>!k^}%3=)%Al{88d#bdBCioHytbb=JwnG`mrMMKf4){Cd&>a^sldfvk7l0@VrpE3o}OXH7Ix1 zjc5ILb#!AUJl)?$JusJ$TC|q(1aXq!&M+FnyC9C8e{gRbHI=Eqg*C1f8Z;Jpjs8?f?2~`9zsg#G<-#YXCaubq?b-cOu-)D+LUoho!j$wSf{k)5nr; zR8+rnCSD}pDwyX^j;ma07o=}GsgNYm*nZMNJA)R}kM2;t4np>GJ`PDK)jnUy2(kL{ z(bc+!ghzFU=$r~5_SiCER_+e+zpQ~Z`m8d0i9j`$DDPG55)`_Bh)m7%?I5khC$%FV zr}Lrmxw7~B|8D+2s(a{I`6of{m5P-Rh*%*Pf_? z%gDQ+#1doFi=jS{E@{R`F80RIt1`lcN}6SJGQ?Ga_C=i@p{3AFC z`m*Zx4-&wxH4ci)5u*k0oxOqd1}8bc61jRL)$wOpy5qn6%1vQ@7yC4b@H>yW)pNXY zvi$XD%#?@xj%Wj3DJlofMf>%Kknl`3GdBiB?q{2%>Er)%hH07Y!Bfj3SGiu3_B0&7 zO2&(-=|d_mV2M(#Hv@xmMPRH&#GOT`9q~HCoOoTvcvEMjs!?^BPKflL%l*C&cy|O< z8-ee^-RU%=j>6L^H@h?e<@Bjr{`~yB(8@>UT0Cio_;`7y<3^2r$iaa5*z>y4?BXWP zAZ594dU&z+mu*VlY*$uz6^S!L?Szuqo{fPUZ4f@c~mIhDEHQuNtksX%pR4Q#O2Z3@;@Xp?OcW%xXKUS z5*s%9IvJ|f7O9)xNGz2Y16#y69rToT?aJurg(h=`hpRx!9b{G;mp@o{%))Dm%_ibq zYlG+CLH7=hKIM=@m;2wvFbgB?cG(5AXl~j1UD#0(qM3#qW@X0dVM5v^Ox(56-x3Y@ zvkvL5oW;yvokEz-M2X!kJY`H%@djjcwPjP9G={nMDZPjHlT{H7Lw+y+6Rv_HgUT!n zQ!W^`)T4*Ur+=?@PeCiI4Etq&ODj9xS}cwAyDi~-^AVqwh)`jTpR}`8(7)w|WA$=< z_b!g1?=3n)CY<%JAk$s$gfsc(j3Vu4n3M`(i#+4;HzgqeNZCQQ-n7Q(Xbk^<k4n7D9M3w9^022LaeI*VRzJ>gng(8e8`*Z)*&j8Vtx7lgcU3@IZ{)6HnunVd!LsjE8(SPI>?!Yn+>j8KL~u~v_YVvL z?R~Fsg?)-O1?!3H50hw2z8=Lj@abDb3B6rCKM-BP1oJo03gJeF`NA9CQAaF$`-s?4 zMSH#U++bzLk$x1mq?X!+EmPvpUr*-E+NwT_KNM=PJbAL!#8U9T0wYFBeWU=wxGR%U zC0;fC*8q#~55ZU3G_iC2th-bOJ)NjDJ-O?;h!Q?oRm4k+rorSY75c*n{PBA%>XztF zFEJUh@{*38B9K0-U~?nT6d-#S6!$Q3?!CXZyk7Vw!LLbMCy86EOz+6m1f8@a7#hA> zW8sYN0B(v28G2<<@G;+g&Q8L1K`CLUXMm+0ai?bwvLrRvnJd4)vk;sbrox zjY^NHrl%rh6~qhsnQ~sE;64h$_oYB{CjYgC$K1!YxeQ;WKg`Dd?8y%$_)Ue@il~0! zgNNUD=Hr4TN_Ckr3FBDw4RB1HN0{8=hV;G@R^(G*3GsUM|Rc=n2 zbAcS&v^_^Ty;8wK+@LP)+tOmf=3@HNLczT} z>_%J&0;VtR&r6`Hc`^4u>`N|k;fQ-Z{jkw)-(5VsAI!4&-{-(tehb9 z;PrN09-zK}Jf)Tf^A_MYcU_#}6eIZ#QJkcEEFxa$g7W?SU(=AYkyiFjI}e94(Wb-7 zIG9@cy6Y{1$WA}9`Os7h!(5l)V$%lDc)z zu$W(TXL|<%qbp&h1alowaLQkTZTz zWYUg_gMI>=SWcslqz1TrZmdG877ZPg+JAPobk?ExYz5-y!`2J1@}KD`g8554#?`lr zo>u@+B8v$9g?`m*Uk!2%MN)_qs`v}VLfoV;%)(ffnA8*A>(|xevzvp0`zXk6%trp6 z1(t6WMX7XK)sJj~zEcAjLHU!~LZ>xQ54ic^Hv?jaqD6*FZZBS@xU7%jjq5G(VTvjl z#<#bMv;kC?S_x8Hi$YZBi`sbBw!m)##lzj2IP4n+_f_Q@(QWaNEfb)Ij9F*Td3WSM zd@lpnkL^F3W#s3~)?=KdFt(hdf=~Kv)MJhMhs{B>Tt(7s!XzSyc>K9Al*ZnZ(dzp? zpq+~#cPG85`E{*iy7;@z(aPd?}9zjI*K^r>|263I@s#$FNB zX?8Fr$nB-t_a&QdDYbF#=ojS-Fn&ae^~w!b6J_hHTlK0{&xl{@0F%Eq@FZKf^2HV+ z<=MSW+J30Isx zO@BKJh zc}jlTge-uJ+e_`i6TodO@BlhSOAKWE_g`JT8U7qDo_WUFTMuqLAw5xUy7LR>UR1rB zY2nqc81hJhdO{mTSME{mFZO<;@2~vm6m{dqO0Ve+jz;CpiWc`8CbJ#X8?Q$8t|Hoz zQ=TH3VCELuj~p%@xeZncSi55ZRFdE`%3I<|=Q*Y#>zopZW+6HXFzou+El7QPOF7lr zP~$Qb!D+pu!F6#0fT&BT4VODrnlE7+jnt1lZL+#p#dbdvS=kH|*#B8F{DoZ-eh^s15FR_W%S&;7+SnKq-B)%sXYiunkuk&i2TkF z&hvN(vaf=a(MJCrw9?%^^5E2X2v>~!W1AD5C`v4w&G@kut2w@dHL(dNI`tPfI)sU1 zD9b}DW1met?ee@GTzdoVMZ+_I4Tz|we+A@azfW$!Beis>Wu2u>uE!b$yL8fwWC|2N z`2{q}BpjIzM86^M_0K212w0x%i&oZey2a^GU+~K~!FX>=a93!=^nDJgea=PF5igWJ z`-c&^u#av43w2@y)4FGpOTvj?W<&l7*3S^nS2S!TUaakt*6jVY*0Tptfs$<6zwA+q z%{SH$8YQBz)_Lgvao?ZFapta%1SW9s{alC-<7s4W8)zHh=I7yWI-r%t5yv!gN$18Y3z9Vwx~#jWujh=hz?zD?;@3vv_IX#$;Xy zM~7~F;`ti5PQ}}}{f+vu@vQYy=8aW^(n-HE^7z`5tC((-g%op(IxMXmuX)Bq?kY>i zbB2h6OkBn!FQI2X8QX;p#a7mA_e&+0EZE)HZP*RMiJkFi;;l@#rRwMdn5BT+$+Bv| zh!h~RxOlhV`trFr#u6N><3VScDB+qJaB`Wd{802mpcr|QLEnpa{Kz0g ziW2HI@6CmB>d_tJfiKsKt`zn@5mxqMizEi#I5cFPJH<*i!W|2SblQV&gExkA8S7g6 z(1z()Y!&wBxLkOGb&5V%Dv)+kJ-%L@9#CetIcbBiIg^WVtRq6~yk?vI=Sl2vjCh>X zLbWaX_1?e~k+VezR|Ij1KDH-HXV3@EppQy=b2!~1IL|n&6Ic5TPR)Iwqu>43M7#u9 zt7oD8Mcp}Lrh>@0Y$%Dg!ED7@dWL{oi$VvD%lJk5PT z5>|;`O!N9Ef%W<_;ZQV=mzjr}jg{p-s85{(`TrLaETrcAI9t!hSO zCf?unXR>KN)#;Iq4J)>)&u~iomD0}l@qj!w{!h(?)bA+WeUO*50qQh`S_rOXIh4V3tfY2y9~CX1+2b!EMS>;l z#q6%B_tsKa9lM)LjHP-sE#krQqhI^@KNl3W;V_AZ=8#Dl6zfieFN0?7SOUwgr7JYZ^Kfz z-5C^grets}-?Ky`qbP$%gAuXIm(mDLCi|aL`hqu2W~x1;ktwS29dH#tEs3E$7Uf=IfQ$KzFp-s-1eQHxh%F zG-7w-Sd3R?StC~ec8V9IY-hfl z{6a^Euu=yRn2jRx1LzK1JXjiv)4=lnu;gbK4Ui5F;_D#lzlXW4mrq1ub3;m%pO}-c zXD-yLL7RDZDr3k;y|TW zo?kC3H8v4Qk3TYDjIKnj2`dw@!oq|N{y!hL6xX$!S;8)eoi;MooDaowSJqIUe1hIc4FpEd*0^>m>gwZ)D>e0{B1hwP~<3CUSt{?npd;r<~|RB;~2euq93m zWER=L8ixUE{2hm!R3sMApjGXMFr8eccGIKxP{nC52bx7Z z!|6_Qn_Oi2i9w+?^8NZ?HpnVf1BlE~4D{>|3o%MLMwz%mVl4aeLop0xqP2UwV2bcD zqf)#;Cv)bL8zV@~m6*t~piljb55e&$17+Y3RWQ!&V+}v>Lu_+EMZWOVm~+UDqYJ^f zfIb@iCkvH~Zt-;E4H<2T;$qjL1VgZNN9}0~3tcOvB=xE0xSYk3mHajXSOfrdlCX$X z3lX-D3n{h@Y2SsgT(#Cvf^)!&H0hb)j;6~WT1=>(7X9yP}8Os94U>>r~Voz!`do_&6Lh?PKf z+1kc|@>6IKoh02<%m7^<-`qZ0>F23@CT}Znqc7;m^}gIM;|s#$8hcxo3J{I6I*~x6k z^_aVn11vfB7{vJ(UUNl%ZL)P;xO>3bC^>}vkh5!j7Kc|uN3yK@2L`fFK9&q5gJBoD z$XlD3#%_bxs21B8bP#=)OfijL2q7+>?+R#{<3hr*{*f*dmAy=3;%0lUY*l~yfuLB; zL);DrFUc3=m@4ZD>*hy2vf8 z_W;zdR3GhN8^~I;S*N((8jRo!sR7_1ujetUtTRqhF9$TX(0;S&(_fxw4OXKy4xG7X z+COpW`uwp^GyfP}Zgi(Vz_IT2Bk~r9-)fGQ(%DRW--Q{^*5zTI43(1B)kqkjHXT4E z0N+(cUow(sCjyAvfT#?C9^JzhFm$m+>Sh*6VJ2qm7NR^&{J&`9KT5Sf+^)Pd%}r47 zj>zeqp+vtOH;b^JGXBYDy0;Dc!VNPIpJo?2GFBekdyJ`WrZFAmzIYjZuz!95C3pv< z6|He2mZueSu&oOrGq|EpDc6?_;;yM>nM7nbTgH%Jezc(0U$Tx4$IeCDhD(QA?aIXNbK6<2c% zmez&sF$G6Q1hsS=GAydN&n&$0 zVV>$KF+X^dDp{g#ZWKsmxjbRk4O>V8s9$cdvO-*UBE-&TwQ`R894k>3^n$k7}95B*vV#gY#tVpFDJ% zIFFfPzbk%#?ANXHw6Au{m1XwU;j$Vl!sx2$&xyec7`FVXX9X9q%d7|{@6tk)#n1wc z;de>OtTsNNM|J7p7k`(cZeBIsi@o&ocEa@Og_pMoJN;{?+cbp+zm*#O0D<)2*t4u2 zOqM&5DI~jk`_!bOa}7}@|9@J77-+*HWOnSb-Z&MjkF(s%oAA306hwCZvBhIocG+SC zrHfm+lmGobpt8A0D4`AiqJ*F??vwzDI!;2^{51+eHgCoepTqwl*JRJ|jrKUT%qec_ zhcEtLuXdou>f=dZ&!Vy!jb2{~mIu(7Qj5mWt-BuV4^QuH?9UR=p1&9JWzC`keL|$8 zJ9KxEoR%j9d*&3c`PM1_`{tlk{67FY5U2*!Y_ZOfx$xH~e{zz$gngXV4caIvD8MKS z_cYV#95qm`T~Iip@`PnB%v|q}*i?q!)&q-swHJgK)^n~U=}D=o*MbM<#Roi|kvqk< z8N++CVi4&gOn$0Q@L>raKfds#MOr!z{QXTk4gL>jgq;=hS5ctbG~o&$+M#~TIX5#3 z7SYBP`Fph^a~@ie)ZC0YdjTuY2JD4}Pd#LTV4ys2m}R2>{NmG`fj)l2AOG37CZh5R zY0rIvk1JQ-pi<~BFeB8=e4YofrAQODcsyc>yWU@I`s$XE!$1zN+KP6E;%BFot5p95 ziwg%7-+mNyiiGQH6902!1fulERAL;cQG&q%Emig|oAs_GQ>wi2sdv~=tx10miBAO= z^5U!1bAg-0B)Y}ygheFPI3!vTfXg(R$STm`+9kt~bmqf1#Gk+}zp8BI8$wi=^&`d! zSEb4CJVG@i;g*31`Ekr^07!8#gN=r;I@lw1q0*w}Zd6Cse4cmm2$wp@bv*LR%v?ZM z%9dQz75_@*zuTBm*$*8g#jX6r0C%R;Z>RU45Re7?$N25}SDs}R2ocH}|10zIciF9* zQ-CZz#GyK_n+?8issgHvFW{|uCkaZafRv4g+-y{S-|>ZCX4d5EB1-jdj>?1# z8s>;S?sHURyvjWdFQ5L;%vovZm-#|~8Q>CLnJc`nue%YKFkNpFK);<1`!fH^MA*g$ zsg7o7bFlG?UN zk(I;n06Q{*UwLn!d=GRgjj0d6@!rT>94j0>djttVe!XKjlRWYC#{$X02?9x!4(|FO z)SD^u;-p5v|Mej}1}g#ubt{{j-S9dWCaYgwok$Zq6Ll?6ara?OXkz;UKG?4ilb*sp z;^8X<^oa+$`of?pVgYd~9kSbrASa&v<&>J%idO`PaM0U*2WSzb@6rgfaJwc7X3rY2 z^!Dof0vpRr5CRy0_E4sg@=`P^cF>huI05x?#MTCfeeziD3$WQ!gL&@apYs;Ix6G^U zmn@nR%L+^@T7kDK(<`^#=2>{NG<+Ztft8+~=d}M$FFVP2>5;xL+i#blgW`|eu+mPn zOgRtM5%!#n#%1kDp;cbZioGBxg$^X$wvUgCEQWd3vz^wTeWZ{7njMF#AXq6-fr(pK z>@=My+O1O~Ktg?D+pCWF>%m}Ml`vKj>*;0~QoeBQMKb8%yMVKwh!CfCNHRYMF{8S#T7B-!b>I?Tis+HbdaUlbG z<}#0u4<4z}i<*;K&hz?RvrtFh0UL9L_So(_AOfVZl_%#CB_%J{2bg(!Sy9v*oCmZW z^+s~nNdxGLbBrMm8 zD{m9Qds*kPoaYNK`^pPfVcJZwRUO8-!a1dApD}eZL9>)aQi<>(DXl{7g?3R)2C_;j)zo zW}}&P(N}&i3-dze-LE^aSgqjdZN+QPN4FDEfkL#xo1vL(q_lQ3)t_?TCEmbV!e&j4C$3zCH(jU~m| z`iE!DOQU}gPKC3JOD@vsEe&q~*n>)a9e&Y9XNYAcNM9~{S_7KfM7hRY^(P z_dn7rmVeI*=q*A3A#e(CL3aHfxcK@=h%P3LyJCYf5Gp4K8fEa9;@lbt1N#Ygv2U#IH z(($y?3zZLylWr*I*fd_AppUk4XQ&G~hc+(&&2^~F0E}+zb~cyXphp7X=wj7PcUgE1 zat183emaqCw$?ekm;O|36u$`QJI+nGxZp}lYjAfwdYzB<`2q38I zinklQ53ebUAt<6KVwe;4oX)MU#&_Gy$T-fQ4*MNdGTXG=MnOSZnRCOGgemp&9=wIO z97Bj@B|@Y?xR_nxVir<@gFdJ0sH9dqgNDSdJy%yAo<*BrVUfmbpX~rS{DI??m||3> z;8%8Lfh(C;Dbw*iS=@-|do$nagI8aZD?Qr-W0*1}ioo`7(gO{I{ zS(QUfK2x=w7`U+jrp8ggFSnpZ9nJp()<&o!Z!bPY`16nyqaf4)S!J`nh?n8wX`Ilh z(PiB8(myY-qTknq^F+U}^ODQ~nNS9uW+s}F-2Go$+yDQzm<5DIhSNR&?8W+9baQl* z*h>l;IBXDV+snC83EY;%0rEJOUaVphP14>f{IR?>^f*$qj2Q< zhz9|b*TKn2@H9apaIkv~dS29i3y!~uw2!I+diZWyF<@MHeC4*QM;`PcDIW_F)IlB} zv@Wx2V@F5Ifm7OH+0I`sC0G@l*(sWfbcV&+34hye48RgFKojKp$eFa&JGnuBfR>R0B*s!Ll) z#bo8b48FGgE*;*$4DMSR%V&yD1Z`b&DT{J1>Ta}HcaNGAvzMuUh>@8^H$5lPWBl3s zDxua-SH7y5;w>r*nXJfCs>z9m;F3~I+jq-``!zV`+(*`ap!WNH7TzCs1TROwTgH;> zhmMT-|wx}$M>->kPqsj;~nlW1cWGp#s59Q{ z9(sX~;#gIJQ25H$AKf?quelfN5Y- zN`i%5@I4!oFVPB7CP&w2yu0Cl$>l-FC`(ihpr@VKYj~x4n3N{M;^Hr*BW_Z9j9qAu zp2kP%3?73-_dbW#32hi^xW6uaf51Wz(aQMV2_^p92kh$#{fwh2-fiEbNt)t z+faA%J><3JIH1yA5Ex1L{eErHpHAXf^k2M9=io8q%~nHM{5isCERjv@k=~s%1t3wK zKt)5lo!CF+(CkrB`D=|W>RdpBC2M*3mHPN#@mHc$+a1*Ch)*^}Fp0`mGY=HHmK33f z`tJ5rj8LjRARWm7$PTU>Lo3c*G+!K1*BD3;L$@ewSMJ|p!g|fFVrLPf$ox@G<4@uX z%Ix?;cYCLm~(!sgaW_faH%tkI~J{=scI>pv0tL= zF)JCtqXO8Fr_Y>pu@q<8cwXPlj3bsg4fJ9~Zi8Atr$H8)K_IzrP{FgZsGTlNyg_H+ zU@7jq{T~DXtrqRdZWkyo9YG}*s^Y9KCo4zK{5_ftXem_4; z1bGru?=PtcKln4;(ylbDOba*>9seCh8u5!@$A5qc3G!}FmTg+|AL+@~+|^CrMr;1T ziNQr#(0-@rKwUP5ZGF`piOW}A;$Qv-mh^${$*O-FlrMnh5uNS;wQr14+ZGknOEny{ zP8wm1KTGC|=G@FMYM0w+?Jh;f%tU3QJ)I`y*T3_>p4LYT=m*4a&URNlUm|ZPX& zr2|%VbE$;!9?_MO@<7T&jVJ00N!Kbn&`?G{*!Es+s&aKqL?3+j++|WxVH%bA%mK7y zvg0r1ey&{GQT+ZzM>#f(wovu*$?^A2Fl^xHE0feg-W*u(>d@M~1z?47XA(9=y*SA% z`N7mxP*hMcOW=?hgPU#V$KiUGo^eFit;yl{cOtCYX;QDJEi<|$-~0LcUQ-5ts~q*)k>R95@+Q*>D(uGk&1240$;)$iR?Fegy5Ayi5a zvZWZ9>`O~%r-QCkFv*Jaf=0#Aj=kbzN6`gvlrKFw%r|P^Z_{V`e|{_c(q!_uSgg4L z#x9sNKLgeqKfDVB*eDP_Sm}|8rv=(P|2pxImdLlg=zEE%g7|WrNBGCRuGmu;_<_)k zi`uB82E6XECQZ}t|NX#^_vYHDlgS~*Q6Ay?!~p^Hk2c=rIYC$lkF}%_pWY$T>irUZ zp=1~TxjgB?r`ZB*WJoZ18OZmJI;6G3NA2d z5HLzH3S%T_IUH;Fl#4o*cwynx9ewT5-`72pc*wx&i~`m!vZqGYpSJ;j4(*|Kd1;m3 z;2Q{S)yD(0V!(F*izmqpA2Yjx$QbU>MET@$$seAt~z7xRo?UM(2{_9l)Hyg@G{;Ee~SifmdyNIO?XW5VsG)<;I zv*RMW)WGsF^47;D$mim_5ggQjPkVT$uAOywa*nSy`9{~KWTNCPH{J6$+wJcduKu>* zJ2w9D%v9ddzJZJ0Tc4*gEq)u6BU@dGO0e5Hdp%JkydHNGK%Eqz61xeQ3nXz8&E9!; z45jR=@6VFleh^RrQx^Dy=!Y^T=}I^*4!?O#rf>uKpw)^gyg4iOW;?ZAW@$n{V|-aPZxl7c^bItp za$Jl#7J8qfes5f&i%k+4zk~_o9EO3o`@Q^G1u{Zx_-MWSRrMM8;q#7q!b@)Z29)|* z;0x{AoH@37)0T%SIJqut{It1{rP}hFYf9&mpgNXz>@&=|3>d&tsBA_X(mgunqfZB8 zum`RS(x3i@zaiRV!b*|GhkV-lL;2u@Aru@8nTKf%D>0F`Izp#Jn8`m+n~bcK!{8;E zoBhiOzub)FV>Qvj?3JP!K zzs4eI0TnKbBksS&cZ${)k5Rr{yIWc_Q1NGA8W~wxY|A*8S-2tOVv3Z-o$OO^-j+V| zgs#y2^4d{}qS=m?B<;F4FU#dTQ?@#LkX;PzjDC4{xE|>ecI;_SrinceBMz3jZInzB``EH}3!II52DfYx}PR)!}$Fu>SI{OwbV^I@tt{4nUyqylPyglcl2A{Yz!8 z*c9Y8OoQlQkNf8UY2k35C{65&p(lrQxz7PD`c)pD0-9*^aH!Ggi{MyG7&*NbXr{(~ z?-;hkidA=ur0$M3Jc~l7lhF0oy8e1J$hK>2PDVQsgPV)>yc|+hUsgz=O;4&^ zqfY1S&TVMrjID)cSdK$`j#vBg-FJayH|p2@K76sWl>T+a|Zy&;#MsL8k1Ynzt>+B0h*dL3n+J=zz@7Y zYF|IFfW|u&;9;4=G8HN%f8dXN9di7HQ;FtDImY+x;GXE0HT#r@3yhAxP^ZLq>ZHq? zcmL06`jvU67Qk>!hfk(!I7}(FHzwEqD&w>pks)NKTC+txplGQ877iJO3Q}+aUGOykN6i2()Tcqk)-9f5gjgs-22wDtMDD!V8FjzHeb?c7FXDPKTNHHfA?+ufUqW0BR7 zm2`*Zg?a!5ljd53G1rxc=Yz$2=?Hsp3VNrjj&7ynW5A&f`v=Nl+&6s=o2?b0@MGur4KfRKx$^K?9>Q&|`%U zmddVV*v9p}+0Axu5p%EE7;8P^9sL-mqz9O z3%@l$`LpW|7W){gsYR#U8u)l7t`oV=#ED^+tkI4?0@QI%r;w_T0yAFtX-vU@lKRt6 z9i`%MaV<%fN1Z79+~(g^1)e7B2YRkj$;U?8mz1S}T;s?}Ljc2fc)LK$R4>wHC40-tpi z`@szBQu72sNr)8svmfy&8Ec}Vt)k2gqdm@2;f*AH@=_;X!Ms0R#J$uoKky)mLuLQB zClFjCrW7~+fj_`kU@q`*=sOEOv&<^DP~PNtDlqt#_cQXKW%I=!0I!#rGAcD0O<+I3 zTWr6A7QpL?OF5d|IN^F8o}K!>ov9VDjiH%A)QxVWfrd~}9cYV_yFaB{;w1a&Ldv$4 z;ewAB#)Xf-3yD(;w`CRmd(Iv1^m~xtdeFpwFBYPnJS-a<1`$X=TjA;Re=^3juc!E% z9OylK=5+JDm;(C0FvlLZ)F67fj!@or@aS}2<-)0i_$yL(@CNWR@le+*r7*G!hh4e0 zoZ=RR!1cR&4L68TSJ~O`Rv6F-(Ih&ls~`o)B58iyCQ21#C)027JoR7`FF3M-+O$X7Jw<;@bd$c~deYFGHfa^fK6w^WS zm563m5~TmSHEv7N!Wb<0J5Kk8NBNq}TjcBj@%r;8`6kmmNg#p5CyDXqf>+#KVoKK1 zJQ0&qS^Am2Y^m}*>hq41+X%a6|D&H!rzwyiK6a&5nSt4G?)w2Uv2pmj@+%l8)TAhIUNs9(e>`Oo zNeTme$mnSKkkVsxdT_rBYcV>fsX5<}$gu8Iceq~04?{apn-slzFD#V*blVmtsA+Bh zsiUXm^Ir_i8VZ;>+2s<{+2+LE%U(G$JNfs<#W!?LHP^8gy0UsFCbL@$gvXh zvV4xef0uIx89E<({_0LH)04a$E(?f9CFOOn=PmjfGH%?IeOoZWJs3ksWRD}L&y}oE zx`%Rf;yo*^C4omS{r ztQjgBq00y0amo(V`3@{mPp1AS#X`g zmvD+$GYvf%*IKWoLPCd;Kci0-G8NG&%jk+n>r-RyC?HVq;T6q4!|O}$u95VU%Tr?( z=k!tECU;IFxoU`Q6VNk{-@NQgK7HxiSE|`^9L0-cho~)4w3Y;6e8jqAIK^j%!R*(2 zsXdAJcZ~TIU?pJWl0^+7oOo#w*!TE2_ty3OKS&wlS2~PGbC_c5{iQB{emMaed-l1w zr>G5}WRtjd60FhocC1J{J;sfx$wbIr=b?{hL=&S0(?ROqNy4#8rWe)}Z1`q>+#C^2 z0Tkulv>$F*YN7|A=_i_qd@s3S#t~{5S$U?EGMa-El~oZgUVnGD=`aqXh!l=20lm@N z56`O(MWf?`$-=CYE@$^CFq-e)c+h1?d~-z z-64|gr0EFhpjnYyl;&(Dp`=}mq|x^1-s^*b%YiYS-dTovt;8;abEE&p3XQkP~eo zff(t0EeviZuLYbOTU|2D5n93!X5V}ihel@z`)5p-I9 zc;WqPiKiVBHfP`BRIr$&^a>cg6zo;flE!=}!?|Cz-|F_CL4EGq55GiO((c&qDCb|W zELa|2S-y4hW{ytd?A!UW*E-DyySxItXm`bnJd7f zJ;%B(S#|szZ3vW2xms6Oshsj9obUA+vn%VUKj$tEhL#3m@~vXG5i#Nfv=r~w-IVJ+ zn;dCMdQ-2ak37o`j^6Rs6i^SqFDx-w*arCcDt=*GFa#1KH3@G2m8&NfW)?YCuHRMl zFB?p1&AXXD4AAG~-~ayE*vLJ=~g3}gcuThveFVKGc&3;3uWCxQLJ5qmumKh1SkRmX7i&;+f%lVh;;z^5xBLm$MQb@K+ zzo{p(sL*uVT_H~hpjM$!=xemQLb0zh^+>ev>SO=JZ{h-$@xA%)YSuN`dFBs&=ovyx zxU67;6J;4klZcn1F!@H8#E%eoRGJ4^ZZa9ZzHR-O)2l%IrK(T;e-!R<8JY4qtl%K6uW1BKj6gU_0+aV>_agiGUhY>C;u(s>U9MT3@BY zKqOo^e{j%J@vj(pn+I)o_*uWmIt*p49^NpOia$C*oX6aaN!>)f4YYmpO<*-IzR1}o zNCJiwC3tWgf)a&kW}(Z=AaEA=D@aPTN1Rvtx9}g8Q7LKV;4{B3voB@A+jE6BhLvx` zna+8VDw%F~ga_;6QOI8{D6nid15rDLjYrh{K5VKs{utaQmBD%A3bp@7M? zBGL^^PrV<_c+8U5u+`?HG=yNyjjoBxB)8E#+|ZQ@KJKq4&FU>Jz)q=w!8Onolt7=G{Jb)y-1Bz)x2D9*_l^Lrm5?J%{fc4>OyC5Bu)qec z`@~1eE#TPuHRq^BFH0rMb2KMv{j2L+ZS*pTqCvuL zN(#T8(&tyrH?V7<`nmUa0)^pboMAH(Udg>sA(c+y$8~+RYB0jQ~c;9)xo)K3HHkVyHWFL1yZ|HdNoI01}dZqIIk_{yK#v&LaFV6k?4GRn1fT`iW#+(|wIStn zmmPcAL_1A(6M`eBW6mV;bkbrBslL0Dmh}QZ-SSR9-fIHSk-L_*Ns$Bv zjSjg4e51s`=P!PC7ho6%U@*1W_Ng+NPj{o`^k6l{IsRDEWeWFf3{)qZuP88#CREc4 z$s~n zuCDJ-Mr+?aM|V(1J|qD4iPE$p_&k%Od!S60194D-7IH)B!JZ=_{dw|@WhM})Yd6w` zZ-Hmi*ehR%9gLFiY)qCzdSg4Robx#S%S$n)+%0YLL@8(czQJG=V?=Zxn0qb}6ZyW4 zW0o#_2BV9oP(wd*v44G2*k-^uz_)cU7A=HYpR|4Z=TzN_((tvLDz~v4HKxF97JqLO zY|?TaxNTkm3y}lB$XKtAInD#Wz3yNcsCuCS|HVup&V&E1NPE5n@8w8YTYgirW*Sx)ZD3veIr{548v&f^ zxZne`o4fQh7?h7|!gt&`&5(*JbE4nD>5q@Ql7VRh&XO;eiFiq-@*a`ingX*Fs0c_} ztL%rESjH9#?p4}^1^zu<+m|Jnok1{!#2`A~2RAIo$*zC_u0dib>TDpLufRW#E;e{6 zq?3YDTq@9-IJatdxCs(VP=v9=vDun>$< z={OCnaX4;8MM}WV1Kd(?U50~!09hk^oMQuN+B@NO{Y#PxbT z(BCYGVAbaclZHjF$yNW16th3dyz8)pALSLW$XQ4@pKwMjgMc9NhYR-M_yr7a}jstq%;AhH)bKT0_qVW&?_Gbkv%2myv9F-yw~ z`MFW0{B!gYa{ZfEwl+GK)ORn=oLqZHju!_DJ*ZNAPrQ|9!Ru+{_jK9S6N;~pApzo=oJtG56v+y~$SA310|DJbur8x(C1f>I8QbQ!*G_U2lWB@yxE~Zqo1y0QXNL*O zFU1J4$5Bplbt{#46c})q#sjeb72;I$6@mik68F|ZI4}B)1ANY|fTBzkHz_GyaqB6p zUp${zAh$aC=FZP3McUICUno9z1K0jD-GOL?QSRE}J^T%Xfs=YStGfMwYG-%Ry5l|& z7hgT1k>>D|yN}Y+Q9>;l0wUO@M*Fn9NzKn|vV{O`d8ENmu~eYsvY%rwdw@A%wZ1+K z9yz=B>jIY>q4#C#foBh9BklzAGwj^w)na`6@=or(x}M%2T$2!pxav5Mo_ftst0$;a zf)DroUAiFM-q{9Rm|iTF8Xj>)WF2S3F622_@_I|FffzKq?}M2e1va14MhJUuh=bbr zF-(jHr*vxwJ3A(h9hpoltlTl|$StaIfm6k!D_|wc=^aUYyF`>JYv`}{cQa9}k}f8p z*IsVNFmMJ#NM>SLhG$$!&HEbSpEx+=7+yHSmN=+3r&PgV^fO1QMe-?aS(dLQlocS!zJR@D+n$*43C~=~!?VMizfvyoTF!S=IiOfOuoD|M~*&jSNKz zRmh|_;PJ7IctZONRc|3dD`~Z^0brivoK0~ez@*H|S=}C|Cr>^?)$_80_q#A`PWS6q z=kKn!9PY<`+-Kkhm7(!B`LlI$Z?JE6V9_D0d7MgRppV|UNzi45QtYc~pSrBqC!s95 zsHlXl&*qBFDC%)m6{aE7&d@Io4dI@JZq8PUSWf`+*kY@flsa?DR> z89^`*a3z=_;5@jDMXw}Z>WxlD#NJ1j*tK+9C6=)nd}s87nqh?sa~6uiCE;>1F(Ob@8->M%^)8-J{GLwBw1v}%_h8Gcf5rsmDp1FP8=ykV(Q5eJ zhOOf<<2p6f*1O%EAa3`$CZFWx9=F(jQyyeKy`3ahroYI4OQ zU;KsB%*F#hTyjm!d=6w9bYOzg%V-Qm zhBtI4z&{UXD`U~0`9zC%cdOh-egv}wP4nQ&4mfdhBtyUwg7ztrUx$JtjiDFdr}rkK zxFW*mRGT|jM1ve=EqAZ1|+f*pe76b~AFlaZ~S@Y&)ZB}IMfaD!%YdK5` z6R$eK{t9-UW1;uMV>k^FWz)7g0n;yUIq^FPw!-22-@cwPAo4Cv5u&1Z&V1wx2WI2! z!6rmSiV(EIqCM#BZC*SyOkM05#4p}|AyQ^ACwC;QAN2wPj_kMXO zW=R_@eU?aweU0V3nQj%^t!HIg;)fPKRe~DpkHqFD zk_FGHR((%%I#@`9nlwJR!vsI8yp#XeO4^ZT1V;<$IuC(520`uX49uPa96)h48#k#D z5=sv@d>I?<;Tr%U@0Jd9UjPYdq~W3Ft6Wg9$;aH}ruk2IKMT{NS?u$|Tb`=qVxkOO zof>;RwJ1UATzN%qJ}UAN?3p>t$ygLt^6`s44$~c(cWw3EoBtwuBCdZ2RqUviTi;Nj zgjEE~YA>lfH_4^4I3fe$JZ90?%+0Mq@t_u0NlyB9cH2RCd-kQWD&+~&OkN~*Z^Hzt*ZM(lHqPs-D5)IWhi%tv=dUo3TD zWX=r;d%42_l)^IXN44dR!G&3c)osFlb}Lntb^Jxcq>FYeB1+_M(lpjB&wn4)+~ZWU zrc}CVDk7Lg2Y%O(m)4J-j+i0Nx=6l>G+S30t@Y{@rT4M6fP-!3bz=w|3Pe2kk$Wf8 z^J;AfZqX19VS*ZQ27;q#ZL~RyDWxF{%Dm<>!bH zrkgMNz4+>OHf%>*z<=iTI&qjnub-`D2}9DM+VlxXWmHtv2LzQO&(;3xfX>UZrJuwq zwz{(&-K`okaTM8IKP4uNM58NRltv3xk083R?}F^%-wL(*H`$NmvGCze|0cNc)Y%oi)XGw1jj?^BKcd>)#@XejJ{fft$>t%j~m;SAa;E43*R%lx16eh(x-@= z%b<>H27lYP54HMwK*g<{CZtEncg)RM3_2$NTU5eY>-;us`y*8(z@GIb4RFIT(aH9~ zckuNfW6V2>h;CBNildC*=VJh{8W|hu1_NuTbk-8MUCKN}hhzhtmdyGuGtD}b`znZq_VXoPS`T?2&~`n{4Hj!DM-M!i0xKBG((0o0hm z-k#SKO^v}e6Y`&0O3y%XD}v|DF-m9A!fNS>-5`{xwFGzWe*IF8|VRM$aO*d`P+Y}?+UW$$_* zxaMM2pdQA_<^(Mhc4P5m_+^ol|dAm1{T98cNC_l>_X-~s;~ z9T8SZTEF~P87-#Q2AiJ(X3u!)O_l4c1 z+GihT@NiJ#+}p{PT{(8;BiuQ&kxdU4O!9z5&oL77C5kFJWjx@$!KY6LE(FR6;`(yo z_(8d*8GJzxZJ||d-kji8rbAw0mV~nwC^rkgUX!=bYGRFo$wZTQQ!a7|rOY+p{eJ`6 zVAF=d+oaAjM`C)qu!iFy>fsFM$=DTPUR%!zCy5?g5asY zt{%sn_1{*WVyn8!ucYwZ};`-+4hF%5ihRR{acx;(8+xFA;7fCuyPNaXA#`aus${PzK0;N z0}Z#`OJIMRTW6#6fNIQ_|JUZE?$>?IF8^O!Z_XF3uf+KUpd-feww>WkD25Fx%q*za zT)|O-X0ZEc>3Ke4`93_eS*qiy=1$GSRb z9=+{JZ6Uhk!L)1H!%kzuJze~k8y45oryR6W@}#}gj$eO{vz2pv7-3t*23=6o50xg3 zk&Mj3CKdM2Q97O^r3`JkdYZZ1cpsm!O)DHyfPd#|gV*?gWY|lM zQ7j7ME9n1oodcBY#^tJY6JRvT0#_}O|NTFQ^y7m>%vrUG2whC|R^Prq_qg|+oo)Z{ zJ)YyknnQ6|r4l-QL%u+<@#jFJLDTIKICzS*%(7 z7%>1?`0uKRBPp5;{={dqhF}v4u|I%Z=^6m9{mfz^ai!;4vOl*Y*q(lJWd-rl+ZQ#H z7H>&kMDlgqzUiPsTWtt{)y5Tt73)D4nZg(qr{P&k0YhlF9u$vBChB4BWI+Yj>1aa% z22}?SzckHpUy4tJfcBxHJ#6E1Wk33{6g-f*ETTF3RNKXD!A*g~1O_ zwLbtLGb4^?F)RejLTT!=TiQP`vER6_%6`dP=i2`SOwYz?`%TG1y;+J&m3dg{s;`Gu$R-FwWTr?rrv*Re|ph$>c6MFM0?B)yk->SaM&Wd|7 zi?{NWVli^RpQ%&W?kTRbL98WcSihyfxYfYRBzk=?z&7 zwKeU1os?2)L`#K!#Gz9cfS8Z(^X8mnd*?WAg5d>|eV@V>bOr)^ob;c9IZMq32WQM$ z5MYMfczGp05U|tmb(KtJ_X;Jo;(T6iL5?&3naY_z_OtqI&%eeri4-&$DSg~qvG8u} zb^AXVmtHzq-29%2i!u%uK?`0*X{qaxwmfFBR$k)i;V~u~x#S%DNiE_s%TID9vCakx z!!7|xuMlfY(3LLcYBLxvLw@5RTjZfoLU0MFSpB<^EP36y{|Y*-{zIt_)7St=rSR4r zQ#)PO9oGPV+@DmX=en7O9n3%pqx}QG$#_DNg)-t}eRr*UKey--OM37OjRv|X36x>! z%+h1^#535XhNt?ke}DT{MxQfjaOV3J3JRFym9c956xZ9UcW&3&`xG@`>IpYo4lmaj z>t^Bfw6=aP5ouQGbi9_ZA>n25c+in@YQIqYA-0&KJb5kkFo!23&jB&wLPy_~JX}rr z*6>Y|f-!SqXD8jS4ssCUgLNULNG)fQOFEio$UYpclU2tpY=9xOU7*q@d#goF1BRs_Agt5gyZK89S zGDptex#k;4fDZ`z*zK4tVx>YsK03;KDvjMwMBMS6#|QIJ^Xhz>*4BFFS>~JFYpXQA z&jaRss{Eqvto-WeM^Nnf+3_X?Z=OKM!Aap?w(knbUy=qhf{P>kK$BebXYe|GWGyBFVvPdl)TcH&$} z;gCdOf}{Ir0@@wDKfA4uPcfzqHjbS{hFw+fUfR?fB|@8jTFYLU6^`Wd5D}=}_-|wh z&dQ7S zjHMY-C?w9QaOyZi-qFd}FYTJW1%z=y7JO-$!CFRw4d4HdOCWUMWA?8>^nf>!@i##% z_^MIlrV=lbE|EqL+q|!;ZYJAJ&$sw+@MgGEB(43WhcOkD_E%8CodA!I2FSgixp|%v z(y(w@dseJuYVn#U@p)j>@(+$d_#q#zxJA7T+4|6Sp)Ex4si>f9Wd%#i?A!BWyyq4mr?*T7c+sXHyS$yLdEX3{k|H;k)Z8M*5zWs(+?+cig(V=}1qd*HN6C5CVGW!`kbX|#oq zV!;cGr@-opuXb(v4v#+BpFJ~Addx~aX9csne6nXG2Zi;jFnInySkZeV6zjm=eavb@ zrl`K-Vnf$1MnKzd{@@J27gq2>;8}oRD{u`wT}b61NH2NR857~`8_GO#*@7NRW9jH? zdh(f*b^d?B4n5^3{jV$uv3o>|exwnvNQ-;P_<7ed9$SBo+-2bXkilOlC{P$hS_z5; zORr@9pU)tNyeP4bzf+VVDUHD$%Ofw513G*19Z%8?OurbQ_k+~3Bd^DORAD?d2y7KA z=z8)Qu(GMW1&BfI9(d2y^505!P$NuKV7f3r#lRm(SjaIy^NT8$MY>@o7Mdqk#8e|l zR_=~*&<4eLT(&*~Co3Tq0kDCdE(#?LUMLLN;+vxvJvRETg;M40d*QvlmcacM&H&!xL6C?;qWki_fi zypC&0bJ=5`h1uYnW>H|SCLfso=XW(}F>WWE9sAM`7s{Dgd*WwWGiC7^Ch#t=SxmPyi08BShZriEP zoNkzY!*^z+@D3`>e>MbWbfHH)g?MbN=wWV!Jhd*p+grfqf0$|3f1vge3~P&dp1$$z zZLqLcV>d#?veeJwxky`Xeb>`9hjV_cemlG%3*G1#x%jKFyL<$VT)4|l@?OkBJ=;f_ zUh_%Q`OUPIVUGoJpEEel?a@A*sZ(zABDa4wRMr*#!A+XcW6WeAq+?lS)#yn)b&O9n zFU>8`7QXmf5!_&lC12n){Na(QX;XDbO9&h)^qut(bc5N+S=1>u@ILt@*l=#&5HmCI>MMG=War$8XnM0e~oH19# zoaI%Awr$QDf)2#apYDbyyJee_e&k=29YNZk3Vm=-oe`)j$G$<8_KRi%CbGtx<^NWY zj#Nbj?MJOQN~c@OsY`UL>1v$R7U1b7fi6u8;6z6=73k$?+N&!d{Ux%|D1wqL4J_~m z*-r$;lf&cr5JHm2w}nbjRpqvM=WU1G?0qN-*4uzmuj zTp52L!0K0;1ta$XRaWlb?dUDQD)`hDC;wcb1<--}UxBH37M};m_92#=2ez_A?f2_E z7EPMU-;MM|w$&Z`4E(MMAsROrIvBXdL#>8B2PUYG@`)5$Bl9qat_VeC((nCfn1hiLbR~2h0G~!`zBvnYu?#i=Wt1bTwHLqxLaD`w zG$^iRhH~v5uTlPX9avO`kQ5bm8yJ}*_Auzzw1bRE{Hrz6kdTNl5i=zp@bKjX*EimR z_ZHr3vxVl_vB^JNQyx6uV=6;2u3(NIT8Z? zj`~32?guo+67%_gcL9*3YzE-0N2BKhe*sx{!b^=E%p)2%P(QupI`Qk6))CZ52*To1tRGQM_;Gr4>B z(p2gd_hHVFT7INQqlOEbfptVC&pop61Yy)dj?|UQe@KF&)utP+i7h&O-}1j@3|fVV z2HE?=Iyf$)tq@fk;elf(D)5YC!&2oz91zS25ZncWkve#BSu`~r*>`d&5Ll=1O&M4D z4Gg}CbH z4+G6(1yvr~SK}d)mo|g)e)#5#D)Bzsm-A!&)O-nUC`6lvSi2WaH18vce=snF;1DW? zXfUg|-Fy|n4+$|XMJ;>CwYcohi|)D4*1iL{U0Zo@7D!ehi#KolE9Iz>DwrC+x_Is2 zXg>VXm6w3z`lyQyX~>)pyOdM@;pquA22Xde1B4BEt9s8>eK+5CNYCY-_*ND4B5FzA z=4>C$Q6*?xRvJn!wmLLNG`_#wbIv*fH^k&$nT7bH;R?*v_b2``Oc5S+R-c-Iwo!C~ z-$z1^l6Dveh(IG+<-+$vH=xX_bnUOie%#98SZckvNA=z$R)(K={ls|T{+IlSF;$=O1VezE)N%PaC6KlWb-LMWy--%`FhP7!wrPJQi zP77b3db>dnKg?P>wt6;npN0slsO+g#H@@PdQakCm|_J{O32vx<1Eq5P` zKW7kA0)8E@T$8jP0q{!;aESkq@z4K!B1erO^?QA9wgAa+%P0QNdtL;=(R*!l zsw4qzWTgnfQD4v5gT7;kWJbRI+qe3FdY$akI)!JMLS&OUW?$6V7)6K<)5$;7uWy67 zjDP92#=qU~ONF?P#>eXbc=5Zq3q)xx8t`jNTX3MOI{&HEYe%BHJ14g}h~+E+8vP{* z#IU~mvbHy#0(i~%n9^&XgE>m~5;$B=lDgh@(o0&-{hbGI1nV3Jaz)j}FyL^pzTbRA zSZ)X&yDI>BxhsO=2nV~tc4;3kO&U(qXMl;&d>72T$1}Q@dq);ZM?Lrf`m*7eNH>1;n=xw5G|Xo80KP3cCm>sk({dg2|*V`hJ8aliaOzx%$SW z>@I_R%t0A1u8?BxvIBltiO56u!URqBp_CNWap2fec&d{^+(X#quQ}*%^LL({wcPa~ zn0M5{6~Kr?q8(ApSDvK*p$ikIU1!y_6(1>cGI>kk{%{^wuI_&4orueMsZzgPlMVRH z`!u5LzK8s#Y$7dIX&~IL__ye7a3w0tKCSM&3QZRU;@##{<+3XIN8+6-R|w4S(OR5) zqa*{-O@e4$20rK8W@86LWY+eOZ&+aZBQhqX4Oh8u53gfQ= ztR5wdfQZKVvR=)lc$A;w3JIQ+NQv5z(Zv%e!OP$AKAuBFHD~|t@7>|NL3PGP?+{xd zPCx6wvH^BYQX(mTia_%`fe!h!0N~cWvKWaHf<)b{}?KuJ$ zsotN13Ro0Iss>&NoPmXaZ*z}xZcGWK%cq-{0%S`M3y6ChBNEX4kFx)vS8}8}OuJN= z0kCDmZ~&z!CA z?#jPDmI@2N6z^YKrbADZ)-m6&Br3Ng*_J{OmQUbriQ)dv6nM8{kHV4qq}K}hRfm`A zg_5+_t@yvce*`KlZIGXfpVsyBg27C|4F`N{3^s!S?^ZJHE#}L05v>Pvd9oM3H_Pz` zvi*Re%+I5pHvVC<+)@O{SC2RMP3k?1kaVDE%r_tB-oE=rUzKEYCHC)6XiWFs7kmIM z!5K@;6?WWZ=9^q3zj6U>JU`X&4Jz&p_oddcx~JgfBcpHGxOV|T)k6@pDOuJ!R`J~J z*z(H7cEFz#l;8T|nc3tjAFZCa{UD;BI&AF+Fq6{X@2pw9IEgHO=^@`?K6X_g=h;#F zzc)}0=?M)cG8{$(k*&L2C2wDn_XokxOVR2T9#>m^o!T4K{> z5Ab%eFuH;%AWkz>(ch7C7|v-x;M*AQuf?5`bOIJ2h41I)q{vp;&#h?#hZVMft;&ko zD&kl)FyO64t=1Dx+de3t}d?3vw}Ytri)fi2v?2`R=R2oB-ETMzgaT=>;DL zK6XE)yHv>30))bctXN*K&<*8*dg$+nIF^)O&KPLXp3BmD>PSn%+QoR^@lP3mx~`5? zkoP^}jlHGrt}(lCDjxk5yc3dTzbWn-XrxnUU- zq7I<3<(U)kk@$?+N86qh#9S_3fLdztScq7LA{?nTjOzvf#)CX?R_ofwKx1SF^1U|} z8&=cqn~9f}y&1n4ydTg4-H8jKUTd?uX4w+hkoo#JW)i6b#yHi?3V|j-!N<&p>!r4` zQf#Lj4pCwmECx5-bhhvuG-D-Fm${f9f&r)}(9iY)tmx-GS=a_}ZTRf!NEpvyeS6qS z>}8CtEMo|gG*IqQzK|u8A3@kcd-Xhs$h#JM2=vBpBK#JB{_nSRncj==r)enhKw)1t zg08iQW8Vop&7I^2cyo??Yus(Q3LK-dV?mvuY6#}mE5M3=Oi2#Joz|E=JzdoYt{Bq( zKa1}Sv(JNbQb^>z=eGZ+{i7~IqO}jDirm{;s=f)_5s(kN=iLx5)LH#heB)%ht25VF zZbd>uPxo$0hx2&=Qk&SCpXYB-EBx9q*sR3E1u^J0!N3Kys8}R!?k3?+P~C6R2fmCo zwIU|YfBO5)WsN20!3E9)BngG^$=3p961-&Xy+9`$EJDQGvqCM7`AtU@0rITUQJC;G z@JrGM$uFfH(SOH*mI+@>5HB+E?fc;#3c56c0M`Is`He(+s2UDWELfr0c?4D!OgQeg z>RAX_3iwjeaPd>&D3doJGd6tegf$7@7peo;3Jm%>K-lsJY2`_p-OsM0^KqY zoI;s>R}`dNE{qAs-^6F$Dd%_)D0G5xBKl^>UTHP-mQ@ek(mf5Zp|+&SXGd&CsdgCJ!rgAxq%Y04viG<_yY*)zQH+OJUYcDl{R2%NyU=D zfn#`#@`=QG-l}a3F2NLpMk|)ZCnvYN507X%vH?>3u#@P*pf9Vf3x&ymO5IDc*KXP_u~~b)T(h(f)k#(ft6r1zhiE52Dn*P zaH=#k1}13Thf|IF4Ib>>V_sB9j97e7V=Y_q@sL`Y+majfB;g zou5E8U@$PLEf#n1h4q>66RvcM(T`SrVoLsw5w6QuadXW&6rxDa!D8xHYXgK`j{wE! z8I4jz#x?ix;zaO(

HW{qxrT8bineLMOXPH|^SGA>0NeF~~*h-pX05ChwJ&lgkW_ zKXLpfp!rKQl>;$=41wZJmcochKR2Y2dIk$M#f~&ajwGWFZ)P^m75gN!weJDvV>*!| zpxwL?pAyBDXN|bliBfe!WUFP2vNa6=y=y~hpFI~frtiD{OcZ$1Cg_lI-zNxDf9*V3 zCWC0V}!Tw+5y@*9pJJBB8a5hQ5Id>)77uY@;)evFDt%Bg9h(xL#h&GY(TYS`xppW~+ zCFhCI`nMs*_;%BaVhBYtG3B5zxeLX`2HU{GHJ<~1i4&_K(B#}@-xorUC+KP*EP)i} zM(G!W1dfCZ4-;U<@N{R-iX+FDt{HH2eoq$#H8o;4oA)}%WfsLjD8*5z0a>oZ$)25F z$1Gqf?$&kQN+T`qY$cu`#8;ldw_tr*XtAq|AD~ zAD6?9pC_m_L3dWdQB^(reWRS}oundF@u{dtx{x0RfV8M*MR51HQcnMkmIqAmUeJeS zh7b$@&dI{Ewr&6M9a%%{cyPU>=-7pz!X&z)39%;xs!39*I~hkj`30bTlcA1XEK&@_ z^SQIi(QznFG7%kqysT;GL2Rz}bNc9M7Zr$_a|ei#Q$Pc{F}tEBmOUgSr9{4{d{;2` z^9FW0@vTa$z^^_0jIv8kDOSV2N4v|<_Jw-<1f!c$;+|(>x%b#@sUHE60&)(6Dby{! zTQ{vWRvb|!oh?`fFhO;sI$2phuUY}bq)|J!9xXG#6sAgVW%>HAoR36TK~$JwncIKw z%!scDBXK4H->N@blpcNi?F^SAg_@Jce zBQQ-L6fVw<3zYG075Xb&LDCOT`6|>G2L5!#lV@8FKOHE8Cxr!TmC^SM4}fn?WE{KP zTj)+nq_VHkJf{z>mR%)uOGjS8)WjFD@{8;+NGD)hc+)@6khTxa{Jhf!K93RKO@zF> z-3B-O+@dfc%ayQ1I+*hm-Ez_TQUJi`!2hQfXxGNpyg4lsk8*8-9!=5!G8C$ET;p0_enB8vNxjWyrQ3RC`nzxjWE@pgfsfghrXhNKR}_pa;*5Z3y4r9LcjE_Wt$qaP+-5v$aYSR&P< z$%g^zZznZpv=34bTtaSJU-H_6nU$b}v5gmsAwTBV?$yiPpTX7EJpXs&)!01&iBL7J zsKySoz7PyKRll4*iSQ$o(j?o$RQ5Try#df9e2n8d#BO zGc8#ZS3)4^W4DX!9EBi-1ycLLjC(up@*w4pSykqVMpgE=Uz~~Z%Rkw%lm^d)Cg;xn z5#WzIF=>9=@+8Rm%K4e)hgS3Q$#=$gCEf15lGX(jch`9Om&6v8~Nrqpgtdi-8*l&ISf#eXXU-4)C`FMW@CqHD!N&dOg?>)I_(Em4Fu)8@}6gV3fIhgdL7Kp`3n zP)iTQo7D+5_kOK=&D9)XCYR%jY0W-IOQ$F-0 zZoKP0sqbV`LgIK_5$%))%dS$30=I*l^AUIHB}Qw<{VlOASNF$#JtklBd>{Jca0h6K zL`l-MpnZu@)W#^dnC$?l2B{J%*Fz4(U^1b0^Wg)0dIv_+)p1!}MUpC=A;C$){OTz= z+LrI1(o&D58CueDw9#yh^+O%FKqHmxv|WFlb!+zAtDeD+6-cs=tl?IU#1~+O4EA-d z4`AGB9p)(D^3nvf;j2eW&wl2()O|Vk_nN6MaEdV<+DHwS=f1ZZcuw`IdqZ<-tXe`8 z)Nysddm+i|g&+!mP4D~cr2^}w!^#rOE0ZuSZU^Mao&{bNXaSL|@40@fKE9D(v1GHJ zp)HJVf0+r&edmk7pZmB4VvT}n5=08Ra_w%Q5!)U7+j}o66G^V1#NeU!Y4W@Dy;&QQ zLNi>7Ixn?Zh@wus%e%%tBTw)1#RVULyb4jMtiooCAcsY->;l>b2`6K0OtSk_*2#UK zFNQmJm!1N#P!u4F=WRFDY~RuL=Z{JkI-Fj1!KIejsp#6`>&-PKQ zGA~@=PmQnff^QO!^Fa%?uMkM}(gN}??9kK*Oc-LjZJWI7*ZnFFMUKQDg5If?`a5{d zE#ILBSD<_<2f{)Rq4Bs~en8!K(7%Y)^H)yaCv^GUghkq;Ew3}Dv{~|Z)dmqVKxNIR z_ZWR?^qcw;iju=FZ6~e+oLgHFIzQKZTBbQQPVSj{X&T-Vnh=E{Ks^vw1lvhthZw{A=?FZBPkh#^{5BD2lCu z_+erhabn1bfHy>N5kIau0P@O|v@AV2GxUBBbe_k-cvaKL6F^Zwkk7NQM3^`@@*KC9 zRMI=&e?uly?9G#4e;bvZA9h%`LKuchAxtpWOMqZ575lYySS3t&pN2}?H43dT%%Pu98>Yd3m2p9>YL7p=zM8}RdmnlBAXE5x|a!1K0+~x2GL?M+<>v)enExdRkg5o`zk%pkj zWfI$urWro%T=H23N#MlfO(EynG>r$*H?(&4AnRr2kwB2?DP0^9Flo~W7kG@FS10s6 zN0Fp!Axxh2hRKa*LB52W$#0-B(JGj)n^TLq#$$Q-fp9;rPZrW;R1|&*^e2$|8T3HP zvSMw4v25%$Wb+i5$B5KJ?n)|j@OLPd8i%TX7z6H%D8nTf>bV5%z%91HyY^oo?C8Zm zI-+a}&CxB3Qz?C^8pJX7 zklB(UeCtWB<@UDZ##1jI-@ZyGTO~!c>zBQ~{RVbAlt3Hv(BB7Qh~mJ)woNsG$S&XdL482hCLbHc$PXb>$ z;IKLn@k|GdWp@Pha|fxFo*~_rwf~UitKA@gVLMOmf+a_<%rws8@;#iy2zOEdmQXp@ z9&&jvEe8f`9AY2IPktyv9c+JNFQQ#~g1Zxc7c|qfz>DBoivOm~EzsHTdj(RvE=@U} zqy5V8z++Hn{EM&I!o&e7WX)X%;crKDwyv+f${mI^%HxQBiih{XpX2f5rk<6_<<9y0 z>z`H@eRx)_%-5|*7KL7`DkD#I2iuMqw7#hGJ=E@$^VUAmo;uie=_>v)@+VF}w~lt9 zs4TF2-v&Z0!)pl0)V3!TXLuvGCb0YAc3_iX`7#YQ1)4BlKcTRV!R3wZzyPsH)pF;S z#n6y5kmchay5|Sf@keDi1FJ0veeWx1@}NtIg-y`peHYGdK(9$uW6Q2FNIDAXz?<*C z27LX98?xm*w2ywS^TLWLYYZq^4+K{DuLxFkpjkvXHd);w-k-KBH>oswUpqq1*h`t3 zYy0@bI?f_~;QCPZo^fdJkt+O}Z0*HS4jreCA1MKh(&0EwkNddv2VDJ;>LP8uU?teD zB{FP?ybmYX;0Uz)4pa_vvjrE*a~s8e8$^aY^~5CCYaprQZE?rG4Za`Ux~L~Hu)32@_RNS7hv$ol-islMGpv5gkO2Zk$uVcC}hi1@r`FtyMO8}tAXAoE~S03Vn! zV^|}EC{qQ|H?$r2@K+wN2ZklH51eb_>vw@pB#>2}vuN=*;O-8bcXwVAOjd=a2)j{n z$^*e|d6Bv-tegkVb$d{zI0Vbfmq+)ooX7%B5sjZW`x- z;eyXFY78 zz1>iHP5Rx8f$vYwm<*5GHYOSk5fet|xdd?Jm_ z+jrW6a=4_x=vP(6x74a74|M0kScr6bM(e7zP+>C8I$S%9qiMCaO)#55{e|c7hay2< zuq3rEGOWH4f>H6^%33@RRvE+=gxY9~c+|RSIZT%v;%v5^1aVT1mu(_e`x+~f6iyb+ z4dLv&I){gT9=(O~yogk04D{5#-|WZZ!45T_S9S zd6ygRon139?gKZg~g~+i`bqtql$bYV|$%O1b4S(?}nXg1LSTLvBDQv7+wgS zxDlnUY=Mw!&?hi-7lxF=o6U}Vv-2o59&!(RIOd=N!XtVo3gMUmaaR!adtI<&c#cSO zk<5?xHmbef5P#!6QF18Ujc=>=vHM+L{{B_80@)@E6$pbR$5_m70Kf4?Hhxv_mOsLf zp5kqhb>i>0Z-8DnGiSl2rI~1Jn`e?uj5b7*+iv!p=;zY;D5e4+@m>=Zoyzn`s6V8nX zkv9e=0UQDaK?u!m$ui5=hbh9sKzrSS#KJ=f%H9Jg#RSLnL~#W24rfp^(Jz3Ir9Kqo zL!z1tbdSUV1UbJRFLz$hcX`e@64L&PQ{ai7feGSzF#IV6iVYa{UT-k5kro552$Cv! z!OML=>}|=!O9LEf1HD(G=v~G0Z}g5Wfot^c?mk;#kauDf-7J5C>xwqu*puT@>R2>J zzXRE7V-F=*UF*(UXv5CJMAwrB+yz)gZn6`n;HG$a6^5nALclVbZOSN0`?!QY%&!bk zzfbtVi6Tu1$w=)4)eYF((~rMCE9z(b4po7x?z;6$I3V&$*2&I)KUsvUjOm}!KGc7B zm=+_M9KHiNY7n5YL*_ewGSre9Gl043U@H9pT$xsC6^%e@WgGg6CNx2@Bc}`P@Ui)x z8fO^;N=!YN(S*U*6Aj6c>94J$dQklZFaup93uX`w<}3K88}7pEBULtuG~D?nU=9As zPQ)ryG5^9F;^g40AFF74g&@UfV4ZWn{4c0~?BIs(P$NH<5rKBWI6f60hg!aVduV}> zpnwbB0&PKL-lZgG_ya$LRUdf3%gSM?8Nte4iLjvqP*mJHE|I$c0}XITYL_drOK&anu67R6BLxT z#acrR=LS(4mB9sUvfP-E%WSMr%p8N>5MLRuiLV_%w|SUf{D9RdxG#4;hOq+JdIMCJ zLnobp1N`Sap$s||-5-2NPrv`bI(&qHx*B~FNN@Hkdtv)Q($72*r7Zi5t}UW87u>FzbYIZ$r{uN`v-h7o zBHUiW!U1PRDVC%@x?L&kej)po+}d-JeAID7z%HiQIGx%P;I8YZ<*C+%*po5`WpsSIPXm;lq67BT9ltGnC|$I%xRq-z*)-7vbJ4Nl61daicM6h;GRiwh z(zwH+n1*8nJh9fTl9xJ|J`=2+Jh{Ugwdg9yc-0JIDZlxqt-@hQ!7Ej@W@BIllkPeZrYj@3Z<}Qvtbyit0ti|v%vmAX2*o(D0zhC zz+#ng@(~q%jmRR?ao|TZ9Kf$3;om!QgQ3`>)TopDS49h(6{&iJYY5b#eYn%y!*AuQ zX^>!hr;ikMiO02yLhIMf71tt6Of|2F{If@IEFTDfQTnitiXQisZ!i!W{QI}WDnx=O zWo5&zTcKpb*rz^?q2(+-h-LSQf;C%!=a_S6-n6#uddEk=qOWxoDvGGi?Ud>0n0X;~8>!8LOVKibY3VBmL4k(!QPV;<-9fUAu? z1a#f!Av%IT_8XO6AGpS5bI6@V8Tpf=KrbS)!PO<9&=OpLLZ3t?jYhs*`|9`X>-T}C z5U$Ej-YZZgn~(&yeZ8r^zgRYQ8yvSrFHx={Vn)fPuC7`m z9Db7#36&`|Cf!KS8nz5>JjHaq0D{j_lDtT_0;dVj8&P*E20@vu}!dog!UfsXVYM1cCyo8R)4!ruD2$&O(_}g#?4mKCXE^pp7Ya*OmX9R z%D@hBjm%j&2r0s&Dq2*PG>`6XIba-WRC?900++e^A22O|4CGw>T~RqXr=H)WXE^dF z-U4g09`K)ar4yGw|9i)HjU^M{C)kh9Tz}p6f4$+T2mnLz*_WNCHe}6B`#laNY4L;? z7Mot%cxDdW13oylVa(OuW#bV-{=}H@Ljo3UuBUo``8SjJ_d}*5V7Coo7i6WMr;0A>bt z7Xs(W^_i1T>0&zSUIa6TBx+hf@BX#bMXPtwLf2ZJoIs3Fp6*ML`y-f3M1UQLYwkHr zW|u&Hpb}zW_GcYrV+n)d>&cI<62(}}cGW*mBl0I&7cj;O8&J!`WI%C< zpaO}&O!emg`}vN%jIem1x5UrZ-sF*QR$qtw{l|hSkQpXn>9^EJ?>!K$I{=ANage61 zrA=^GJqcd!RG7>+gZdG~P|qbJm(_L_YPwk{IK4ShJj$`019@X#Hn`m=WKyRSq=8Kr z`#H2#Qx6iP{=|X7hOpn;!A9;e+9Hnb_-8B03^5Q1T;bVT>}ij3Szt+K4&+LO(2$NQ zm;wd6uC(mcr3>D3O}gJ`V?V6{>mmh2MI6x0#*p+T!>TeDbr?pKK6^PTfxr(UIi$=? zU{&5tzm+!%Ec&%zahfXwsQP5eU73hR+yie|4|q`$`gY4u8_eQt9vT{m!a$fC7=Go+ z=|Wc&TGv{5b?pWu@@e+lN$%mAeQ?RVi)ZTg46d(DuAc?w)&m^`0Q|5hI;jQ4Z3qB2 zk>SR~xINE`G(Z&=%Pw1phdxHdm;BxNCD>cH&QPIDs6M!)24NioD>`(a>hoQl zQlEdB(+HG_AY933VcC3$ick85Gas#zwHP>AfE7EM&5e9w1()rA0`)orPpIk=9JIXTEZ7#pyL(GF^AgOg&g8oTbza6|8f^7E-Pe#ddg@^AbjZ>wxPE3C;P*wO0A>5b z#FdGmxq0V5IiCLDgiVHafs7g~>k+7h?4YjkVXmZ~c5M=+Wf;e&nB_UJiBfpg0Zk4S zd20BhQ5!&)lbXD@DIKzcNh!i8i;GgTbYIzxl#`0B)@V)OfwF_J^A3bBd7Il8a6P2} z>$a>^Q$vILS$@Jy!v#ObQa+Vw+=lh^NnubnC&CtQ-F72X{1Gl8?>LcI9%Ies634$@(~%IFx(q z-v&^}|ceiureNtdUD@yC_vOx}j?7=)xlp0I1L(b;i% zDq+R~9H{*1O(Stmy>@SI1ar%@2AzZ z_J<6IUm#SD<18XBjdz%8j^m`U=_rG_fdj!qo7KG^!uRAW`}b<_ykpXolY;CN=m2j-?EVoMrBgXE_44N|!F>V? zET-;);XDkDEQxGMh-owTtBHwi~2({c>H-Qj;LTho{}jhnj9 zmptXoP9)jXHWq^oZvG`%g3`VSBP`lVMLkHh?3G_VDXv9*TR5dTklS} ztV}zezf$0mv6MU#FBc7rltqgld6f38t2#=BH&|hJ3TAb0mDG*4Fxr?i3bpmX#iBi9)91s=&wFwPjkH{yU)bu zx%}(S2q(k{fH8nW5S{hxu{?JkqyZxn|11O&wq0SmSQJHway9?@NhTcvKIdG+T|7z; zEmS1>+jN3jcns}@hpxyUyL~Uynr?!awUCl2gyvxN_17okDQCK2xKcL85rzk&tP(o0 z8`7kEAmZQa(Ky&`{wfc^g?N`XCg0!VHL_j&sMhkr_#V{r37N$8sKHi>Yy?X9S&+Ps z<~y3s4*;w%LZ#uqr2axrE@Fa<;2Fs9;oaHT2-2T0P=2a(EdU(R#@HyZKIj43E9>5g z&f$dsY~g2!Gr9dG>Y8-AEp4HMCW(WsSC0u{tEHl1&(PCSH_=%fCgmLjBjG(&-Vp%x=~bMBeCPxl^At7bK?=bEIi&@w&<^G9!?0k&vP7D(O4BUuD7Znz|0 zR^=)GkYzR>g)^_II9UKM9y)$0arto{9(NPbp*0$v#Bgk?pl!W3Csz-&7JA-^%e zl%HH)oSeMn3&XrkM)KWyEH^NtM)7?$>frV_dusgVf^sw}acnpLlIc-*0y+{3X!jG< zFzNfsz(cR@CSujrT}rhm@F%^%xuEWKX0U6cUNb%g-4#4#YN=2C7XV6Agt!W9a5Llv zodWN=HeWhAJbLVs`o%4yTr~>+xzh=v^xwgYgtxx)?s>~>livo)WaE@g^8;d3tdeYPQU$0V8OXW(JU3Mw$LIt>3RjLAIP3Zp^m!G*X@> z5BZ_&D4xu`g`ZZi#_?;>^Dtd4A->0``~dVvlBKj3?&_JEUlI7I-kP;g!zzgo7^*Wx zVax=|5;^e;g>dE1m^2Cj0s8%_mY4lMjv5}^*jw2H4AFv2`Bs+@9j86A%labZ&=^nmTWZ-r2X1q+Hq$tN5{ZOh?tNo>^olGl z_$Mr*GunaRuY!a1W%^yZU|x8R9(O;6yUJ-FGqeVaXz^Hkk={=6RJg)GOLvRPAo^F( z`}tiV?;BQ(%$e*A%@1JlnB7geYQZ)0S;=eSN1+u!eO7X9x8i0Dwig55H-IP!5derr zUh@gXxj^R1tvjgve~YWr#V$C``ho7CP6xmMau-Zi&C|&Z{u-ADaoS6mzT#glP+VqXj`tc zi_R^=rIH2~{J0y8l!?R@8*=;NwIcerlJ20`%*LS!N!FI&o%Wwr#5kb+MG@r(bnMiZ zrd5A^DnP|pHXBmCRN*0L$;^worp6juIJQC2gH>i_VUJBBtyAs4cSKrs0d3$Gw*{?0 zR=B?AiepQ965a>9%Y!1uVmQ$R7d*}T5ai|D2&sXaCKRWIU|diK^U-u7etvIsnpXY+ z4+KQxHrfhzv{VI7-uKSFnq$HHlUS4wO&ou2qpc!>BDA?%!~%HO9Fv$Uw||LKmJU-V zGIjx^m4Th$wB{`K%NwC!O7aLTvqPi?7cj}6SN1ZVdQDz^nEPX|W_XiAU|1k70aGWA z2&(_ed99g$dE6}f0D2pC7ZFw&`}qowY+9^Qa;ORm6dir!n$+y^H5v0gHoDgrtQe({Wo0RCbGH_<}O72hB;Q`=jWC6^GDhiSd>EcfU9+1qN99 zE)6y8cq*-1LVAqsg@_%e^`7pBuaBF$So2!NW~lE8Z3yz0t>s`#AhO|XsP7%-q~nNI z&tpvX$H<0VPZtYi)6S%9OE#vJ!toM$HTGfY&PF9s2ub=B*%ZphnyHn~^)6;sZhei>C&k5C6p7@`+3zoW9aygVMRp|9STW1c8Q^zAfUKflVHZoe2_Raj@@Fhv( ze&zkZu#g2AakNjG=9Uk)qaPlfsY zA9E8#;DSqoj8FA}QgwA9C8Z8c8X9U^9D(Yu|6oN4{>ug-RMl#@%`X>mP-9xG^gC{( z_!@UGRsZvT^#9Tc)4Trt%8WRHP9o8tH{!1kM!!O}u0NG2a&V64qEYq`&!9fkaFI6A>QW0$@|#14(YpC2gJqPvt3?o8i)3>U|t^hpQfg z<5ST1+^D*~;)#YwAP>%f8Su`XJ437!@9U9|LGJSc9gu`~+OP3;T;nkJro#-oMG&BA z>94-=6zXq3{>WVi&w*4}2u`_n&|KDK78MGq@?nBUeQj+HJkJB?T>*~00Jg0Ib{?0X zDkl;|UaN09-5f#zuj$3Q>VZV@p*z&*qvX4cc%w{6=tP%Rx5eC8Z>g0HG)1{BfL z=H?G}i;A8G`fD7W{BD{DT_Qu-&^puwZ1P!jY*p8VH!KVAQ9qwU9#kHD$1h+lvKUbqVUpEE36mm#5J zxpbmz(GqcUO*|G}{>xnA&wC8L)a&_#UYb~qw_q=G1AX;5n~)BPV2o>SE<{>dOvYe@ zp)bgse4!l99BAy(Ior44(_s;QZAI?3==ckWEU9`Eg=q$Srs4pETSLz$m@iN=|M@h! z1@To~oERB#IOs7Dy1oWRw--?0oCN9g2PA_Qg2?a%Sk=|p+%4$w$G` zb&=72tPROt)Nh3;a=zaEa7a>4bP`V7JC#W}$d0)LFncT&L;3h^*x0}ByC+8n_oizp zfU^{cq8wmt(Nc^djW_n3pUy8z1NNuXm_?M?3_Cnqbnr}Lnh#mFio zuPs3IzZpbpX1VnshppK|6gg;Sbv-jy*7 zIN%vi$S`bpavqFkgu%H3CoC>;t{;cz9Qo@(CkSu_5VoXeU7?ul9tFu?#{N?x@Y&v+ zI@EjQY;0{;94xJ z<_XW9GFhA;E|xLM!2>?h95}A7&(AFGOA#_8Qd5I@9DR}ogAkCC3)2#O>LS}FrC+Uh zJb}lcohsEol`56;_)J>T7Iv6Em6Qz~4LIM1i2r&q2hF}6*WP9*qI7PWrs3!vjjoJa7wT`Q`qbDzxj5Hrob$2)$` z`aaw|qfIAk`i_I*vCNC%Shx6RVHF_w&-m&%&QN-jQk~7trRm_zeA2cLn=yI;t7!&N z%4Nkt3rrdPvj}L{SG4tSiYj5F9mhPq$a8(Dow3%hq`>U?7Cf~lQKT)Y8D4ISB@V8<*U)@$86{zZ^5 z-GkZmBzm)pZRbpflisPxsxokdAIX!^@0v!v9jx4rTc;s^wm4xyO#BR7qQ~1Z(XRjm z^$*ngE|46Rs(&1Cc_~v*>K3tJ%G}&VH68pA+{*<}5Dvv)V(;u4_00QWctkjlV;aQh zDev?*2`|UK+utP08*nPVKb53+TQ>i|iO^55!jJPa>EivVkiWhgu868yO8f792VMLb z4FsJ1mDU?C9(=ek8F5kDT`M*sw^sjYSbxi1_P>4)-OC_R!P);PsUvPGQt-+bMx#r% zx59p`+~xk)dyM?4I7_e8al(iG{W<}UFyJVR40k>WZhm!&7Q;np8JvIy;ykDef4w;i zvlwkrW?oHAyl@)2k<2b$%u|_qO%v=B*V;7+4JY|v>`?6-PPbcz^5eYMSNCh+s&0pM zy|hf}Ups181fiYkw!7e$t-IfHKD>vlxro@+Ddstj_rECffiPeLd@J;a*ue}LJauNR-|VKtCAe?Ln(XwX zivs6QC%{li35mMNCmN+VQ7$&^k7?%uuoCyf_7JNr2J7#HmKd@nZeIr(($*|5&}5pM&Z*Dl}&DobuD zyW+*F0~^mgeRl{ZHsQ#nl_mF`s#l8|r(so9aBLAP0n5rCG-!dmQfve_(ag{9ygn1R zB#=`|Z7zjfLwI+1SpB+u3R*!N=x_8#OXaXxulr8A|J)Lk*K`<8O3Ejg2gb?+6kkzw zRn?N_Sxwvb@$z?0`NHPC^IN+s;r-`}qY^ zpi=%s6T|o6$wwr`cJ8IdJ&XzLmM~+ueYEw-aD^_lzUjVT`CX6v6?~U3vL0j9J_S#G z%6Zhau>j{et=6gTp?6QWeO0RRh9=Rn!HZ8S6!M0P&-RW zNxcu*x@#2a0U?84u*|AeBAAy>+B8891f8?9!;Om>cvRRw0sXy&YBS# z@j69RslMi5o?0ZP?Q?G5o1H5z_MJO;uW6RaV5Zfp@~(@>rz5J3;F97Po|JFf1kVUb zSXya5+@hjo5hXUI2*)5(&HtN%%*4AvLfbMs9K?oZt`)Q%g_aM~2Oc8yZ&8%zLIInh9crZFI_>Wk+94R-M0yx5-4ZTJ7T-!$==pP zfIw%DIl}l`nT%sW40J4}$#QwIhPRW6{zubP>DX_gkjmLPYrLzZO~fd{84NQsj*%F| z03aK>eIAFM@d}umvLJEnbGDV*7<54Jmp~`-jmXGCP?CK7kJ@DdvvrSJ|H@-H>VejD zD6X6K9O{!GcGTxXo&E&OfBb5|y<`U-xx2UGq6MQH4@4&v!UF2!&m-p$H4j_=nOEWu zSpME~si+ZJNg}mA%Pw%9^|SDDoNEa-!k+^hcK2sx6_u`%UJ!fgA-l;L3g3AV_9UwO zt3N$@#!~7cFwucA3&P_2#f3APAEkP^ySp!d(AEhu@4yk>7G;h4E%~f^57*}6HNxZ7 z1~@1hM}%AU_5)lh=iO`W-~6?eTO-L}b^-yrSYsLxW!*##PRZcckf4LwQ-8p%Q~)_; z4I(^OYc^wUN-r&(@v4{y4}Sc*3BpWo{Uk>qhg{1Pl;}*b`R=zGeLU!!$$b8vg&0A6 z1^y*F7)4jcXtV?9E!`Ka5srBe@AN&8s5D@4`HLOGS80-ZlGrin1B(HU3#8 zsHhp>&UHV%`CC&3RBjU>aNdy|-n_#Eyk@{TQZQZy#R8i^Ka4h~#+X>Y^RV3fBK83l mwZw(OSf znUA2Mfd^NFE=mex9>HPZ=k%l__knK<9k8b?fFDQ$o(s<)3W*#5UyXUGKpKr2NbzGo zNTm7tl8G>p;42d7bXo}eX&y8xJUaLsKQM5%t3wCui3DoPqfQi3pcnI8JW7-KEYR6z zWQwO3a%T;k7V{e)0&+4uk^nCPi59~AJ>w87RhaXL%qE>i1J`3A8td1{Nj{;5WGXb- z$UMUL0lf78|F(HH=*^t|CzsO?y;1%dY1)m@pqtk+XNerQdJ_Q8QY2IXnCV*X#Lxwaejn0^6 zG|mujYyf5hh$qB@WKs)S{~9i12HKwUUo&q?rumQq>0#iaA-}i+3s0FB3qv5s=wvD( zkP^(!RRowhJ)1AbE~_oc&Uyp_JX;zCAQJ!`CSPH+aLl2?(uTppe2UyDfEGk2lHp%L zbJ9Myv~x?WI-c#8C_Fq1$Sng2bWd{NAKVgf1Ob|9Mkw$ZqGu1qOpt#lbcKd{;Cc(bq5>_H4w!se3muygN5gW9lngW7u7DW=u3(T#Wu*k6JeMDcd z>U}ha7GzcN5HPfFqcPz7_W+qY1!2Rx+b9TIaDS6{I21eafRn}KnVw8N3x69+dpnH< z#^yfWzTPu2&feaxD&4J4Jv_)bB3E@vLZq@)>Ep1od%KoIBla4iB{8r9Kn&8(*OD#x z7L8HGGUpq)Y0p-YoO7muh9AFu6rFR4(FynAC*oCY->%cJ$RQFbgl9;rd4m4G8}v4}wgW)$hi)wsX}Z`+ zMF0(^o+fN4;Jy7l==L`5T3&Qo zu!n)$%m5XbmDp4zjtcTq1|BpcyWJ!O;VLAi z%PQzSFB*lUN!i$ngR;M77)tnYqd^OzJHgpFD4y6IB*Eg@b4NhXnYU#j2!gV#1VN7d z73AeA=G+2Iohjy~LGJDpUy!!zmvo5u$g7p0GbkWo^UDzLmW8c)oSbt_l)>ZZLNRA1 z652p}1r!>O)7JvviC{vD^Z-h|g?;uLQA;EVufcAsqqtx~k4qVFC0++I2AvwuB5RPa zWKez8G8LqW!WkIhn+eDfVeTCAvB(zU5wk)za;$qQIj~T97^DwjU5`O-k7$g}capJq z_TmvN8vcyjvS+)VJy0j z7dp(t-+jJ0*)k}EPNZ2_IwO_4ZCP1=rl{GCEMQRqh3rm(6tWL3nBdN+-(=Kik^?9U zSU=MO34tu1L3uE!ZvJ&XnM7gzNTm_I85N$8OAVLwvYgFi#J(^17%aP^kJi*uMG9~b zG3u(D`TwvH-U(+@=!)JG5^-Bl*kWQ8m6AdjxEiJ&u!}4fB%`hGGNT%!RS96S1Nb#KgMxD~3i8tQnBne1_MPeE zPct#HRY7zKlY-0>xoUh^;d7YSZPV+Q(bP=f9#-mzLOf{LdP8%TDR;g+H5Uf9cFW!? zOmG!dH|iS`eH@-q^VgmFhO^v)06#;kQrNC=VPIpbk4YDnqeQA3q)QVi#O%$|1A%HrPBSXs@a%pm9hI6zXCZ{MLb)Wf3RTLA0yz>bp{rfyEdnE zkh}HBisAi8PQeH-EU?#M?_jx%Jn{qqW+*W$zV@ z?zW>?qaUrw2DBgmsrUrvt26#|Q*;E+0la7&48*-oO(V?^x)JsMZ#B){i-#+UG2-^W z5$$537lN37uig`f*yV^iNBnan(9rqrA0A;`B+RD?r*)BA{x)VIIU(>x&X2<>`}26x9qi$-GN+DCkJoaN9;bvubln-a7co# z>-ZI?!M^&f*}q3_`P=;3E#V0)5>tXDym`j{zwiX|tSwr-W0Be>`1;M$^Z#AH=Pq>P zYUgx+^y8lsbNFX8HnH+S%Wr4)_RLU;}ln$Fig7>*)i zEZ86h4pk#J?b(vbIcr$32$@k0(YNKaBLQhD6>J;UT0cLQUUGy0KjvOy9+Xh19RbJI zQ!HyTpd}ehb?U^@d5#2jbtA&7HUNi08-aC8H#!WSz4H_n4=*=mNp(Zo0xHF0O?4y3 z5C;GIeL#^02;By|&C4@d7)Q;8TCL_C`k1e;$$@|^_UGw*(8Z%r2*@H>-)RYjZZVHq z%o+0deHwHzUG9!^7)Q;0Fj`7!yP@VjCZ%XBdjuJcREr}Z-uY6B=4f^b){C&m0(!P# zgvK`aG>1W^c>sM{O2G!T42|D)y0nEf{(Auhj*X_eq2w6DSxnL!mnL%cxskws*D)-I zT4_Bp)_&1RA5$wR_2#$^Vu^9`Rp%=f6h~?jjjUTkbXm_4MqE7+ZrQpVwEKcUIo|-oyX;t+4nDRg&Z5*1#zdO2(H^-}yyY?Bws3&Ej zTjk+SR39xYfnKXVI^@IkCC!iF^iFy$=};Y!I$si~k)wqG{{Zf~$x`#dzSflKYa=xs zfG@yqmlP_MrRPL{%(!{)bq$W4k-EeC;OMu~d5c-PrCfk>#+o0$uf&VFMhq};b<#g$ zo;M2uY~wVW>%GoAM_SMRGo-^Inh-Hhk5%mlZOM`+5GKD?<8JJm zT7uKy1yU)#&@Z8l8A$}Xw~ZfxNC||D6m?Z~Xn#g_SOhzpLT3Ytaa;4#lHy#^ zxS4~<^guGa-wLDH;x`&t!as+ruvo~87{F}QMM&OrsEe#KVUs2tH4qB6fd;Q&krD}P z9;*VqR|%jK!Q3mLaQ;Cw=mf98K)(P5v_2FtL770O1gLt_Xr5Fu!H*K4N~HOK!^8k^ ziirooheCy@YDEYp`zoLeZ0WRL7Xo+jAP^zK{yGMqsLpvJkw&G_+0F!g$vhZ;e*GA{ z^Msxd{L2`MJLds-Y`=P;BV%F?l)+_ST#U7 z#w||eXvxmb?*F$d!4V*BqbYBfO9t#{Akl3W(Li{PRWy*(dXQvvpk)7_$2~+{!Nl&p zFvzWvsvia+W!n#fj1Ad3$3Q?pv8AB&5>V_wG^`&KVd}!fMlg~^1NIii(Jc9k(K6bUKEFA-*!x*MeG5k&OJ==2R zGTlhrg4HxfvTlFHH0L_q$b+{v-91(daUD@J19e_V0wy*KaghWvD{*mXRUDf2FEfVH zgJ{jo4eQrIeP0K4xtzc-h{CF3u&jSu3J%idfmLu?`&kc?9Bw`>hf@0IQIfsQC0Z3! zg|YsHbCSItM7gPvs;!ZJcm)yQjB^jq;hkJ`yrJaNDl+a5;zwI=L=!= zic@7XFL5okYwt8m;TR8_3&Mt|aHilBLj%{0MY1$RGcToukzI*qS{I zXZL?B%v*>H(D_AjEjE|uyk&2wkmJg)u-vi1X+OV$dI%J21yDzVf5_sQ7Q0on;8z6f zx{_Z34Iwf%2Y6vRBe(3e{MwOK?)^o`f5xvIdr*L8d+eKFaCE-xe7MDE!NEtEecOA2 z-oVXV>wcQU&je}^vkVKC+g}AzC~{vahi%cN8v_rR#qdMQLfN8zaE1!OFII4d&NoqL zo8{q#L55&5+db1KlyMqLodVjQLz;j)={ll0`!eY;S}kq%UP8JKL=*b7)L}3%zV>9P z)mw?(~fk8Lnbzc%zVGTnVF z2!iV0M78s5IfO^rFm@ZSC&bbA|jEiXDP*u%hW zW`GJe;xSWkEtX%w>%Mx{nl|QIHCvjWt)umHmSi3CVcA!*fe-`Bpg33*Wb>Kp@_aya zfw82rT_*zeD+pAIr!V*o5#(&hV0oPZ1QQe@fvOLG;X@&js30_Cz->v|FVTS3WiqWZ zF8QBMgdQ{wHUeR57KLf)i*T!D{j0tRSLTItMruEHQ2r1v_mLj2AyQ-O}Y6vHqg6&TC?O0WZXg+er3ug5E{TRCOj+ z0|~&S0-aC1y}_Gid#c}bZxy&ONKyb{FQllr_W>8|*tQN_om5!dioa+M0svsCs;>#f zm(&k#Y{=Fs_n|d4R5joL1k8X~q|tKkW59MJ5?#xH0PqMtejtwI>mC68V?}eHK?ZA~ z5g>GjG%j0rr+)X7=`V1%>CnjfZG>fOnz&N z>o7R>>;e`s$a^2xaa&5!ja+9QKqdH+Yzcv0{hZek9~Q$==z>M|8|;0Y$E~!p8#!-= zsd+!kb&TcM!48A;DDQnN*Lvo_k1ZXJtYn1Rx2hn?eW+C&)Y=N_+X^^`qh%(-OK=Xh zYjw4exxST+HPla(*#PwSjK*e74+f9eoWMV0EerT}a;_MT5DpgEvXX_%HDUfiI07nL z?|C^_jdl|mzv?~^E{3>Pg~_uJn`@w2G=}YP4_kzjVAe!f%aBecfcD8K^%-P88ofDZ z+wbqNTAQrzx%)dfWV02--|4lUI5buWZCVcn{b6kYlp1lMlL60}nAaezDQNW0f z?%4_$iPm=+vKa@{sD_4C{d@&f2+NB*U*Xa0v1@d1NS-(aTdA9xd8iQW7SP>1wVVT} zv@jKJwFDgYI?HyWNR6Wa)stz{h%isiLpT{&_7^Li~uh(wEh$NfOKo~3H78C{Jc~t zv;a+23W$vds0Pw#RIt&4fggaE0?$!jaU<Rna|_xHwIA@W5!oQF<1FBC%gq za`r&4kQ;ii8AE?)Zrc)k33M9R_r$~&yfrnYSyHU~H57|_YSrS#akSd$6&5#F4W+>? zyUGC>ZFh?1-WD2oiJjOw7!ghcrjdrrEHcf_&_ptge+8hqPdB%eA;)whTD8Y*Ud4Sx z2m^H@xtYqAg}8_T$O_m!sZ%ufnQ1ps*Rz)WfClgVTUoI$RC1z_K$;Wm?*@#Qj$5tV zxuFx~1ct%ZT0XM~PLFRE$bGIlV8{ORD9O=2q4#%+=FSSOCjjAyU_Xd*OK$f1{Q$Xh zLnneL*Z!-0p;I*Xnd6pX>^D5wI|{)LH>_TJf2U~foX~Dqp@Vjyde>G==m!LlecY*( zKg-7aVYsUY8GT&wsxS15=E|ig1nJx80Lc=l><^toSl+cg4sgKZ#$&X8{b388W3cMI zC3Mc6;kZR+n9)`eB^XR@`z1KY-puMPiqN3oRELVtIvUaaKb{6D7y?7+7E`&EY9cEN zkxrXICJ>toi8_>AZ?6Af5P|*L-AI>inYj)ZZ2cDX+2@XH6&KT^Vm}d3sekHOWT@jX z@asqHGum9qWa|YA=?T7`(9ZL$JKT4|E$HBqoNJO=p>FfzAWs1(-pZ+LjKJYM{5y50 zI4um44~9Jjp!-^cthwbQkTgcm$wvUK)LKcOA5a)b!2w_mxOkdM_J9Dqxs;j<6t&QP z?JB<(EZ8$cJHIcNhGBuU03yK;R5ms%v{q!S#asY!1f*b>=6>3M#pvy@AvC|N1=5Ih z_KSC!b81TV^P=?AMU@Cjj8G=T08pi4;z97CP$6)$B7njj1(X5jba$zp%A|`Ss2uWO zTlD?^*evacQKu^}1}?U&w(4PR>Tp_`9I6Ygj^g6B^%CX9EjPeH**?{;o?{Jb^U6Y& zgdbF02~7cOrW`regUpWBV~|Xzu6@IUT_M5-Lq}nW{m0E%;lLxFAArMn5L{GMmV7dk(gUnkr|O`VbZ}SWzUNIGYwZ03TS-}_F?n& z3j&L3AoJb?{2#2aVQhQNSaJg$4g?_*^*}G^R6mLz8H&rbW3Pe>vCIPp()@XZx0l$a zu<6qVp!F90(Ry^3h0JWZhn~ppMzj!Lu5*kd6)V^dtvBtBasd9yB@| zm#fo)0-3_`f5cn=QGVAz_T5Do9%<#wg76w>*of*e_#R|yY!@8nhw;RP5gmdZW;ptx z{kiVE!9msjqLJ=hZ;4e!{i+zz;;_xJn)v2Ml`P!$YiT7byZt))R}2cej~6=3!{2?r zIoUEOgifScSURf!=7RXEEtB!Lw|;{PM%b5IDxlEqZ^_fgHgv;rFc5p2c^n$0igdp~ zxT5p-UDiE6TwU_H-1UR>B zj3;6G4K@VWwD3Iu*tk4IuLU;P_F!>Mn)jpa2Q_Zi8kha{dWp4$bmL;7HPdQf2`+Rl zs*yCc+_{-+9v);IvEL0_yK-E8hU#^4co3S`8 zpb8klY#8>?(3TNUR;TL{3V5V4r?p%5283&iMcUHkf7p_NNt>=MA;cparXcjuR9EG) zO&L;H&z7a+U>;90&4<}#1Y#+uQU`6Upbl->!2N{=TLYW;gc^c{3+p9U&{ppLv? zVjL-WZ+{QEy^Xt;7o8UDVc<41K!sbBpRMNO53A;9g6c8~kpQ0e|0$4%W`o7R`}6D( z5!63{4I+E|8XOG5NHA|<;Mg>aZbND#S^4U(7SnMT^$vtt`>QRwwb}xM1{;5}wCKRI zH;F@`FkBXU{i>^A<+oo)5jEJYNXDJxT829j%wg8{{{j?bm1k?KK>K@iFuKMuH=_cR z$Y>VOsXaRWU4X}dp+lpG0Uk&l^KaFM{QIBuzD%x?qO&=kvLz) z0q;fkrFw^=G!0Y`h1$nT1~|7AN8v#m3K-uJi0=H-Cp0HBbGd78a8}ynl(xK;KDB_j z*|kNW3E3=x1)6%CQxjt8Xd0~V;9}~5qBy&0Vdk5;g!RO@3S<3+GjPP9ungma3-m!1 zs^esQG=jog&0iWpG=M&}>d>+x9oI>KJ%>I9bef#oN@3xe#I8d*Jd%Teyi*!J4wasuq%P;}7XObsNYT-wPKw7c4(EUn1R59N8~z~zA4Ih z4R!~#g(2xGSI5p>cov-K)X@q zZ4i5f%g&Yvbh8F8{%Wv{t9*ALR@P&{>L0Z``oZe=k`7$?y=B353+4QWVAU=7gMbTy zKZt4C>-?qGBd72>_pWW`PPez6frM&-QQBhlF*wHFKZ!%J{X6L)PFo&@gM)A^EZ5`m zLz;nfQ#XPs+A7qBs%~v>WEx;<*^$M#oviYjH=bl-1BYMCwLdzmu5z*KW?VY* z2)`bL!$HJx;20KJ&W+ru3+!+ltO5|1S`#+qkX!z?9gd8$b(b1CqrE&kauskU3*FkD z zybebZTabzi=(Ww`h+MD;MRS}9i{)^~5E*9yI|ON1srOgiG49j@3z3PFkkpj0h8$g4S1pvHRFyML&*+C%bzzQ}F*)THJ@n5TV~7Gz%dS4aG`G zwosVQmW#yNvwb@CQpMXYaDkMKlz< zK0+Z|2D1nPqI*~c0Xf!#^bxmkIJ3;p4*=b-pg2+)^Z&Q$2c{CZ6npR7Kd-xigt|j} z7IN<-d(*gs*W(H^jq>FenC8^Mv1_Yuv)HT(YWZ1mnHQIym2mA)Ps6LhA30*9Z z<_C^)7n&|-9vpsJ?R8LFI5UHKvR`M1TE}T znQiUR28_HgnQjcWT>z~e(MC}JPmZPn%{S^61qoHYv;JY;6tT!acP3cxurS-N(ot4+ z`*pMjDNSx+Y8J3&>N-OjIVb=DHUK(o(Ejhj4o4k3ii%X?Rp20$N8Vy8xTU3rA>;+jDRx}cseF+)kTM~r;j1OQPF@6mH0KgCkFhfV|D56q4 zJTg4yM*0q+bKmC<4t2bH`uLCC_0j?nH>Mgb7&4aM&d5kdCY%58uAT1uB}=bAc>eQx zlaJo}o#DAueZnMn{rD&P87rQQ4-^sDT}nHvruyF|QY|UT)9>J#71&psh>wV>+v^wP zhep~^snm(iX5oAsud5=}JM=b_FAwB6Zl*WHHZ@c(j%NtlD|{uNlCAi%=LC5@UiMeS za~;k@yy7CMdM*Odk?|Z?;o*<2=R7n7JRC21yppsD$Ax*tl%5P^zZb9kyzv8uu3woN zD#`wMJkUc6GM_kfOhkHtQbXfcSu|QyK|#T(Q9@XFz|y5lXXk_pU%7GxS3gy`DE(=D zR_!cTg0r(E??B%8r0pxupFclIS3+Fe)s=sp+yD`o0V3%Wc*WA^^Oh7RFrFE|PA@At z*EY?1?19;*MDlIMlnz_T7dh{I_yA?=b3Yp=)eNgqnXhMSE982Nx+lNt(J9d-OO~8G z+!QM?WXOe!ZW6Lnr=LyNm%Go+Ntl0EgGDcd*b=OR{MW6Up$#V=+o-7s`GvNe$Vl&N>)z2G@UuY z_YDET1a*?9GYD3m!1wk}=5o_~?dZ=E*NKfWZwd+vFC5x&>Bf%5gRlOmpFMK(liUSz zXHB$gKGz%$EA}`r>%@ra0aHRg#svF39FLcGma~>M7Mxa{nvC|mKIXveVz1__@k97hS>Wje|(`QR!)CQk{^0M+f<(nqN?@{|%%YWRbY}LZ5T#4-KZMvzg{M~t_>K*nbnl>Bn%%xv$zj%s(+|u! zBYy1N9g|(FVyo|#x)Cm|8a~?R;nI1!(;6G(Dn%ruB0ZCr40P&em&n(-KBx=o(w+|{NduHtej=~ z%Lmz|YS&~E)4)UJMDXwzNttY3EDs5Qn!M1cQBnEQi!J5MP%CZr-dtSw)zjOw{hO>dqxc679K-qB`orsH4*!z{-&Dp-2^1*N zsUKtXmh|vx>MY5K=ffE+{m~+vw=z>MKNvh5;P_1nHxB6=#*EXIjyfjdID2N6%faUX z`C~&;?eneiAs@G8u9@Orvtq&$b&*#;j8o<1p8lwRy6)rS)5O9A1?SU?CVm_pT$+Tn zm9@zpWFb3SLUp#@%PW?X-Oi3V=3Saseqni#T1L#5|L$bDjhl7c_rgZEB~gIR8ocD- zFQJV4pPT^Z$9wItS;kCx#AffemLy5 z$B$Esztt_+Rq8k}Ola6|qS<6|b*`(Zy zS(0#kw!GEe|BCjWQM5_Eb6HFzc7EAjmFF*BV8^d~??i5N5wSVMl>U=A=yBEw&N5{@ zPBZqG8vxc;#Yj?zvO{yYw`i)~sk(+kK^$1FY6Pa)VEKi;%W#M!K!_*PUpYyW^j>m8MSoJd3V6)0N_SW&9Z>UGeYjh89V7l&EkU& z$(9nuYBd*j=8kvvSHDjv^)!aYP;LcA`ULav{Uj#Fn2HF6oxU0B;<^OOXCWWGl7Zvr z7XbbBZu6)Am=7c3HxYc^IC=zb6`$GpO911w7ll92VSZUWgpYsF4&!2Rucw(h?nL6V zDq7i;9apbf?*3?I{I<@4H23P!oB=p#^N4KWF*Xe4Bn2Lt8lGnlBgwk(0{fI{i`C~> zCZ3-Sfb2^yx)5J>~n5x-fd_=HM*|Srg zDTW(W=N$5VJcauB_()@8gNgA83CkXz2_Y6uvOl!o+Wh_R4!`6Zx*mHp zbTTyf`=&q@jMGiFVzTCVzK`dsmc%S2Xyd5&!<&9IHAzTm`WX59AHJJAZ0Jw}i&C*1 z{4?XgkB>$rrdj4ij$7;%q;qeZUPSgP@pTYB4)_Mf_5MZv5r{J3eA+9{A1TpZGvtNh zyjvCKn;L42wpJ9bkexjFKT)?WtJCsz4)WGMe3zw_B*I^`FVsZ~dZgLL;k;L*mcjPU zf4m}#MIuDP7JMD2R(7po&$)BEMHF}yZZ~XmQL??hyZAJ8qw@KXts%_hPgTS7`GW;M zsTup}P5i!S`m?HG4POgN!m~aZQ1;z^*jSL*6lI#+U=um)&ef2gQJZN^w<#YsL{Gh| zqCWSlxZCn?dM+{1gCUFIyc-xxs6CUf9fl3iHQZn@aJtaQ*~<#bXRH{UBy+VvM>bt- z{8EASs}sS0>jl>PN8c_Q9@3=Q*z}=lghaZr+s}d1J({v^KU5iM8Yw1+J9_!Kw{7ZV zlPd!m^TT^(A=9o)Z4}`9nk1VSdSl(NNhYy}L|(jlzMoGSb4YA}NyWY?7LnuCc3k=( zI#$(o>np3vFRR|Ch`ctNAf{w`i!=v}`iV)nuo6wmziuNmYZ;ru&y6?wv}*F*C}q^{ zeUhJ^DdbgMQJ$fn7pUWDl62sWGJnx%SM3eZxj7qw!dRG}tfL4KFHJ-MSCKxGr192R zV#i#|_@V19OA{$!na)bKnRsQYC`6u<(f~f*#SWQe1hI^-mEJv#yRo>-g>icu^CoHR z6xfk>dV*6F#*uF$Z{@ucW2J1M(;Q90c^%2=VdEirsiyMETG#&PKVQ;mwTZuR`lz`F z?7uwPW1ISjy!*a2bVmAIV7R@lEZGkMczBtpfP{av#^s51g8tD;CaLk0FFqd1U-Z}{ z<^w||za6g^ksIuYt}#N-3C`ICjy8j6L}cr?WnZ|i}J|J*-F-T?k~vi1!p

PktyX~e6I3EmHQi4;3rDUFoKzd&R2N7dB7U-G-g}D!bBg zW~%(#iX1b0JCVpU^64CsN}?Kv&@6&Um_b@Xg}e_u~N5yw{z~Db5Zrh z%-tY9bp4*)p<|%=|DFVVgb;q{ZJgdpzUMI$7y8VXNZtM^gK{9K=GZ0g=vcE`D`&fn zIa9vwsO8daIiW78*)@yp-e0|uu=2&!m+>0Pwv(JPFW+%WbAPcnw&42pOII`=gD9jfa*3DF_Z662!+Rn3XOFCwqD02 zKk)jde7X}RC@cG_BP!CU}VU_7uWU0hsdYHF|n=zhI} zbB{c1>xTKlaZ2_H@yR)}ek{Udx$u}f`A(mRx%`H@L2U$?Pf~5^80_UY?#Ws9r5X=z zS>M+3KkP559jd-+_}Eo@t+vW4D-Zqo^C!-6^XAP_AIeS;@4a_%J~VFtO3mY;!@50+ zcBzvUuDpO~d437tqHm+J2QgSml6T1}?Yrw#4C)7kdClB)V{ff>_L4VO72L*H3Z9P$ zlzUFHd~k+jc{?CdRwQZLwgbE)M~$Lv)Z4Rnufg2w8x)^eM9?G8tnmQ`W3?E#uFWX( zy5$hRyqA~RTsK;>=FX)!#RpFdALec+StQG^xW82OeCVl-_yc}&%&>Q*_4V#)7cLJ`GfXBepY2^?p8e*7_Sn># zz>D{9?<=@6BVplX9+#U@^O|PJ8;{%(?4ZFLn=(l`p=X`Q8t;^S9BAl835nNSpEE!S@&qE84qHdf%6zY!gl3XNe2f1WX@x+v`Xcm^k@) zLQ0ct?i~nAIA*07Q)~Th9Umco9}h{wS>x%t3Z_mrJI_8fE5Z-f$lHG*>82fvTz>eK zB@gL`=rD%-0fvS_ZI^34)hO-X#zBt>&=nL=5kWYSW_%(CP#4CyR-T&kb5cuHxtC ze^W~j2&lcn&>bIvR;%=t%d3N!K717Kg`e3k_X#~7T~T)QWntOg4}}948-=booVQ;R z*k&`r;+Bf;C@3yDQxm6_&o_ARJB^yE8NA;I3W-ba<&EFIeevA_K0ZFL+_Ov0F0fv* zXWZTI_x8)a z{c8`t&-Jj{y2EGd@%u+Y1eZMhd8Sx{u$(b5V*nYo3>Kw8)AVR{c-CQ^#Dd$ku>r%w zl$Qz5K9s&m#3Xd9_A##~ytxbOO|EUh+PRYo@1OEc8>;W%AolQt&$`9m>Ir34WwvL6 zeP%zcUy!tW&-9J5pWXM;M~*uAX75(N>T><>+0`2-#cu_E%uZU>*>tya(}17i_x-Z} z zY2z$c*XY%&1scA7-6ym*$5Y-FgiV$%{Y)D(Ws!i(#_Y`1m9H0?IyKC_J;CbcUMs2h zWhZR-i%LobpFMjfFnqYNweQ+{b7s)Bh*3uQ6*}=#qhdyii97I*p38tXWd%LpDl$kY zZJ<}NZ|>1}^_~9GTckurmOIV2D;c_GvEPmRuL{@ul*|$_F)}uJ_QbA!r{n&a_mfZ0 z8F#NFMK_CAA%63$!H=V{-lg*%HwK-2Q8(2=V2NwB=*&#leK+N`mRznD4*mWq|FFlk z(K+Z3W5;MMFxnjQ_WqFE1+}+Qj8iKf9V|Qke2v=Kld^8&!`F%miuj-QzCY*wv4s+6 ze>Qq=kA78Em32$r4llf6)}SXlm&yGQ4?Fd&QA+0dG1426B(-OE{pp;+L2Ob+7jct4F;YDIxLVMC`eVArk>ZnSy|YV`{|| zkx^3@Hy#z;cW6E&E9wxSB@3?uZ)FCYkg=VbeZS!N#x=mg5hhu_II??Fg?U}rluaM! zeAB*beIVA~Cic>m=@Oq_&Z5bxsoCt3krn*#@VI@%!&3o@7O}cDLgz?@B9rIlNS+GY zSo-0PMf%FfaZ4l*x*or{&pz!?XrcS}>KB5Ib7#-qHgu+m$*Q3RyVgz1m{Ip6rrxii znlOD;-kbc@JW;O~fBv9(|0G4uz`!8z;m$9bii)Ok^cVT-N>|@rKhgf^({;La_!VVm zs&dzCrJWshLh#hsttY7Wex5j$IsDezv}Z?3bT5TV4R&aln|r#};i^#r>bhrnvT*%t zOyT2&!QNwQj@%jlK~iV_(c{xACjOW`<(`jo+D()=G#}~Kko~#4#G7GAEEl*VFv+BT z)mN{ro7NqYDN0muR*1h)eP-3JSr(eHTc{7R`KJbd)LXFNw8^wpg5%cS-{hPWRXJ$` z@n`*KtTIvg$G6(Z9GuUS6%%5&riWe&T{=j0ZcgakN(FcH%V!$ZpX)DO-He7}PYTxp z8Z~a!SKKSy^rM#g{_zos#f@KX*4m1Ea5=tcvq_?|Xu1?FM;OIlbGu~Pl_$9i{aro!N)oSJt`kStJn5_$=-z#x7noS;Sl6`Sy^f7pw;Q8hrM+=9nr!e#PpN zYnqQ^jMqp^+!*~t^uhtN)ZIEY$(Fm;;{=*A=bVELR3_dHvP%f6DpospC&)6vQh)E3Sl}>&P>bWH(b)OBhbT<@@x?x;>Djl@ zu_9p!kxdZ;2fLRf>x{mTaC-Khf;*Q$`1g%io>sui@~aQftJlO{O0=041^D_!t>)Z( zdz)!1L69&*ea=Zg{XN+y)1A|_M~p%lCF*~F=hk%bg5hB77Deyev)gJFT#r*-o-LFm zB@MZgI&08Ay{J3JCMIvi3eGP6mhpbEX4BNG-+Ytz%aEo5(9Z#Cw`ls{#=_{Mwhe*xmdH;D=+tbd<9(>z-!y1MdZ#oGmGD_n=1zW45( zmYes>3l~m|xK~e(-6RlebZfV6soT=r%z%WYC0DN+Et;LbYJc4`y|K3 zKDjjA&d@$p)90bXH#s|{H}47)6(=pu@|`h9=hgkgL}d?=+(ezUibdw?nJn%bT7G5(_;X z#tscj*G{d=EgpHW<|5_Ia8R>GdFAi6ik=hhwHY1#ENnSsB+ z`4SUtt*;apnV+8oqJ}y2`wuAKMj^8E_G^N^wVWAR(9!cdwovm>Zw!9GDt zWSaob3+Zs^Op4`mwO zb(0^L6>b(Wj-V$^UHw63KnVzn)NDIUvH`-c9T$EeEJMOR=Hu`KYqTpRzRt^^_~6|{ zK1H>M69!1HK5|R7P}(Hh?!C{6$EzManmuOPq6UjqI^ND^3#xbTmba48So(93mG@a% z(=_i4zd|AzJ@3Ici|ps-$91A)%qJLlucbVO1R^vA1qbnnI0QUU$;##Rog|w>qYYDG~U_^k>ln)FBXkYPWHX>z-gQ- z$SPWX|7H{;sbRoRczEVEHM7{!S_~&^3q?_;WqYnasPHRd-X}XZ-cb{1N z=C_u`U-FBVU%2mCV1M@TIbWB>H4{ItF)z$q;HJFq{0ZfZR}TpZFETT}ZD}mp|Kv{Q zty|}!pS!JXdNHo(U|LSz`mdVP8yoMSMb$rUmcX8N&(cd8x#GmrO)E~grL4#%+K%^~ zzG-u#E13D$R6e66JO6Wu#*!v#AL5Jnxkms) z$6l7I5tZ;*v23yu%CSV=R@SoSj80|f-W2PyZ)>z28;;IBOiC&pu^4|$FFW(QVBS^3 zGh>_<7@ty2aojg%j@*BhQRPWf4u}oXqTNVux+uN!+ft+!p!{aj3qSpaPe+=Tr%Y>l zx_^q39rebx*R!#zBL&E(g0{x1h&Fgv>7}6D7k(~ix)8o_Oo`{Sb3Y@DkNm7Jy6+_a zYy;{{-O1N+?mxC|nS|ik`J|?EcB!Fgz!-7iWWI_ElTj!6f?`*U2-+DfJkIsN%?ZLv z=G8*;1lFerTO}NQG5Mou>ZQ1iH=fl+ff)3)&lB7~hp3$y7w7Hze(fUVg!eUrmh)ED zolu&}h|AhHWc-I!gCO`i-?HKQ^Mz*ZHXl2B^t(j)BC&Y|B5Pg=`I+jcTQq(wh~#xl z3K7UM=6wS4$CWz7Yso;V+&9zPX?;*+=naj1dPiP-Ez`PtJl{i=bD1BEAEmiEWm^70VrOJtxNa zwSV4vo$1n-w>{cD>8}0GG1DK8<-c$kXoxj+s~i+&%v}R~eJcQ_j{?i@QvsO15{(P8 z`tVLH)oGWu)d;)CSGK${yZd-%NX|6=XkGw``8KL$mb{L@%W0h95K z*ka?wMBNEatF`S?w=H}DB1Sf;2Tl?3D@zyw@|>-}h&YpEU`S4y)a+wW_(X zXUVigbde-ySUhxc zTLYeCr=H`}SL>W60Qd(3PrKul~rW&*jfIyJS^Q zy*f^6?$&SX2B%eTmb+XJ#$X9WnkArEL-}alD3d$pJ8cS0#ENDYPO7=^X3Y|OVrAqb zgK=UT8ua7~?#T*9R!)=IP%q@;LD-;0P)bR3NeAmorl?6Z{#cj0dT`#9jN(#iBe1bkNTed9j7M5SlI#hW3)R38{#9m%k7E|(! zC+EtS=K`uenOgTA55phD%)!g}AzBOpDQ76j5pn48zpIDkQ z>qvZF{*^pJb@>HJT%hC@oKJLmx>CcpZ)JC~K%$oV^z7Ip9wir-@ehj&)408N!DOu{ zUh)tHX%LT|Hyo&DX2LYwNhofxYS`x6vgTov)uz%-Liue%-t0b?q6-Q}A_oi{G+TT{ zfl`6%gjvx>hmWWE=3iVnEPl(D=$+4tN^d`)$G*jz$}4MWxqLM;F?pz9d(n9A_4SGb zqlk>PQ&*PC4=4d}So8i9xe5ppRp$I)RjJgl8xv)%L7~Atfi*h3L0?|mr4>gECWN>; zHG~C^cAaU-I~Pg_A2Tdj3t|-xM&W|9b2QD)2MAPuey>F=OpXiRF9fppMP`S_0_~`3 z2)Gfm-@SBNpg-alP1*xnT z0IA26tj7XvaZ(FR zg!C9(EF(k=&~`Hqf5w)<;Tr*8`RRZ!OQl4QuZF*zy%OBU(PSAgphf&ffiE|X8j4$i ze7P2US-EP(a9O~SzEFfndca&pNEJNbY$%Z$l)p*^*ti^ZXK~?D&;wqOIUEI^&Jobd z@mNk67HA>p1>)3VN+F|Fjwf-ZCc(+RAYM_q;ZmLnATB-;++lN|izYlhCpE!bE?hTm z7MRO&K5!95J#rE}1Lfnv3`nEm-VcK2AJB75egYOQ<_k}Q5(W=fXX}EKgF$rtL2TSe zgjV^W0we7Gjcg+L?l^e=z+&wyO9Baz^{nczE@Au*9qfDEFs1a!@t^N#EX z-)#WzUGkz#k>6=T*L;zFqy&7I1m1rdG@dbXF!=Plj!p(R2+<9u;@I#qUD1(&S6#$$_$ta823G2!ENv(?R8-CKLe`-(4IjzQ!q%a zX!&B2u2uXj5VN|-|ho(z)X1znJJ6;9xQ<$(ho|Xf4-mX0X}O29;V-p zAi~8GW~oB~r80O>Y9MG)J{OV(>|Ck}A`QSnaB#f!@)#ry%fQp9tcr?d=2T=N=0hoP z@%K4OU>K92a-EvjhkU?vLoZth#qSoUI6(}r4<1!wj2JEo;pr9NO2&hhU4uArEcmi? z)Ue~oIA+I!FYlVH_y7i(9t;?M{y1J7BWMUQJcQ5W$W_B3hF688LAK-_5OD@Oyn~+M zr{v+4-~k*h0mEN8AEt%OfE>7u=IF^dh~ei$BVL8hVgL^~`6D2F<=S0&IEdjvD!I8T z1YGtc7EIkN!EsPYLHU2+BA&M7?eHMWM}a{mYpi__GSA@yA;1wxbVL|15@OE#;n`Ec z$#oFJm&9SA$AlO+4Pwr>1=()kI~?>~%vwcc6#3wD*mt2L%fWZ|AcmhFlaDZbJrL=z z66Ygn;5%{XyI<9^;S$g_pQRot2Hz!tYkV5f&+z>W|IaYIpJ&9-p+j$7uYC0?U=zSj z02bi2O%F&mU;2Rr0E)Ro98T!*uNpKy$3$$|MFVBN%_-a095}LBso<;q`bCQu-^`sA za_4}H=>C~m7dN9l$4n8LHf@@jnVB9iY;B!wS08;rDJUrLdyRD-1*7B1DF7SxuM;W* zMUA$Vd>OM+-QSfOx$kfE;D1|jU9JF6cS$fxKD{UKk|;67He$<Gy5rJHwgn5E z<7wi7D%wu`m%|@B8f#$ZFYHm09GfEky&|sBWe9SIVc-sDV#LbK!5zrZ9gLDI$eFA4 z3XA1!cbTlch_=z)#d|E|>D)yi@7Lh_GvuV-vt_3;lo#uQIuf0QUoOk7(_XedbE3e1 z|6K;w;-|OXSCnDD7X*Po}{J32~w zQMEo~<{B9rkH=z1($0OxiL1_D{hhb0)F{a7>41`HpC>ogS7pU#UWyca0jiwKzr4A9 z@<(0SVS1&-)vcO>d-BFH(pv!Qs{;Xk+KeOAtpntk&pXhy!rku`kC7tI)M(K#zR1#L zZhA`5!vf+q1`XYK{X^c$#D&#AXe*sho_QW*G1IMKohUU*Z*%6|497{@p-syr)#veC zym+ztrs=q6*G@dW406L%pAjG>Ic`EsN^1PU=g-|ihE}lHX|ieJ1S^azw)XBm<@tN| zh$(Fy|JjJC7fwKf-nzrx5Da>o)%m9-x8*kwKpHXwm0uUbg3;${F_y%lMs-1jk;x zwy-8j(J|G)K(*MNKo}4Bv%IQm-L`FGEW^k{JtI8-uFjsdxn zu~UODgL-(s#ZTw1j)}Q=lb9cL+SS!{G^i60l9pZzVx$0Zy1R4 zedxSHF9_t9PmDP2SV?+(I%v~0;*@&_TwVx8(O2jJn|Cn(E>JMw_4xF$yCMoT@{4oo zpDet2+R@u=bihQEqfOoBMxPh7O&Y{i)~{ZiwI7kI;KG!x@L@oMovJQBL;>3Dy5hWO z?bwC$$D9uSkb(MAn3gi{=5u2+8=I^{14N#^d75=DQ&8lM>$xR5PZqtiiZt3eW}@w_ z0kh=ax`rhz7$L2_TF!KB`sA_G=#tQaizVMaXr2*qz%N}9#*YkkGMIZ}ST+uPRurwS ztzS;XNJ);42~Xu|kUaNol-fAAlu5!O-dk+bJbYrLg9QgY4qUVM<%SgksK?h5R$JKG z#_g?N^7GWIn+IoUes3tYX$ZW2&hPuxA>Lc6HxlKZ?-HJiE~1Vl-`=!!sAeyEDynHD6W?VzsIavhzMR(QnD?`*| z_j-&uJ6nD58qLY3{^n7)2VWc=k&w5kE+E5lr*cAC$+?CxpUw z>*ne#+~HwhRpIgKt+eC5sWXCnHYf;RF3KA(Bg0esuGG|y5O3{7LX0IKhGJC^V{BBA zu}e#TXo5SnV2D8Mh{&uI;^o3S&#VtVeDP#fq?|5VLbORk`03*M?bjx+`ft>-sHG_a zO{--V)m@@eAN)9TEy!iTP8D9qcq6SD$rYnR9xWN1H*K2u^o^=lj)tTSF!?#;LCV{` zTYWEra^}?sXPtQKac*JVM|}^Ey(W7rl2SGHnp1*3&ov-(nI3QE-v$NeM-tC4qW&pKV3hB zSX+~%Y+H8R2OVOuvg(9S*-U=rJx(Eud>XJop!!uDo>A zC6Cfd=T>@xnj>W!SuB0Hsh=3l=8b zs*slTE>&=g4F2$NJO*>e%y6k)$yER6D@=+L5))&^mY3`)8dUEqLF?#`*QBtH^LZ!Pq-ow4meV+fk1I{qx48#87obM;kNN)14 zcH({56-5AvRIGp^m~CW9MHNW!!Z=c&c>D>6(*YxuYO-Ma@qDdm&oIlg58`~sxL0SX z);kN!DX(<`^mh-?RC6S5{kV@~MXrB}ZfgaQTr;*ao=$wa8J6f?Zmp32w5h?Ok@x<9 zI#Iy(<7-~7mA@O9b=?1deSKlK$R8L6 z{gsJ#2?-D;5uTy;@%aTksd}GD0z6L{Q@;vzeP;%#n?NkXR9@-aqTBG2`^Hu+*Uy?~ zL#K%tN`kPcurQ5)ShkqrNDfZq1>xpI=m$iOuij}ycw~GRz1L>SY;`q=vr_ZDK-x4f;1B;twEa)2*f4AQtYN>fS9^!Km?4%|89{INSyN z>mh3iw!L|u`=p4uMh?{*$4X2^9Q&6w%Jh3UIZe`+Jn#uJ)tDYWwP76}57p2M(Es9u zIc+c@$820t8z1_i_*v{C(Nh@O7|1T&rQ(I>q4hSfgb|b&*6PHmCxTy}%%% z^V=(RnRTO3EmbT>kh%k--+O)CvWVM>dL%1z)g zFsJz(AG(VZDQMW>teCfuH1{b8aLq1qc6P-NoIg~p04vcUZOlCoY!g>ElgnM-nrT%J zX5-?&d$c~L$w#6X_l8E$CbX)Gzxf^P9ayWP!lMz%FzR5Hl{j{9l7v|ClA5a))d{g4 zPwu1_Y&{%i+#hkwt~`%id_H!wVD_3`KkkV72kg&+JLbx4h6yH_C?+Bjy4<0>)#*Kp zA}+z>m7e||a&tKbWaMO-S`?@Es7H(01x)_(S&Yi8&R<}ur1Q;g`=IDOcNw|#zhFF; zxPe$c=9+P+X*!nsCePet5kFO`doLjNNRmXmz%D(HGi6Hzp83PoANxO|;T|9b>VKoY z^uQJhFfvI@!7ysd#IcpxJrYd12Jp%B!+_ZCcV_xt$5WOrHj~*san1t_wy}x|2*jiI zW)Qkjjl7z+EqnR%moOX0!wQeYpn#?y_>>^Hh8K>zfd@LOA&{#pEn$n6zR=Kwg?Bo?)Z&_Xd}Z zT`)qfS`;U2-AVZH0j2j#ZtgJ&J3D*BmyPH4&B2lPgr>=TVhN@=kxZ_!UmNvCEN)}2 z&VOQqqy&=^Osos{D`MED26bS3oSe!mjw_|9u9cRzxA*pYAqw2??I$*Nk8$1E0)i-9 zwCQ1y&gM&_X{H>=obORnlUlGsFcVWfwz&JWA@|l*`{8hgNhXAlVLTf{Z%Pk`@Q(b` zrzAA|R;XdvM%emzNrcPB#7m$kh|Kn<6?o-J%rGN?XFqw`>su&s@~j#L2MmT06BCpC zr2VPKqiR{$`Ve3ApD*$ZBOW;<_9lMlT2!U9f0DPv=l+AA^*@4x**w^1slby*x76^Qj_*SEJ5+Dib_I7c*|AN2dq=DX$~N=5fo-`o)NQ6 zrFX-S)YQ}*NHAcIs6M0PGvCqAtZ}M4%T3{Xd5JO0qnK&RBtoKiaB!e`mLl#YnAYwGNc?}(q-uk~0T=wX}xor!(grIJ@w_fib7wf_(mg0At8EpR6635CS@_%>DiSU9eYMX>FW_{T?@mFv?b%d)av1O5 z?xa@Q5dZe5+5Ad9>IUbhiS4+=J_z0YV z)p`c_^Agx%*yZCdq$>#vSs*tIF1FW2h;QGUZ z-@UsH$3U;4ADm|^EaU7>IV5K^&3cnZIdk&MaVU8V5o&5tYSD4u?#2+uBqOB9DU-rC z$(I-w$|gR!NQ0SWI;~=#>8STmNG}3LOc%BY@|dIF?gl^uLf?l6*%^589WVjka!vf; z2P6)qB7IECCtGE5v5^0HAQVuv;60tA0M7%+7U~4759i!DpC4psXKZKVV$*V*ZnhNC zDbWvpYZs9`Fx!Edxg!s!T$$c-|lyPBTO-yzb^BY4&-bEjCBFRBNhIUVRj} z{YJcNt%js$s$FOGxn5LIf!wC&XR|)i`m?i@^r1#|`=F3_fAcUUGs^Rkn)yQZbF>>1 zW%4#Qh5cye)(js11g`$s*j&KtJHGd42>gj^FmTE6TWz1*X3>u~XS3KlNl8eMWv|77 zKcCVY*g_+z&Mq#PnZ%@-B|lX>%s`PP?E#Rk$$F6{F`!J+#Jvh3XQ!w6iMJ&VA(@pz-R?pV+Gj~{cALhPrjqSj-MP>2xa~Pw{yL`CfSM(&giA;|>moB z)(E@Jzoup3@L7i|{JIYwQdyQetv4-dlK%^pNgPzcBRWA#Srd-Y6zeaJCSbS9`&q8dzZTN;P4X&Ca)oOm??F z_d1^gF6*!0f1I!4U~W%**7=ujcp0YjrfxG0E*$gKny0tBqDur^(QLCT-FMVQHTH!; z?sAU^`&6P%?ymUdZU;?Fto6H_^L-(Z`M)&$3&l4202)xkF#AJ0cTUl@ z5f~6&gvp5?F36nEVKexm4CEteF%^#hg?_Lq{adBnx(06rdBb9tod@T<--`N3tR(EQ+<-y`u|un*f{Y&KKxWVn^haUdU6 z^FDs8S;pA~SY}N>A>4RGweS3~C`cNV`*wJGi|#!CEa^ALxdU}g{+~s_AoW%YlMmfu zhVn2ON5c~g(X}b`4D3G3g(G4uetZrXj(sY2IoNT2AojTcvcd?@w_bi-ltWKzjBd#Q!D9fo`u;Wid3MogI3bfpv+}vVWt!(zBnZ| zG#=4=ez+DG-1_j_Xo03c7$WkN#?$*Y5R~Ob*gwgtcLu{0{*OZV;BR|7?2WsZ3&0k} zFg`OPnUv69)B5{fom)S}P2vE}DDS|+{&MfncXu#iGTMHCjn%0v-MK^9ktrLYW!xz) zP%0e|8(=e2!;M8qb9ZV&aa=+1futhZUPrW~RpFZ+0$rU=*i5=Y!YF4L2W;KE_baRB ziP;UCaGuM?_~~}zPPy~c(~wub8)HHgqW=l}1h)eJeH9(hePGN0eU8ad?^2Qb@Du-U zK)qipGn+mB1@h}_RpRWiRwf(cCEYzg+W|Zts@JrfeayisgnjEb|5)iESnTWoj^%KXhg(FL12oI{Wy%M!tu4oR;+lY8Gd}WvQe{*fQaj+`U2dGhme2kbyOw3^8Q*;0Ncu^7YY+38IeqPQ$ z0Yd-ATb}j;r{W5>Acaw=j}pV7Ym`Q_ntVw_Lycx$O3RP2{xw5@>Y$rSHJ`~AjRj5M z;3`SC!Ypi-gq)N#>yPtJzSmHxq3G3HX~DEK3Dbt$tAW10jvh_ceu21u5QTr=Er!_~ zEPn}G@BwS$oY3;-m%li=%*YVIoBWL(ScYQjz4JAGohkTvN$vbF24_Wda)~TSjo8iR zK|@#-nc@|XS9y!urw$#e7&#A^ zm1IF{JVoT6ogI74N^_p_QMv+!*|hK<5lMn$8pkkOOr@XXVV-tBOq2DEm}Ko+j&ggW zO3+xi%pR{>rh8Bj6B9Q$Rv$Wcoa0+fRj>fxPvJ#+4+Yvk`_Dot_!QeBPIf}L?)Ra1 zqf1hef)y76ICsWcIXNit?l?|%GqL$q0y)?N@c{{S`7@N6U%6hTUk)Q(k=^riy~-*Iu3_*PdEb=#`KwJ|onpXPB+eQWXR9(aoAgq_3Qy6!J3-f9jl zX$RG+M+dOmr_cctEZ)zkFl#`A&~1Cu!+D&o%&fs$IkC(ecL?{+aUBxrZ*Dp8as%^% znX>7Mi*+#1jp4SX;nqmC@ebAyn_%wq2U9Vc$9h}yUpG}k_^k_j?t@cQl{PB75M+{O z@!(Id`e|p(7sNA}T<7X97Aq zQOx|2#O10?(FVKMTR_tl%7c{bisHGWEei)bJMN9+il3EpDR_&XnmrjshtNc{2I)_bA3A~ z8_v1~!4=W%V)AX56mvIva!2K(>BiW-u?gwFy-lYy?px+&d-z3MD%jB9{n72z`~-MW zdZq5`;R8*fBo0!;G3JV{ zH5W?`MH@Bxk~|qhX>o1IugT7>ndj#WzvCI6yiL zd(xAe4iL0#UR3H~0Dzr>f)o`Nvd3jT-QCE5So;a{^Kn4jI&QA=g@lt~_SV;aE;G9S zR+^cZh!*Rn4T*sf{B_I81S^gE{3^hds^!1AD>G@3_l$Ur%#S!P3D*lJym%nSMSFE|BEO=R8WR_nrFXG-@F2aC z!S@mh@^HKpDlPwP0@}B)_ENWp6mV^Y<7nlZWz+#Q<7EQxj+9;N;Qe}^!-XS?+GIm7l4n>lAg0IavEb?aBP*GF3lIwncJ4F5lfW`*^{WmVfg^3U))=NH=+l-X+rZg)R)*fGKjxs+O`keIEBBmji|-d zk|kNf@z~$n8yYa!`Zj%jChYsCjv!={WehCo@gL*=fSRjw;ke=u%b#gdK;?E8PsrB) zwC212qCY74$tPceJ;yx-Rn_m&l2`dh;l|DT>zv8u(#JWm^Go82GCX?N2m1#^&yVjCJvc&?{i!V?iX5XwZjz; z3;kNJ>+BDGI1B-RnNU@rBM!@i3on=6#fgF&%<6=exup`j6hM6QJ6jjZcAf85b8GAu`iyyX!g9FY(&P? zbFY2C_;z`Hu|p-{U03@2^U)aBFaCvGCrlfF#HhSCjO`}Np|<0tQbpq>&ix75*_Mn@ zH{G`HgvPS-^ZX;6pZ$@Vi!%+!i}(QA*QQMTzl(s+ zzTv)S1r~5t5?)rMM5j=Xl&qZ6B08&v_a?3M<>i9qrFz?8MUFL}q6=U52h_dyA%(!n zxgPinM7FtlOCN->+-7tM10Kn)f&^`7 z@#6P(T}yx}(8c_OO}9*Y;`F_*@P#t%69>4^1)ax$!qxkpPgG1=x5?|Ah)-3j06vT4 z)Bs^@32(G^-nWRDO>f2>puWwy%6m-Od1|Q@R+k<`xOFrEA};Hzl(@J8q|R7DFykqZ zH-BAsja3D}Y+mxcx$F&GbDd!C-xEL>Gyt4$R`c4*@R#4U_)okS%Zpt4b7o`$oYsE7 z{HxH`h@%BV`gUrig-?#F@wXkO{V%K5ee5aNl4{&;yU9Wu<>B;K85SlR52-7$!m&kTEutR0o7YGVvRrtv_+t^)Hnq3y3YKX40$pG+wdP zr}Y(4XNEtP>|%T9UEsGqT7ZLt({_ia$sct5{YVr>2aB0((Z>>Jx-0{)@4e)=Hjzt; zxk7a`!csLfC-uIoheI<`wRfcfY2+;PTGB{z7V#12oVYv1I zg-D+?lh&*0sv#cY#RzPQJ}@_MS$=1}bx)|E!~*byzOgfMGVBQlhxTww8#{(|=o7!R zqnMuhkoPTq+kK*&<`^iylU0cZ6t#UXBnZV%iK5D0g!LPFzvD7xY^=P4xn363VL}o! zS{P02V^gNcOl#PrjVkT0S+Vp!yPvc_=d~l=bklxeQMskfPXHiV5TY7Zlmmt9|xmx)xxhY~bH`~O{fRa* zH0&lZbf@Lw=d$u$>PHx)Hxk6So$*aa>E1!6om#k-lz5|KLFh6}F{2HH>s703WFE0a z9$D~&PlX2KcNGPjG+EUP%DZdlYRHJwQtTV_DmTpydxz~uS*Q^f77v#FfFVgO{b{?+ zuYA5+r6yUx*5*)=NzQ5emTr`awN{E%ksg=s@nX^!I&Rf?ydSoySks#BcSc>LQ!>o% zEfpa3#P7G5&3JynbK)N)o}1{i@``o7U<4<&ZQC$7#YI#D8W^5Bh&7tw>| z$Sb>UPpLDpP`EkZrT|~d-wQJmv%B!QciXrCpqX{2(wUf|9`G_Hyid@crWR_U!t9O2 zj#g!bsxuptu<31pxTG9)edy!sW1-w#;jAjbNA6*8ZvAFCOaLWc2^!^s_D2rn)zgMT z_|~~cz^xk3gt@N%w8a;yU^Pm7G&6@2CM{g#o&wjd8Hn?~;GPefc*Im68!)UB79JHo zjxDBF#;9o`>sezrm8p3Q`+@+OY?sIou}yFx?2|ixomON5VL-%(T|`8`)rYbmPSwkN zJvN);6IW7&Uy6!Qe5lTPlvcN;w%~SxHS%fnC~Ib?G}<f*j`pVr23AV(~JY6VnS@=%>o7+~1^2qesh zzdEz28NrT6JHj?G<Bxx7gc#We++i0yafHt)MF{rN%4 z#)%{`OyD7eU8Jc<(fLj|jFvp71nr2!IK|+7BpyFl8L)NMA(4IW<6Zm$F^=A=FQqrY zR>{Mq>TA{#Jm-jk*(=(a)C>dt4)@W<@~sW|1)0>39xjS-@0+~^bN!e*=s!}#{ zW*6JX`EU===J;Qa&cZ@s~SMfmrHJ z2eW-wqK2D@H}foYx?kW)YoevgOkWyN>Ll5h>Xf9Cj#$u36~j6b%%))xa{wW6pbDr! zBc~$D57U~iu`Rl}iOpsUA)p@?2z=B$R)tl#`MocVC#Fbrr%v$T)IrOp6V)yr2bd{RcsGD{P@V`k>GNDQ5F+mXX~Epo z=1n0Q!CAThmqTr`=w7X0V}xX>_08qaG}M0CW*=V1TCy~1OX_&;U~?m9mct;g=(vO? za2>n{pv~J3Oq>1qp2YoLvZ5KIORK|pN$ z38_&Hj}3c@o|?CHriBC8x86G*lRd{&ise)I_%LakxuAok{{+I^><3JG#nUSVb}*%L z#yc}7Tl0A!ybi?xRFnS-DUpZgoO=?dCVl`(nA3!GfYU^bX`^BrIbVSeTvkMIDdE7 zVU}_^YGwysY0{QyE3jS8mCw4d{!hOx3Z{YRmy`P_!0Z^b zq<8Xg9){VOoTqu=D>EIn)ZJ{R2ZVA8AJ3BjLG0u-RLa?{}AdhC%rp-7i__hGtaL%YVJB$^N;{v5FZe6D< zU0Kh{UU=52Fvk8cvTwfnsJ3JZY*gjxa05G9eb!nz(32w&S+CMwnbRFFd1kK*vgxyT zdWnP4mTWg;W7!_$IT$vlwd^wWEl#BXzp|4KYu;Xb=%u4g& z)FirQO}!r}Vu-C?dVATIx>v%35#vY?C97XOp%Qx6yGX4M5s$Hj{;nRgbhkJR)Bb5~ z;vsxk0d}a24(Gy@W5p#wY0StI=474}XjdpnwhroJH>1Ml6Mbrk!`2 z4z@)bO1};}O_Vq8PG&`C087$`)sG5E*~CT_X1H_i8~94`Iro8bq6G)B+Iyff6}~tA zPm3T24B6#Axc>Yxy=s4`Oey2NqMxZ9=WnQKf-7njC^CVe&E`So?f72B)njZHh6oco z+8iRiMJ3>Nyt8k)!UB%ZN>Mtyd73?rxaII#ym%Z4zM5%XJdc?PG>iR?3)`Ap(I*Db+R zo&;w$rF7A|(Hb~ytZ*83CwlIBOnYq5WOh*9o3?5%_P!4*NL~Dhtig4yaO9)J*&!*H zepCLiPX-AI$syEnQJmR@=Hq7d-uQF-wW`w0KTM6&ash?3M(%cW{F@GUZ1}Bg$I{f# zrjnjCZBf$BSL5Zd97o0&qHku$m5}ETokWe}#H%nG)cVqS9&HBGi+kqpetz7NecO7x zx7WqM+}d`2cwJ3u`}cUpzH7E)XS`bbY2Ie#aXU7{s(u-87^!8p*(kPE;y$sZ%Y{{( z?Pd{fz3ak>pT*IG-f}m(YA6r)L^(r}NUh?8gsV>#cLPKss(y*}aSFZ3R)`mS5=H%3 zj`En{U%PBz?;z+Q7GNpf9HytIAUJ#13`&$aRBx@bLfr8o>Pg#Go<}x~ONz-#0fp{u z-dE@`qZO3vwe`dS8RI$J>gqGWGAX`@W$KwM_2EZhV(&0Qtl}T>NDk=wS^B|7iG?bw z(Z)b4m4edWTWerCz2Bwdtx!+G+}r5zTe|6bd1JEXRZ73Aa&bcb2%Y?UssJ7C=j>%C zlI})X`qb3G7Tvy&D0{Hbn*=vEW<$R4@v1lg9Gtc)s`|^eJQGsjE$2L?iRH3vBs}$t zNop*M51(x2I|4eqBqm19O3*IS$bD><3^+T0{_g2o`{?l!{SNtxcq;RWQp4mh z9@ef3*d0N`$**x{`dQh|gkhmYyYGJ~uk|i0wzzn{H=B;@-k9v_%DkljgzpL34YI{} z^C9ip^;>$ApLLw7C}+C@8=OGKH_LQC3%829cN}+lNp=;`N}DL@FxPLsoY3!TYt3Ar zDzT&<5)I4TX+D|LXnU)RB>Lb=T=s0J$an7MI^Df&U=FE4OlFrK>TEsi#TpOE$kB}b zfD!v4?ef?yIPRW_({j9U54+dwr@RM`y^f2;AA9+ooR{uY^vbKv#UJ_~`U`uZd`x+; zkiL=IsupEaDFI(trYU7eYY=# zMO+W5cpd?oRSOVVkQ^TbU8*5_)z*DHd!eAu2EUHzue0O)h7XL3Xc!x={HeM)XfArg|mX-iqayJJD_zdo74jK3=BFUv6ecsQw(#!c|2 z)oyk@Lz8HKW(slq)kDJ;L{84Qd^*4D=Hgk+64b2=56_W1k5|KPyj4T^cdKDAMO38^ z=Im@r!3g73?j<39;yKS)gvZW{{Rdly=O2gAhh|#i&#^fC2F&Q#vJAQJeW1Vny1x#A zu&{Zz2#mJboG+WX63C+}v~?ZEm)sx)ZEbK0T!{bv(NPH+5=9S%8#cMeL}Wk_ zFvBQ3j@h%Okr`t8`e!yfNvcf6WYeO7DU;{jqNLOV_ufTz+F-yMurGzviXrSb5oC8Y zk~`-j3sUkq-X5u@bxkWThToMTmel^_`Mo)WjA5#D~b1~b@ z26*e}BppAGmp-2vj81E6xqXsx#TJ&dq-L*_tgYfq)~S(7{bGFuF`AWG>48b^Bo3l7tQN+Y-gN46>rWx3jQRGX@%gN*= z3TXMQEjZW2bcpQ{tNla5k>A`yDz+!f3M{G9y^z9UEkAAH(Nj;0-pBENtmy~t4bI!r zgkrl8<-{`m>Tu7))%H{X?4x`uJTIZRrkOXg=dn@pSgw?Wz(92xKxCeaI2cXo>xPdN>Ie!xLUWH~U2VxQY z(Ko4mWB+CghR7l+7>49q=1RyAWkUQtJ%>kBpol}fdZgt*s(BmBL16Kfwcx--ha6;O zCm}&o>Se$bn{skqt@93f+X(aT@BEB~q3m9s>vN-?jJPQ0lUDZVct~&jT*515a1&QI z14;`@pEy5UCnJi?l=IH2H<^A2MmzYdSolrdr z!@tzr-hLq5n+VH)o&GKgr>hMZ(udq2xF)>|LKhCtng zAnSPHC!iQ#?-cr%nvCjOM8BlqrBq}y&11u|^12TPS)~T`(qKQy`rYyF1`~eOvhPOo z$P7ooe*E)rY3^R7)1Pm+;jKjWqgOp~IoO9s5PwoZccA5VL-2%sKJge?yd?83ve5@! z1Gv56r_bN9u`-stCUg&Pdf=az#C?aoH%mp}8CaT=IU{Wp(Actmv8sOrD;Ba3eHMa3 zVD!qUtWIy{6FI^Ow9=say==asZ_ZV)k6eQtdl%*zA?R#~%zYH_a< zhi}HfuJd{1?XLvT8n8Z*#}J|XBmzf6+5~OuSAE=_z~i;q5Ri%Kg{*)5ey52|t|Ya! z|NI_jIY?jG|FU@M6s!Nb^4t}Y%p@9}iL#i*NOa*{_l;Ax&%+=N;NX79n@f_HX}#-=^ko#_PO$E+rVer{&%od# zMf`#0zy4L%UhkRQPcJr8OtrR&;7+bx_$BFI49Y&gOv8Rc&B^DrFKaso9|$Jk$@WXw zJ!l^kx+Ei~(0&E`EgF6xmDl+WF?ROHjUbA^yG5=Tdxye5tvOYkAtI9dO|_I13xBN( z)P=NO-3F1yf5Hxp_jq52fn4SzsGuB17u*ThZjs<^CWD3JpDdlz(Bv&5iMaRD==7t z>ckSEG#i*GGa3lDjbl|^4#sI4>Hm@9yg#alhl`tGX(`!9vP#6@p9pD7;x{Lp6n&yG zqSGAg{td@zZ&zDF>UZekV>EaePJ9L=oYLQfag8hc_{-#oYzaYx3_`-O>5D+Zdslxu zRJ-`X-@h_G7l%1{>AuNA`sFd`=e&3>MK44BvwTM9@a_<1ZEo7>1NCZavOa5knSuHQ zhZ0L(i%hoTB?N(tLj;kv%9xS8Vmil=eTvZLF>FAzmYhazll^*`Y-0?+kIv>jzdI{a z!at}FbuRaS#E?~CFW^8h_sH#NAoqVazJE$=JV!6;`+f@Yq}HeLI3dPQD(rhM`uhIM z5$L>eljbA}0>Qk~_KT}K=kqnarUGeJ#9fMr{Eal2|D))kkN@w|X~@Neg?WAlk^{OW zsK_K@&|SLxambu>I^WZG2y@;Oe7(>^Cz_?p1>t}B#osDc5Q0fV$)x&1T3~QsBHGO~ zBdx~cz~X02JDS%;)D~?(ThzmE*e|?)%Qhsr#7SH)nx-&V`p_<{Q9j7`Iw8e?l^@s@ z=X{=OCu$b{5w`l5A~G(zh$jFoL-Vv@Ly?TYcG6rN6g!xq&^b`9wv zzMmg$bsdd`M}+rik{Qj{+p<#n3JNurS@Sjj%JrYd8N#VnMPsWLx6$7z6u8~1pboA7sM3x6!DF58*_Cr0@JXKUDglT)7}G{=})o* z05adX8LYJMeu54H&khGAYpUqpTj*gCfG>MNt;X$W!0JK?U};?W2dUS6pvyez0UI>_ znlK{Jlo{{UWB;$BcUdf-@m$@@U%HEhMD3EV5uzY18g>-UVxws2!K9GOffB#{H|R)A z8WM=`YI1V+Wji|dV|bV2WDj;Q;r)AulBy34$hmQyS-*5A#BgZ8o3XnG?LT1{xCwkm zK;1ilFX$|BwW*1P-ueCtsG4g_^*nD5vEEq?zU86dZIj=oZzHqn9(x7CJjHkG9frPo z38D>_;i{|L2L?1H6{T!oz@!k9f|1wzp1x+{2%5GEg>YZJW=Fs8jqYgZeJ|#j%hj|? zlmcVyXswZv2(YEu^IG2>@>3P(Z!)zmu�%Q+lXdLTm*6AoGyjrod%qJT>6ia=%J0 z>01)RR8?bwz&TxokCGFG{uWhQlIKAeagbhGxg@Yj98h6T-C2XCoWOY?Y{UT|rd2B_ zdOzP0f|SK_n1{Z*&*F3Ab8?`@O3~4TB9A9QZu=TubFSRE^!YiTWOM`0wAqVd9~~i* z(go+UG@t&7XbMe~^pikmxfE)n$Sz?NgjNgsb0A-g7aMj&Fd-ot2$yTigEIrJv>fXb zf48`0uIJmN!%9Qi`IAJ3*P$k-Ah4~$~ zaS_sgYCjlD^aTyxbpUk;l6|m_csb;;Imw5$jAD7&CCSJb01hJ&PFG8I@q*m9IWC^! zcUXW&k@ep=O9ax%K*=DY`Cs1{eO69OOZyz>_F(%isOwDBuYKJw+b$)A#a-X<@^3z1 z@2g&RU`fz#@wQPqi-|Qd%e^;_!S&BG^1TZF+VPe87X;4Oz`wM{epzz4eX-+yfCIs{ zi3L=ax8dT)xdxvIq0hVXfV#`-e=c~i9GDfyEMEg36J_vAW4;KT4phTw!%m(Dim26} zI8t_A1tdm<@}!X!=e8ZK=Cz4>=RVduVZGci*bomTAOUBTh((>BYz%N{J}iBB*B$;$*yo&o zMBnpdTeYXB#~YNTyq>K|VCN1+!t>GKc#6YMup7UCJl)cFsFjGQ>lP{-#?F-SYV$Kq zy{7v`2HCXhePNMC4uUs&)=Sdg+lTa^@U5nM0lm-cD$k)rB<+s_yK^j{;rT=J=@+7X z;xdke{?s0Cx;G#!xDihmeU10R@<#_+FI8TdOd~;6i2z}3Zj^U3b+kGof(HUx0Z8sL%S`td}+pW$BB0Pu#je`4aO_VX*(4Wj!m@79p?~$z7iXSa% zj`C--GX$^pOUx{fQ5yu0Q?nKi+oqwSB@p^tjq#!!FizdWSq98*xWM$id z{E$|tjB&tQOkO5jT^MsLMA3Qk`fUx`wM1<3^J+U;5Tp3r{rwvKLYM)rKTE%R$F^^X zwX97+R*0N3oT)q5t-I&D%w?S|Hqud450ROzpblYaJAQBd_=!Abl^PXA&Krw?mo_96 zZ&cqV^CsX7bc0-td$A*zl|GfAw1GZ1kxc)NOYwX zx*q3j!hwe2QfEV(nX6hBe1FfCSvM-g8k2?QGszDjLOVwumzxl{&*7fDT&q9W2)CrC zz&3dMwN1aeveO0W)sW&cCWOQ%DvT-hW1^&AlTp;5i~^LHN`8y^d$ZWg7MOh;K2_zcd7M*=^oO9vrJiE8w5WoT;!H zf(3s`i~|Xl#U~p_8c;&k7cv5!=kqUgmuZFX>(=GwX5%r5@0B)TAkW{29?=zPtOPo| zZMl!PfNA52>HvJ&U;5_fk<=D1Y} zw0sOS`)z;NX>=1)smI*O>Ho;W&xCkcN^X@$KVgT?1Pg^Tnp^=5gSYORo4gS^)%I8= zW%Kd6E!VrCRniu{&SX{ee9++hh_3zPT?T3uW0aBq1*fVf+!N(1-Ud+`RG^j3EEhC| zu=w^m8{mxY7`)W6O0Kn^wcy2Yk-+yy+aj&v+ANfD;i*&P-9eSzn2K{|F;;|q^W>8I z(icrDcBzKO_WgIwIr?3whCn+hQ&#t5B{jw6Rl?OAa6D8tX_vazn91<>7#c{h6*1GM zz}SVeb;qD7QG)cLlBQ(;pexOK(@L;E20W-zda8lvkKq$Empm28V2wd4AVJJPIik7^ z1IrEt5PNW;ce~h3zadD-%4zryJUF%e^!Y305F2EtA}IRfvo*U{k+ zM1Us&xLVf$`-Tz>IJv%F!$Rs*1FEwO{Xr|l}hMK>%@Y(t!r4&A^ z66~Ru`C_!Kk(iCJL@1pJ0jv8^n#M%(aZO{&R{7XpkSGVL4(iYImiv<{QMgR%o&Kyy zMaD(QG{d-5clz*zgT&nx+q*t8-zc`EK6Fw1(19QfO|3akc%bp_X#5a7Vxgu5aHF1YnBfDFTxge&o)+D7~NHdt~ z8&1)qtU|_1`3qK`6i0?6=`{04LjK)q0(Ez zWLese_k*6|BSQ{%4;Q9gL#dV)7t2xOW46!q_;Wll3ea$rsmOkPPW@bvUl$q?*pZfU zK+6#ALfR_=4>F9RUMYC(e^w$~=!p0l&qiB-dz8h-ATf-r{kr2K zu7UB6I*s8s|*?& z^RHJrc_L5hc`R0E-S)f#)vdKZs zm7o@s8gt6ZzCV6Fix0QQC(if9gA^i~qEWkBal~w-QF2#;TXetPwnwXvd<>_2$lf4R zz!c5C^utx-?n$dg_O7c8n-T&=o67lycfDe9WQNKcP?;@?S(EIy3q+VWWO@!9JRnP& zme__DrTd90eyO-}Bqdn2t zw8VY?&fDuY`xV|%zQ2REUqyl_c)SUe@eIDKu0b$i)z`sX+w{Ata`Ep8@#;!VGAv-m zuxy(A{(G0*_!;H1Tx#HIb;`yQHJTJoIsU=l;Cx2n)42HDW^*j@^XJr6aP7378hky3 zJ`>Wgl=42^%i^orSz7x((bL~o!XX{9z7BcgEB*9dm1;NXe4S&U@97@4l7a%V3tExT z>fWZPGKpN+;EX@5(&K}~gX{dMqDg0uOt{`Xw0i_8)qYPQiisr0Os`;NH9Yh_BcAr` zAePw=y43307-yS%@~GQ&vFft1-W`=YO`j^ara#;4B^1d1quqAABm*2iTIp5(K)QcW!spy!&j^vs-->M|{hFLf2ck^^ zzd8yljQnb8&x#N-Fhuyg66IhkCOhzdg6$jqrPGtluY>Q@?Xs@1@y#famXaO9v^krP zEF2r&IS;S>mW(a=L{6re|C8nH@Hrz?i;h`GQ`ZEXfebGHp|dT749g_X4uZO4K6M8?eY7abl6o2URQ6ia!`l33 zaOc+*{+|~lHPT;SeF0p;a0j$lL^@0^%`{y%4;k&k(gQas_Nes;JrE-8%1r0Pcqfs0?C{$1So`&`dUIn=q>u=ZoX-6vzW2j+4z zW-^&()J|Ed31J;tOBBB^8 zwrv*Lc?A!>0zYBF%npYrxNxYOP72hY@vSzge^Tljvk<#E^yzi649H_a8e&H2#V#tM z;QpSN4*91J4!oM%ZIf-%^eN$q>DUBR3c=rl0M3031EpjA%tHec$&I{KTH(sb?{?s z4*Gxe?uCxJYjkMs%>qh!^ z?J8EQ^Myl=UM|KvlQaanXU~YxzY8sRyx05NkCL8zxb&vPN44C4uX4-li@{rQE|fGK zG{h=*ABng6>S48R{Lbd$I&AwjzLE!>lFKHyN2dAr4Ep)0`4nyJxxPR%t=KyloC_5- z2GgJW$)A8G(totjicA3`b8qAauDW{o`0&CMlCUi;LNz2alU=Yfb{9{XAn95Xq1=o1 zG*5uhxscw;hcI6{zX*fg?B6;Qg$;#^D;*O^Pt#FWCg3z=jPwcD91Ubl!4-4qImnMRZdjTRVoTa2!fpou z<=f?ZX(5R9L`eT5gaTCp^uoP2cpa5T=nB(jdQym=y|*uY7x@&krFL3ONQRGs(m{U6 zUo6!%5sVgB{ZY*$e~Y{1kS?&S1+gj;h*2Y0D4i&wG&Nctr>LP%nIDAnXx8}WT{7?c zN4fMUq=c5E9)j%BvYhgaz)sOonegWUt;uN#UCiCk10!*1p#vBNy}4g7d#;A69xu*| z$W|IpPpXm_4)x2Ca9-bN=^ic`Pn%5dRbKi~q#`jqAcY#yfqiDU(6u762lB#cusM+U zD1pD0<9qgs<@=VXP6Q6)y1m-lY*a54VQizSs1ceeZzijBL8lLtcqMYt6d z5>8hdHo+5IToLKAxXyTSxcVmfWV%X9A(e0fREhPs6frxXYK&5cUQ|%RA-z|TX7GL| zygF~2veS^%uHHJWH#kz0wZ0nHHAc(~uj^LQ4ZFSNy7mCBa~nPbrq0jU*^B?x%FL)S zhs(PXCEQXB)sU-`SHuZbjThaA;CCIf)X8`~-_Dk#J!W}S52WHpTx-8Ujyn4f_{=fb z*Bal0%n6rYV;*sx1n=Ar=gSh+&?vsbzo>^8LrmJ1W&ewZL}yI6Z<80XWoum{4gNx% zhto~t`)td9P()G8P3M&Iz67L)f$0y)kTti-%mYJ$tPVkIl(Gt8bUe z{;NtKi;A$k@x$R6^&1kMdA*c9q=F9=se9X8$BH+J=Ggbaa2S)2V;m;yHngW~f3=A|hwN~&|S@0eo;hTBH4#zW-)?c29Us&|=UuGCZ%?rgZjjDFAb(?J-gf z)_M7--NBj;inTEW0XqUm%mX)a!611=1Or2{`)ghek+s(J1mZz)i-$ zfN6&``i!d)OC)Bognyq0((s(!&JOaAWME<^ZQPq#o(n?fsv=ecb_iGWA>LptClSnv zWW90rFHD$ZtBksp0E?qZbzd08(CW`U|8AMKbq>WL4>Jv}iA5C)E0)AyNyV~O7x|q( zb4^YXU4H!KjMw*|t3~0cpt4?i<_6+DRqXQI-;!a?8l_z4f|((92IA_0+r7A$(U5O@ zwEYFMLmed*bL8`u*FP^SSbjQ~Ant?3;=T}!Ft-!wgJN+F46i?^P}qT1V2M)D_YuyZ zuvd(fU11RF2!?G#ltzy7yE$g)1x+u}H7U$0r#Lf-cM1!lZ$~s#jN-RO3Rj2G6V3(LngCh|T(T6S zStR0IrYKltz5(Wn{%A}nmI3cJM|_mndw?7X2@WkE6Xtq4ra?d8w_s4D)jfzN%$pwG zv@mWl96Xu<24gNXLe+i*PbVCUlv7IC_rb(_SkTLC(TmZq;-R>*@%w{}s3Sf+tRZ9~ z4WJwfSp5iLoE7cCd}l@bmv2fV4!4BiPr*v^1mnsn#X^L8DpO4)>a03fNPDd!+!}@q zz{%-`en6O9HA>N3Le$?^%Y9k5-SNEXun(DVNpT{FLY@eI)NiG&@^=;fsAA@IOt(7S zBFi3iW;$PeMc`@7LX=!cnIq8h6!5$8u#7Dhj%lJ+W~v>u4I&;`tr{PtB_(yF9$@AI z_;CxJ-(^)Bf)8dPFw>KD#OmQgArmf&2~;JWzj`gcgr4zLzBAbdnRYwMVq@R8!;J%U z!{1q>#iSmLIGT*CcC~Xw5XlgEl%+eriwAZ3O55ksN+~s;l{Y8MV&`SK5hR%<3rHBU z2?!ECQW(-I&gO)vCgACu$i98=PVj{``=)xeh;)!#!qt}YV?es;La0AbZfhowqgwsP zJZbHyp3-dde;g!s;z?;rFw+~qn!N*Zv?WIP5{0W6uXoAK4%$BBNrWYI1jOBx?u%iA ztOLDoI~$_rE9(QERshxG2}l6kH&)@OVVF}37v-qOM-eRI+unY?v%(X>K@w?|rYB&# zA$Lw68#bmfBLC5vazD)cMc|!cyV-|F3VLw*2=sgFP#naA#*wj;m;W_?*tB2n6Tu<4 zMj%3(#rj2T0xVaDBGG$1DwnH~qJOD~P+J&@6~xHj)8Zjdeowu)$35Dgk9MEmU$uUfo8@9}UEm%B55E3CTfU8u8Vj@+pDSHK8TUvOLK9{~?EX8#{Et2b z2ZS~M7Omy|u;n?o&2z7Y#QvF#Nu#Imh+0?Yg=2+hCHX8#xFDz1H4=MBpUPFz{p?z1 zm;~KEqa3e-KT3s2LL7sJAvq}-zkBFdU}v;oxI#n}95tjmJlNW&TA@j_2rIE@6<3n{ zH)wLUdr7sQG`^-LG1q+($)yxvhu-)61~K6+?16uLoTnS zE8$fAa*E-~@df@#bHfokj?^Sm`RUUNbdoQ9M~P>N6fqpvv*?g-s-ey}Er*BeK~hNX zF9~euBBtCgI!W}PrJm*6+f8QBWMX`sBehbx#zsxA( zdXMy_+&Xcc*gXn)e9Zi?6dOd-s!iqFdK2-yro!^%B5YH_J5+)#kf_e9PD3G>MsN)2 z^aPR$=QJ-B=MfP@L!PZiGLjcM1YdsOAc(B9-U3kpS6GWNTVj@s*#0^5m|#_JJfuz; zZHtSl3P7w2z!bC0V)1mm?-~Efg0zo3=h^}&MW5f>=4=mRzr;I-ojLX~3h~Zls)kNf z*)g0~&xv0^#dPn$JDM)6bn5~_ap4S?R($xaEb>hry0hr?={efFN8IUwz&{yyzYQLK z&AI}`vl`=6gK+%k-mS#>e}Hpmu+Yl=Aj$(aq>2w2I5Qya`wLCIV)oFy(r796mpJlf zOs?YSG+;7!Ng%?sj-Ih!KM(38{Aov|4#fZ?|E6F~Fjq-9iJaQCq8g^%!`%WvBG5sh zsPsf&4^o0ffb~jsw0S-z`ZU_y-J<9r4 z8hy3U^-&72434n?N*OcHSH&Zj6&c(bZZO2US=K*VnPt@7*6T1eXE7vbU;KzJ%nyMo zNJ;&x4Iq>_2|k)e0=1<06=}1$3F})6OmSN{yI%7{E`qh$Fg`yWYr4SmO#|1oI*y;g=A-l+6i+n;1li znn1BvR6pAivFe(a-z8<>Z!B0Au=WJ~*qmAsCzBYqSibp}FeHo$hRm8!iO1ZmoI9kO zzZHR@WE2fPr4$;?d%%!kC)d;#$PSTUN*A{g5Hl-lciiN(?>mL+ z)q*loRxy}Dj4cZ- zcHji2z?_?XoY`Y|gXRW!Cb-yF)>Kf!sLo6Ni0gKr*8&FuSg^ZwC!5uUBai>Ym9o9s zCxL%jIMl;HI_hF6o%rAXoqTT#Y_54j;;>d~{ANs-45MhzTLKm3h8rBkY1(BFUi^Yc z`gtI%$CO{#FLK7+3FvvRj~`ppRy4!4|2M3a4WvQ-ltlS^g}dUT+yW5#`0ZgGG3hut znacz(znGbf8?09!6Q#T_h$;Tj(h%N=Z{@Qg_PEI7j|%qqEj@>IM^c3=Xt7PX{b_J4 z(_*Js|Ae{s@0Vbc2CzotC*Cr2)<-cYao9V6W27mJ&j>{f1ZjG=YG zkcCUWv{laqxqf`*_aQYGi8GP%qbrO1;AVh2yh}nd?L?zS8_{k@fkxm%a7Dh#M>bAb zF+vQGI`&j89!LVPd&sNW{Whm$LVA2R68|j)yiUc8p34Z+Ax7-b>h=bujrAG%R|LZX z=R2`&8Lr-LR(;qR{}3}y1KzPAHqsS{j!IR68->p7C))mAM;BEni!XZp{%8-wApeT| zzK)KAIuVcoC<(JZYcKk@a&%7UTfFsqYI}eB`&7mmvji=DdjQTieH=vA=wEYf0USk% z$du`rc8uM@^PhS~F^cGRUDLl^-tHF@^Qok48r~SmEr;B;2bS zisA0r6O(X9x!(c0uCboV$XQa%LlSDN9X=NhP~y`dc~2{|qPO-)EmkvHsA5ojuO165 zyV)!qR(L9Ivz|w2~sbF+_UOHNjbgYfn61Gls7J49wLu`0Oi>f1a z6=w%<8MP9em!`ZJHgE2P#0KI0APBx){t#xg%}{AOe_!r8&KKST#oHba^upwEk(s;a;U4r?FGVP0;o5G(ZtE=yfq-LSV*FsY=+YS%`fB@rf+T~KZ%^tO+xDp3us z&z<*IO5hR;W0!9f$J&e5^*G^~G)(;``c^x3`f?dj3qzxIfh=xD^!uqhvZCgzE%#3L zezZ&rjo~vvRX#scD?L^oE7y$tC-S)v&nzt^kfjQ7)zKKy)zm%OyC&hsg-Wq39(g0L z1y>}4vJcot9jiD{T*tksDcgZe7d(b zFJzgq3KhLPm*P8Z^F-#al|O)P)W+LpnVY6hq1qX7FnuGd93)=%P@2cO5X=rwr{|gY zE&ZY3WrX|36jErx0q@noTbE!I->UEeW4(Ya)O~s=W&GLl-PhOy+IgnBN}2E%HW=;8zzh z(7k(f5@=xWpHw8R7hQ&j%4d=|A4{Z0NT0nxpg+-3EpRwg))_&Ea9gRYY(!=$Tsn`_ z`c|9?3{(AYycgl^$1KS1h#LY~ff$x-Nw$PxBi@@C5r zaiKntqgZtRg-FShQ$l@cEe1Y_eA!L?eL8(_Y*yOZhXjg}{Y}IGX_2|ncmN5sdSD7w zwPV&g;&TbX<1J(ub=%hiicV#e&;5M1C%_?M#gbUc1xttvD>|RN2E71YpvtCKv6Mq2 z)6v;V4lzODiiIysGt)Yt)#9=iOXvlWpY|Zbt40{cQqw0CHuUC+zfN(gTQxT<;ZtNA z{xEGeOFx%wVRzgv_(XfXG6eT}cUxT6S%$GtUN^Gb4^?-0 z4IF5-YQ)R}ax*$h8vDqK%o2JA5TEayb80+abKyZ@ec^W-Zdj+i;UaO{|IIHT>2-Fr z?Gqu(G)(OEsh{nY`+hbNClA|Ly{w#>NvZ)+UQ^e+9?sB9Ok*l9!ISYnRl<5l@{bP( z$CP8K3XmBFW5-PdPkxQk4tpBzZQw>=J3W-1lqDZ{>Is(LK;PLo;;nL6-A;tmZSYo? zHffM6vsUEjfxqzY3ZzP=%pJ-&ijRsoPln8(0vK1|OhGwxcb@G9UME%A@U%oq3Vh!~ ziR(6r$G?f6b-{>QqG28z@{Ru|r*V~}oTmET17Ca*nO&E=7KNZ+)U_aGqhLQQSlE zy6yj}Le$mCp9H5i4&HqQW6+i@J%&$wV#p#$xO+=bFlP_QTsYlnCx6yCBzQ2@oeH6C z$BLqVyOZpGx`u?6izWQ@$jg#)G#e4y$9ar{4P|xwZ(ESxZ>B*e2V4#dnEgi83~3G7 zA}=CZFzv~$t@R|GIhd_U#4W4u?P7L$B$>QtFRfEklsKc`p;Q&O74(=A;Ot<$Qfe5e z3MaXRs|Yyv+JOi?NwWvz?%6`_@s^WHdp7VoZ+yOkxt-dE;=ix-k&xSk4?#Afps;?-8xIRXelHhZ+ZYFfy5{miJou=8O@)lD@|2j{d66B*9{m2f6>x0zC1C|~RR-Z_w_anpJHs@!NcX)h9^*$Q$`b3f7 z_Rgxi0krOGwyRtXRMln&?b9@!6T~2dBacey^x)6-$NegXq$Io#Z`|AfV!^anML)A?7ovWI1lNB;4twwlhq)nGi z8EPDC&7)4lj!3SGRMme1Z)Mn3w67fOPiAeudWX2c_o^1x9K13nJY(BO8 z^7#I|vqsaf2>-LxS@7IA2dnm~Y)4Bez^%o{c ze|ZuIi@nV4-uZ@eIMaPwG9^(n#Dkmg;)+bT;_D(2s^fTsaB8*J(C2^!wQc6$3&#=Jk z$H`C|i0;Vkl{tOuzWGNJB=$sDa8n9R;by#kdYfgIy+0sPsp`=qHR@mCJn3WVF?5D) zCQUg1aIkFj#F+Vp>>qmO>GsR@R5g1c*s6CKrGE*wDp!g`&NJ4*@ao?(geh|T*0{YX z{?1wcA*7tlWp49daNu zM8){~UmPzLT198LZq3LUpG@kG)!P^Uu1Cqq$%z8aYs%4(;#JKEA|1kI&iC{rK|PuH zK9OmNp<9gxY2ck*ZR$3!FwB!q36X^Q%cF2yWr&Pbsa_0mmwJ8*iYFa;LgFZ!kY*}P z7v19amt=|aMP9m|)WZ?n2`0ijXxo9t2H<~l2_ik<+APY_Xud_h;PtDqwB)uTzl-bY zzpB!|?na8_Fs_4eqHLM`<0$mIpt|2l0|Ng7)o_0kVGpT9PnZ8j@2#kpw6A(!f?|Rw z;VUKh0%7l(W!&bVUY^Syrett-v9a>LX-fBLBa{gVU&VH%h(!XcxDqijaiv+`78DAi zN_fS@5_fiX@sxUl0AY>y&Yd{9!K8jb{JH%$hnAE<;CThi;w$YxIUK|WMB;BxB*~5h zBmc1L;dl0w2LDQJ&d{g%&kt7S%{{uIf3$?}{emA7oLkYfyd+vFfY38>`b?lfmOxx|fo^cbPP6qg_y3>?^6qzsUChzDy zfELX2#f|tClpa(C*JTC?O+B`Kku8s2?!8*R`k%o#ZmZ3b94d;(UIr5-)O9LJG`Y+REGfb)Mf@9Fu$;%#9PQ~0P#IhtOWw8Vn`nhFPAho#+o7@el*HbpH~x^ z0iQSmfIw`GE+oHieFRISjz4d})+UDJfkKH+$${LmPI?ugUrov9F5fS;wT%CT6fPDR#a1S@x@ldT{EKl=Mrj1%Kn-$}4 zYR$0_Xqqx;mcm*?n0oREApOhq{*0O|f|tZdIVl{7uU!=iX~*s5S}5_Ah-%P)FfCqkX_!~Oqp`H0;qN$ zyG8(l1=s_x1z|xbi>hR1=k&wNtpoUtRY9-guv^t8CNU4xcrJ5?Tp9`+^(QApDKqmY zqQ78!j89{IMBc&lm9tmAYp$+_`x?U8>Aa9)l%xNx#^li7#icbBi}t7&P9-3|3uU1b zG}y1x5Uy$})9kiJpDyEa}ut99a@toQr1JSzlRVv}--OxVtVD;bi`^44sdmzAuf zWc}|4;NBItq$C2A&r1)WLL{eC6eSUqP*&zhhwkj`dIO{zXK*NNp5Wz{p#3>w0_huQp;MX4fuz#tfvftt}9 zm-1Dpi*Z0cR{@R6&SS%5HPofuH4$;}q!UBEC|sM0z>Dj)AC|{r$o`jr>WU6%c8Pvt zL57xeStOfPA>RrXE{Z-C53>q@!M_ap9ZsqM#l!+k&Ztb_sziSZex#q=ReTWFYPjO& z@)+K{fAIL@Cr==zzNW}YgCBM8de4n;-C0d)bssq$QzI|RJ&79%u<_jZ$IDkYD2Lov z59~*t(IFv377n)C!Rn?S4y3FMJ~;l|@Le@<%ube@mn6W~1r zw5?zpAC!dN9Z-5n<(^K1ZtkSvqGv9!6Ov-4UZ2`eCB}28pNkWInvG3Yu2@ZP3dgbg zyt8xXVUxmSLEt#!s)usrGA9IA!GIdL6hg5C#6d+7=jd*NHZAti{6ff;Mj6#mief>l=Y=WnhW995zr}~{i?(} z!>}Mv9Y*fYiJ3LA3X;gC@9%l5p@SUtG^>zode}neGGWl3P4$pDJn-#W$2HJV_DcV%}W!>L16+rqM2bqN~iCZ9D+4@Gy%&2IfzpQ)rapCuti%GqCbGeXS+3cp@ z@s3v4h0*oN&dQgOSl3WrLw)DmccK>molshj*=xWkDCq}7X&%b;6Jq@=i>l8kW2oa< z*t8Z@K;k>(#_w5p{r1y}3~O{0xJ-d7bNlz_taK#S|EM`|@L&95`qG(Cw?KK2dY$g)5{GC<#MP)oE*tF*>wlT z+5&>NvWmA)d4=6c=@zWkJUFGJ4oJey>0~aq+>Lwd6EixvXvFpA_N+}O6&4z@OJz%% z#2!Bhf4q7$c2XNsm#DXirXd-`!RX?lC1YlDYl0gjq-otReUCHg`r^X^ld~Q^;d>l< z&}-8BnE<+B^K$XTb@)k3m!A$oE5PnHw>UMYP#0OPUI)zpiZe_|J%daERckZDaq4XeI2p^xCS# zxRf>y;unjDgu{CYhm!+i9E6{A{@hicZCT@wevsFfIQkGVIk>ZIF~hjMDit6C{J^J9 zXGzCuU`G~~7W5p<1QJ&`8UuF$tOC`J>-OQT7*Fho`138QPYhmK9gL)8(OE*U45>4& zWi1_EI=TA?DWNv>(Sy#56LvVX8Ts+z8)uVd}xa~dg*b(X;ZWYV^){ifse?l;?1W)r5m zn1X_WBs)HYxcFg4+KVm^2QI?Nw#lG~5F>`g&Vd2lR;h^Hoz3}a7pc3rO;SzR+{)l_ z-8Ov~YB=P^&~idarLgr32O!#0op1f6+?!QlAE+uG(a_x*PPj;rK;`e1zNoafiEOP$ z4xO?$_O{a4LR1RMr64;-^ks*Tvy<^>={~>>SxLS6Ck964Z@%ZI}Hh; z@6=fD@Q6pG=sP1GN=t-)S}7|JLTW=OwbLhS4xGS}!^zo1lRy@U>gV{-&4Yi^4OKk` zZW13CC(Jj^wE?t@v+E5cy&(Tp?9g3c<4;4F*0&4}km+^tIP8G6qb~L{R?tYxu*@_M z;2u$EDlukBPXw!!YfQEM^h2@C9O@ruRe`TLChimmUCWk4NP;zZ^>d(}|Wnxc81xI&@C(+u(4q zAu$JmQjcM&nc8S>APm6AJa<|w+R3R!taz-v7ZMGL;`;z?pGPQ4x~{(7`Fi_wZ)Uzm zviw!~#O{9<`<3d8zWs)E#-01I(|aAYDnuB|h0a7#<2{=_ZoTDi_!VFEjFshHXW)GN z)8DN3lPtUTiB{B5$DaUaM`Cj~Pu$5`*O&=>s>+(0rPt5G%1d7da*Y;S2{Hi{60Xsi zxj0Qcm%AwJx~Zwj&$huJ^#|%+hvk*Va%PGd$k|ppiSeIeioy7e^a1C_GaO zlp`m<4ak0wV!wI)CJM;Ar0<_TN;%tmelqt<&@;oZ4K$b9BKaxoDW6d1y=nf@2afQ_K;Rp-^H7iL<2MMlteZaF_--kgD% z5C7J(HRh9j1K^c=d$|R?OAjgdpHyXJT%PU@We;+k_?$kJ>_yTgKE0Eyu~F{t@0ThBt>>*C4rCU=vccw3hMy1CrJ#olBeVir z7mIZGVSy%QY|griXQ}?%#myM5eR3RyV#En*-ZSNbEJ3q z7U?et(Q_FjiSWpX#rA?EFR>Ei&k>>ItjZujg{t>f%h?I3-N8k_B1?*7pGrt;xG9oM{ z`f_~%*g-5-{ieIH>40K|;Bc+wA8(2oi#m;@JlILcjKbB&d}-iQe~YmLjg}eX-)3 zK<)kjBwgi-I9YkJkvPz3tX&%Q?$R~C3x5dWk+k=^hOCyy_U$Tx%bq}KxCF$A?9Jd( z4gle*b@!N|KKy}GF1yap^P_3;rY_?cn@KTtge6~XvRxE#O zo*czT*PVth5B;VLpgYQ1Uj{WZsi(N??lg42KA@drsPYeXGURnum4KWlM zYMio;;;RFENzZx<`a@+O>djUerATfI=R4j$dIrPACW~0^ILr;?Jy-z&mL*l0GlRWA z1q)~-`E<7)3XS4~rZn-QWT|uc@mCc37CMybshr}Z(5ky~ zHgA}Y-T02FjnK&B`H&JnPKVf_Y=wdnzX=GV9Cb|%gt1u`-D^HnRS1}&N}`RHKDl;z zT5UgWy_!(S@}_l>&Drhi+b?IU0LkI7N0EC((p+n5xybQs<}K#au#6 zV$Sda^08QUwGkBB0HfNK)J3$t_Q=Z}c`!S$#=Dm4-XzO>cYb678 z$Sh-!drfD)Rqn+`>ssq@li_pm&s~M5p9J|lzv#dXzt5fL)d_CnFovcvC)`DM`}mnP z4Zj)08ef$~djS5j=I4PS;y9B95jH!Ahf5)6+On7WjvVT{vxMK#+-h9WHa+VSWPDHr z(rIm|juuEPcQ-kZpB*M8Pv+ybJ;^M))4o&lzG|#!3!D6s0+B>0VU~t}Q#4DYnB*#` zM4?R7Wt3c5)M(k)RW?$^vIajmpyXmv_d5miZyA1GF}4(_vSa%Ot$vxM-8=1_jzS5NV(K3uX1yKL{?o+DW7?14(Usu7-1tiXiJ5C<0le{C%9mg6@$GI1YI)H^?1-b&PAB0sl1`z z1#c%<*B*Z2=lh%!|A)eLJ)awNnHZbDUNsRsrhPs%A0IG_tWWwk%=5d45rP3TPesz8 zWsMTT%}V(uYj2P6+G%Y1U@z`JhkqK>xcsE-yth-H>cr}hlVBjOEsJoN;?rl#K-?!d z1G9W>d_cRjoiw&J8}?E1CVt;$BeTz7SF(s#_MKl|J$_5gsVZhQ*qFtTf3cvf2PABI{%MERY_M$hhW=BX1vl{~~H8Q`_awB^O zwPj81WMKqA?@Rn1^Zk%gNBq??wOi^vC$4k}zfj(AkO(?I|Llz1>!}dvkKE>(rK2ik zMczm+!b6E`%Dxx74{qFgEz-aLG*=0JD{&d>SS4M-LBi5}k7mOR>aZOF!wC}*iV4Ie z4xUl4N-BcHWmbSDo9h-Tm%O53h!NY$i&%D<*B=0rqi05|RKW2@MLSy#z({YKc%F3G z(8FzMCV~M(|6YpNMzYQP?Ce(;d1vQ45+u~zFPr7eiKzJ$F+v^+F~kY(eXSMA;Z^D& z`0yg&hG@eJ&iySx13c8}9Tpq8#$r(SAK?TY;6VHR?d-`Gheap3x6qxu1n5r8Xj*5k zqs{wG*m^!CXdB>@06wQOqTYCt=?+d}e7xSij`+918O@;1P#>IEtqtMP6@}!G?HTvq z6##T`?+N zZj$m+p@|jOq)5nwP~G)!IeXDU_z+NPV#)U@Ko|Pis(8Ei_y?%-PWx|AD?udFUg(_G z?9O^WM0{4PaQh((vu~2emzX2J;Z2t0)-!N=Gjzh4S0N;#t^z=DBBZH(qT_BS5nhqp zL@24=$VbsRKRMhOZ`UaU`3l0UlrCZjbdnX4J2=h(q?8mx@FW%{JEjT!TWB25)>@lL z8Jn?A!Kxg}9w4u;cI2VA#|$NX2DNNYj>TBBT%Y;BNS9;ouxKOAHWk`~R#L$s<`Bwb z1u8DoFDmj@ZNH|-QKz-R0?NT%p{K~B*}1$(sX7jJ1mIp7?1N}7&jQMRp1x}s`Jn#n#$W)?Gq|z(BQv_-J-^bw1Z#!{7gQatQ1KtF^`AziYcTJLAFN?Nr^@(`L#r(f ztARk>N`+AJLyX+_z)+pZGbxr1VMD0Y)NUDMVqMUFu>x@_N!5fw*&NnCQN~5ZnWK+_ zZP_7qF6(9XyKx*vgMeMJg#0SFMV#%zqn-YdFLW792C|j#ItPJm^+v{Xy5`1@>Q!G) z?i`m6eJCi|l>FUHRntse+{kUmUeoM%vXh2`%BcX(F`LWgb1B280B`?!^4RlhJJ4Z&gV`&>qUaffMY>E$q4ZQDlhqxLS}J?Enl=3rS38SS$JlAy~%xd zSjYophNW?9f=R$f+js9&-({TcqH`@h=WB=IwUI zi%imbI$QqWuo10P`D_H)PF}cj6PQhCylyk9wTLZwey~&mnD3#jz^3*nkJU-J|Axhl}@H^XD4s=UYC4lSCc0Gi^LGxNK$9s ziN4sO)A5EzFMV-$&Ow2YxkseP9q&_M+nrXkmV{+hOsX$#eg{ z=Un#VIRjk5{dKG&+r{&xd3m}Ufp-%>sL_ej7S8?fI+>Z_KiPWw*)=8kYI%~$g4rfx zfSS@O`7~>>*}5&TrHZIRM#fWUpFzyRSE)i&5QZ_qLp#GsFLf)n1mVzXHH0!$l1|d( zdfQ<9%1|#(uL}oC`gPP6{dD$e<7sr?-H1BSQB=nHLQKC4EU09d7|OsDi7*Ow+gGx- z^GR;o+}S!r!;d4rE@6m~u?}EP|BxIb+B@Ab8?Af!x(&fqREW_yiVJIqhfoJV9ji_R z8EA|Qk}kj85tB@g;aZhl=qHzZV;7#5ozkmg?#htARdv_i}J+q~!zcDpUFuY75a-!wa% z2LCmFK_geeWI(U0LhdE@sl{sjvr`g6`e}c zOBj8p({f!Ma2u=NynbHiz?hdY&I6BAf4dUezeFIm)=7=S7;85h-9ZwOtn>y!7dO#jd_>1XKXxblldG)pX?FR^hc716N*2uze@OXvKmOMD z)8)f*B!1?=?{{`gXVMh5zth){bz&4lt8sP=lp<)p+LE5AywZygn-QfMaqWz|W}U}X zvl;9X4xP>HlFt!?7o7IZKs!>pMwrFnN1knfdT~YmVMq%b9jo3L-?Lv7Y=6vx4&0jo z&74QoH3ACPaShfhNOP*6ju4EfRQZ_ThTQ7@-sRMv|K};Jprrt_0VBJjem&sj)-DiG zkm&^saBDz?C9CJ~leio_zRUdP6M<}}QdXTW@7`vZn3VtUX1&_3I<&Bw2Vn$QvV&=v z!4Y#A+kZ^vl2+sT&li{7+J5r|Zo}s(0l{`nal@Rx`@scQ9VzeK%cT5T=QxP{>1yF)A>5bn;y zf30F1CDsC#QMA+~x!G~MwN8JRzXaf3$xdID&~ND^lKGCj_B-tdlE*GFYY>#CM%rwI zGVnE6@wGc?NO^7Y*_)34Wc7hM{hpc**({s5%|-D}X5wY$3zhFJzYVNTTiO8_|KX6+ zmWdAG+>MIV=cJZVo7uvz4R?f%!@(;0Ct@&l#$S}9xRQHPtNZHKe?=9Pxu&jO0B0ecq`p4A| zP&OjwusJV7mW4p*kSXg)u5Y~0Bl4_W5 zNJq(!ZoU(r!T|Iy+gm_gUva9op)|cXw%S``$V@%p@~(8?LHSea)V@6X_@}Gp$AhZ` z>LxQBsXYWdB5bO=``yFaXK@scM?rjwn#_-d5B@G!B)*2LJN`JSs0i;XJwCZ5S?v_m zp%lO_*4I%=ayEqb0Me;aC}}Y-myG(H(}2YLp>UcgN_sJIao>J(dS2zeca#0BNJ(|u~4Ak$G4f7hv=AnrTV z7$3&Iq7nUev&b^ua{Twd8;%20d5+2UmtEBs2Ndu-TWYJ9d&NcFWj@XR9|qj{xrK$3 zHtZk!w$InDNCp2NjjS`J@uH<{x8AhGQOogCca<6IBjXq}5&+P!N|6lxpZ$H_v-$1y zw~!Igk3sh^J6W~#5#|8`p@9*YsVsAPF4bXD?=0#*+|QpsCv(5qjUkI--{J}Im75#~ zK6byRBfFjb?L0;i%kn!!Sl=1zhc)*-{=KSDhX~47_6#sCwO+q6k?>7)Q7EfOj9QvIMu)KL2;|lHmXBj*Cx;nxPPi)o*Pn`x5)Ma1q zhxETrlkg)r-Cm}g9xi70(eMiOj$nUW-qb)sB>UU^T@x!-+n<3XvR2`VZ{3%pCC4AP z)zx3+ocoX{Wtr5WiZ>Wn#QPGrEJz<#zh+B3e|nPQ_;BgLG*QvDeZfxq^BCLR;-%#2 z4u{=K>IG2~*U{A%Lv{8ox?h(5FaBb(w z%3Mg~{6t6c*Z7u9Sxrx892GTl$fiA-onaLqyEtGF36fkw8XH4)kymK#zes@L3f?TY zZZ!SyzEo zhR%lc!J&OSBqA(#IC=lqUdGcVF!T{>`7`jy-Cp_4_k}gr)%TIMn=c5XrC^pzI^Vti z(yrUv+}^7*lyIo`d4PkM^vfHJO(9AUfJo}@0w>PW$!$(f&X1h_8}*0*KK+@o8oS;~ zAvCvQMojJjNMO}fwbEzk=8{>k8qz zHLUICLeTkroj)hy*t3~|(0cyjmQrx8g&FgPkGrkA%xCbqwrIkT43KtkW1@6;a^B|axQS3v zIr?FlvF5bjC959A*S2rlwyAY6Y^k=^b7+5ftdri=` zq4(uD*jm4V3*P90%FA9UM9CFIzl$_g%aEns6jBTkdmkSUT0tyXKeB5@uyx^F7Tmej zCJ*fWDsMu=#buU-*bo6~Wo4zST(ve1>CRO^&WwjR zA^V!wohtshK5?;WZK6;_Z{Vu>b9EO#R4>P;GdR<RL=SMfQX7`cOYub0uDM)I8yWis}!4fdl- zZP`S1nxjg)+uO5T#+&{}GR4cBS!Dg!0Y^IO3flnUTA6?_q?&2CX0d+a=UE{Q)qRf&gEli_#r%%CZ;5elPUO+AyWF~+nG$yOIYoejIBD6zHF&C z9%=imuVQ=!G*lU4>RP#fLAdEMbZtb@=ZL~QXrbVRtgt7O=qnT*THeX3yQ-fps-!sj zs7VZl=zeClaX6adn|mOR)PzE;d^hT=T+J_G86sf%h!B*Genf}K@^7Ar@L^&6UoC(O z&ZJ)-M$VV{h;^|CT$SP@i%GB0I^ytcZTRPJKEWMyQ(}rX@5G2PWV;wkr@{jt-uAZg z?Y{ln;I!~o?}+}zf|yLt`vov;`#*UyQKu9eg-W~hhiRc8mNBedXS&(1Zgq5!1E-rUy*jzT)2s_nyUE#MQiaRQ);;OEIH)JD zk8j|HsI<#=ZpfqMEr5vb;|=HRUoJMC79YPySHTd9{;`JFn3;cl!77|mh_g*Q;jajcYPC+@ zxt0o!=%LiOxw$R}ROTg&nbC372)*XXs+y9UXq@<~sLv%|)ACyYo~!>K1n5V7b3-**M(!tHbE-4Q|HJYQ06MBi;JQ6|D2TJ|If>N?hVgB2})8{4T2;NdO@?URnbviTzpFZAgsc~pL3eTouQY$J{kKn9IL9eaLwPO?uT^v|$@e7t zY-1TbjyqqBeYTOPn4;ppebG=~yW?>co$h+} z&54QIdB3MUv*3r>I~za)@|A-0|3ALIGOEh0jh5~PK@g<7ltyaP4V!LILb?T{-E?>7 z28m6VNQX#CD&5^3(sf_XIp4TH?ilyCV{De|eb$rnnG?_{wCWq7@S!;PO&f|!Pe4UO zxX6~_b+xD50ca|T_;v@3&%q^GyFh!Fh9N9`tC?m^%4O()FBb)HT#jqefv*?6Ph{YB z4gvbKO&9^w;fw&%ON4y{{FVDwpWD?w$~4FyfW-s@7jwHgqD%_A31Cfq`+$20&{!W= zz6F@^J}f0R@%+E{u7i($ko4TKc*!)3S33-4w1Oj=_?zO_@Qnqb@Rw|tgMj-6)jMaJ zOC%D9Xc5vu&|xr%@d|Pd9|j=X4H}m;CjO%5-Uv7C32|qb>3_UtTjc(#%{NEV^RCZ28_UI;HCc|uXEj2 z>B7w)wbmIwf(H-~rmo15N-&Qc0q21G`v-wO{K)Y=;ChK#E~cbZBcn<+R|kg(SbSL8 z`hXMOr9Qh9rXye(>?IpS4FcOA6?+=W#CaZ@1KM=~{4{V`)q?PwdaY(+kHC=o0yK}v zPVox;c}Vr_$n#10KseSLtw2Qll_d!2mVThANewrmXnq4`I-;P@x zn{bjoOP3DCJncLDGCWnXeGC~6-qXJEKJ^g2@93NM`B~Pm=&{GMcRuvDRI<871gIRz z#kLXrS8k{?+^Rwmr$+PMKu=&l@)=d6r>+ z2|f|!VZDk1usE~k%|WzhBmC4mYCdQ1@*PcBNH`@gRwfE`ml_xE=EqF=^=Tia9-TFy zNpOD<=pgDgFi|=P&qRHTvDLF89(Z#VhOtgAeEeCJdL&4$vK@hy_gEf*2~l=dsGt@{ zbN&>lTXLLP`yIwJ0jxSiGJo9KW#N+xAx^CV9DP9}D^(x1#kX5NBX}bBtf%9r$5?lO zk`{`L$?zoy5Xd!lqE7gGaA1iJ!fTlV_e(AzxxB%P@9IMPhRT+))h>|waWrQbpY1@v zjD%QyYX;&#ycp%V24({2%iPER5EC2y48^v5Axd{(8jq$T=zgz+3{8o&eY)-0M!}X3 z()}xqew&y;zpBhjH7SYRpCAT@1iSFOY$Cxkh0weN;aN$Ql$3m0T?;v2*?|}Uq2f-2PJo}fOYDR!WMFIdNH>a=tHo-IEhQN+{ zZ0&|Ynbe1ujuO^2B1EuPa~I<;D=K#(8sLWpcX6&N1nANmX`hD8VNQM*_5Heg%7}uI z7t8iFSIomh^>+2{;+t8Ye7vH@5|!xLo&P}_>Zc^F`D0&ki7#|)mHV;VY4`;ULh{TOp%gJ&eY*9Fj?MURFq2+@!fbbP|`6Tm9n*m za|W+U$`jihj|;DGDW$?-g7>(0iGKih-2Ws3v??*kU;W_$h5ZU**K!^s6pX|gZ0#y~ zyUgV{_N0?08?@EN;83R$>s@$<$a$mOK61CgJIj5C`8I%>=Pi4LTIhAFbg)rE1b!-= zaO_V>&1?E_X4hX#8g-2MKeC@R$d)cl{5T6**7u)MvX@^TBK zR_K#s{dyy#@M(#Z3E^UHzv|b%W9!j)(%+~u7Ll?Z&sL9=aPIVE_nCFKN=!LsF`sKE z0E`NS_OmO)i-x)807#v)dHyqLp z(xwvSF->*~UpLsin*~t+1hjJa`|w>|0Qg45KKdspQzhXg5rvOGyZ-mzbOb)3-yD7| zCVe@r$e(z>DBkX5lHb}1CAW3}gc8DEQXKpA{P9Oyip5_FMDIDKnhl^TN1pWaQ2@36Ymv_~>qV;3T*O3V%! zc&H$9SBBzsS7aY|YoTIbvhdtsy ztbJ1@ngqU#r%7LN3>bGVl)NL>;Z=E78g(Y(AQ94u6F_ZVXN)NWuY8gE&F(J|V2Is7 z-4BP#cy#RS7CTt!L%#M~&J{M_Zq{P{(cA`J0>`2tAy_Vm0D^(b9jd&J(&e@KMFEcc z)KV9FVKTU@7SxJP=35?meHcr|tAJZh!ePv`Q|o+xK@Ts>*3`3kYcPQcEug?q?7D3U z0M)mMlaV59{z9mC-5jl5#Dgj7Nw;jzI2OV>bS>Mf3Vc&XvVsGy=0-9-(wd0mKl@tC zsZhV77Ru^)pwG>9OlIax^KY|PfGT0EDi%CiBzjmBpI^e~c^RqE#WM}%qa`6n+@AkMLH2Z|`Pusvu*gdd97T%Ym&~sw zT5x*6CR3F9p*MdfLp6b(w0{yTjn1kmQH*3gn_t+QTHW{u2Wqsk-U-ZAda-NmcDd?B z@ks~=Qi@yUk_es6JE&}@%@|719vYAa(p%^ya&KDJFN+JWE<6+_8{d?GYm$;;+3XE- z71Lj4D)zbwDyrl((uIz542*X8(JuS(bHUozQ79ZVF*TV9e;-yJ>OOuYfaQ;LrJfs< z=v3u!lSHJpeMcHlgbEvO-Evg;T;S$NUl0j7-p~${o;Gdgj7PO(q3cu_7Rnc3%=L3A z3^BNo|Dw8mhOnGt%wtGH4MR3`phhU@XR9L-=yleAICC)qsKeko2QN7CMk z2CBE2x(xDUHJSTC88MS8UCuS5;dR8d#Jy(qt{*0=RKX13O31a4Em$1Bc3LqH zVLc)NG~0w>kMz6pY2V~YT~SlsJ`V&UNrPvHh7@^lDg>cv6S$y(;fhB%&PSj?fHfx=^4_6;$Kg2U zgFFl$q0`)@IFuFW&)uJmHA$iEe4?7x@H?A>;jN{uI z!WPIYre$53k0Ca?K0nf%O7V=uM=~YyggamvB1FV5#vtCR)o3*VF>#HIYq?>8F4O(3 zp~w+RK0UJs4eD7tsK+g{E%Dg>FvWoT0GK#iOA_) zi7uL8zS5nqdqOeC!oo{Us@jQPFV)-Xh&x~p*s^Ms(oGw=X2)#Diaj+h$kujs_w@W) z@N})Wvh!nq`&KP1)QNs#rK73g!`Be~+=Kh3^ICv}Sg+a%a(hEuaQgUK=OQjJ6vsv^ zqI6dBgBmqO;}udQ9_tR;uYdo8M|09&HD{U-5H1wOO;)G(4;$8W1hhzl1FC$Q&Jx_1 z6-Uf-;t3g54YgHnuIwS+`iZ^D1;8{;pV%D0eCwlZB;16Wt}@2t#&KKLVfv`k2{JIq z3%DK(!V$Yw+eiv1Gjnl%2ptsLZynENoxJLS=H56~Q+o!2p6K*V!Yai@vPi8TYHaw}T zK2tC=ag#>)j=}1qEXIq`EWz_LhOYsmx-Jv77D!7s=i7+5<&;a{Vs{EpfJw@HbV-Ek zvaKdH_j)A`H8^5}Qu$WS{a}Im{g3PG#NlCjyyj0omK!RG2^kje4jYy$kzV{jXO`40 zpj6ZeNOl)cc6Qcl?gS*!&%ES7{U^)1Su=P)TErhv&VK(6CE5?CRGw&e#?1r(*4ux{ zZ~&$m%&T+Ja$__`lgxc^quai=Z_xM)_LXj3-NC4y=97mut^jtF(j;oI4de_21$1S* zf*q$zSE!ytpIbxISv#PzqH_?#RL06vpAX{)U$ArbJm`Fm;-tBFdoJzjBiCg0w1=orq;!uU*@ z?(RZU-8b85l}&^*_}cUBa1skdcv*wJ58DO7C+)mWZpi`_Wv9FBYG) zdo>;`!h%D-MS+A)^fF({iway86a1@|b~Oh*sm!#xY_F=g$JU>d903j9W$rnVaxXr* z-2DI;8+q?>^Et=vWLfIRvCR8L^Lm?;tBJ=|a7>T+WT{rhXl`uAaG~eQ!K{>-CKoo7oGlmP zqB`Y@`&y6X#_@JLsR(Hx4jduIO{@qFNChS1#DWpDIu)EhhnU+gpxs zBb{^X={RBZP)-5FzIKKsh6hR4tfO?B9+BO4mDNw)qYJ$;Z83VFf;Ms-RSha zS}N{_w4DL%1sz8ea}-~y2X1en^&f!{U-T}BAwwdmF&2|Sv9H*$`A$btRREzn5$TKc z!!3IvSB@TfGf2=GC#rMK;Ieq>7H3)&9mXcr1;#Jop`|oH!CgD&o22#G_JJ5D<$zHd zbz>~BpGD0}14@cRd3Pq{ze@30!Y}ptl+_A`di(ngDBp&ey7nVVCc!;i-Z{El8F`p! z&jE;7HHVHvccW=0TXa^d1z63G^?Q2o$Q}Oi>_Q0; z9bQ2zGwpBy(aE-Q{Wb;t1{-!@59d#Gh5R@}Y5470B&Vj+qWiqP%in&r$sqgX??!y~ zvjwQ1`F#3UfRSF=9I#A{HT6?0gH=>1B9C4Fn;g2f%IcO-1f*sB6#}w+bIGxCt>@rJ z4WaF7n@c=ALI#z%eQ%?x6p>zF<3w?HChzMb|10N_kZMZd&(#hyz|KsKLjd{KESOq+ z2s4u{y2t+kwohxB@{;c(L(-cZ0p z27Tg)x8u$~k7y+v$h6>ejSVp%675uqMWbnzXB(?|xMe+Sfw*E|^&puX-?+9!Nrs^J z>mkuy;zSnjHO-Tb!tZbork4R!I0N=8zNDjHi}gIH1z~|K8z@)ra0-P9J-9Eg0gs86 z@=@1~-o1p1#k4(;&L^Wf7(;X_wIz__9tSezt3T(akpiG2SP;Q( zK1>@coF97kEa)9hejF?S>KAbOy0h=>hI<2W?oBqM!mkUGx1NS|Cga);PcBwzpY`3m ze+aoy=tPt6j|D+qtvl=8$Lqzxn=b%D*+nUoCTN<*=IkAU(hdaMv0Vf`Z;%LrIU@R= zI}hbLJck%+()M0q{S7>->~WmLa(*hC)P6Ou8MOS!3PGzC8B|K$d&>d zcS%o5e*_=;=6DAT7G&qYTbFI4^^V-1VR!fT?g_G2Ew^6tw6)U}pd~4@tzI-U{SLM( zvTLOD{j2u9%z!!o}Bit1}{A#s(E7~bQnBHA}BDMl>W&}WCE~fj<36JR;Igs8? z3hu;q~X^^HI^N47IJV+K2kh@sd+%Bw^&C!6yE9^}n6w zB|qd*vt6PJXoUqQGZ!-t?bkPrAVLO+_CSS|4?+$#0-%QF2!Az8%7f*Oj@MkXg9324 z&5m&cEd9WG&&G>8KCZ|>E(L|lT|hio7q`u3_brDWxu2l0-`Ow5J$XoZWJHX*tymCN z;RMi6i(}0FtnSqF@mTs4=}+TcX<3p0#f(1)5a_FzLO}ac7RE))bZL%;NCrSqqPorl zQL(iLrwd{kUaJsazf)fuhoG4e*2oV*~G@6&` z5YpdNG{#U71_2E&1|n){z3}ptj$);3n!fSaeeqW`okCPHe^Ubmw+YHWs*2ns4eZ2! zl^KZtDm$V6(+!V4DM#1nsN>l02Ysw7R@!fVIreT^F@)svC`cKNd0HA}xd9+@9XHx> zZMb-fVFL-UYQH;bxn*l*+oA1ds}jr>@%~OA=NJ0ry@+XawR;ZfX%yboQ0w~RmJ8;v zGE|o`R`d+!mCa^ctGmf-i-yo60stEodI3HROZreI*>iOViE_P`IO?RFD9W=<8t}5u zd34lgn;`wu6ZPh_WQx0x`OV1D4%qo=uYn`Xg@+%qmKEoHn9mfNRp+kf%4GNB%N0T9 z;#10_=i}BWot8?n$cpiM%x4oef$to@?wsj9_1{`&!SY=5gwTNTkVNhHWXjdaA7=%^ z8pKads7>>_r0o`)-MH5KNLd=sXuxq_rcL~f_$`s(+~65hvS68xZyH#V%lZLuKD*te z4<;TFkI;lTLh4m-NAX(|4g}=*Af&gVhHpCiq@h$=cGYNagY=I=++&qV*=aKr4PQ|i zSPlHXq;{Hx8DYb(07npBo1S_32PuF*bFV8YxCBk$$DL$lFedn4x{K79bJS?6DPqhv|`ow?>GA_Fd_*yPzou%xsD7Qf8~AK&sfls z9~C-or3Kc_;*KKwsSV2oHrBU7GD%V8bBx{=EbptnkSnS$^g#P4K75j?8zp%24 zn*z5%Gum=Aw?9D>NPwjtM1m0dQ4Rz3>Yga8k$)>HHhAaDt21;(BD(0uRSq2SuO_V| zn-t;QaguQwINKJq1Escn18Qg(5dvsk*@(vzKlv-&<6^dRJXIF2A0LkWS!F!+f`Km) zg}5feMo9dsA0DZeio>Uoc8y`w6^BxgEsidA?d?_8GxP9SzhtZ&Kw3Rv^bBaP1;%nB zp(29!^7LHCF`hc;*=3syU_lGcZECayMb7pDmyEY3M+F6R+3OaHsU-w_oSO`mNIBBi zNZP8sLuBswuqUO5*hhYCZlaOK8S(1t(5pPO_ad&w{mM|of&DDotDEf{orItS@+7;$qM~mfDEwK1dW@L-}9aZ`{Fo*?4;}p96KJqXQ)2J z_t1lpBBGO7#%!m{wGQ=TUyEqEBA=R__ejGZqPLG(l`*Jfz#QQ5(1M8d`H>JfP_4F> z{I>1goI%{ZGMx;;TO4VbgNqR5m}rtnQVHA!DZZT3S^YCf5?UmQOS$`Nu^K%dCDs1s53fvc)U8t7PY zf%9L1pnmd~wC;nD(jb2TDaQJ&y@@uY-rCN1JqH{R%r|>o~YI?J91$4*xai+VmB;QX}nfMF-68FOx*&IwY z5r>9lN*ob%1cr7kdMsD`qH)5O5*2?*ZQF`iMyM@%L@MM^-s|)M5=dwIOM*jutCN%! z`&768BC8Q^dXpVux_}qma@=YVPM1RE)O)*Tk402KG}i6ai5!29dSvA}%zUcd2+oHK zxpu-egHpB=FfAb05NTDYYbB@fuz`kX&$82sT9Yyx1M{6iR#bAh^rlTls^Qrrb$hC# z@;^zJL`PX3uBVJHI90AKY=(oJv29-?BhtNOrSR>(aN12MY&Jf-#is8lE}%xf8IX(t zZN@|nkg{swtw=n~fhc|>NJD?GiHDq|(~Bday72KKVFsWLqLJbGKxfnbG=I40G#nPW z#^ltk2EFz^c-wTDVe4il&f5R+OA2OmUKyYIn<{L>V~6(LkBeyN0`!D9;zNzBP!UO{ zt?!^sY3Rw_5zy(ypz)ch>HSmvy5`1lzFj$-3E@>&$QKlDKFa{PeSTP#=(!aXZweT( zr3P;RvW>+Y$#KSfLVSb1cXbb$_(2dVuvL+j71$n2^G(%?|K)IrM7dIvRXu3Zdh(|T zWI#BN!^fMS$jIxd^<7e(RM>+y##8i03^1OM>PMT3RQoDxT{I4?1R7r*#?#6U03ebl zd1%!B%Ol_W5Vo~<<;(!-Q+aQ26e;%>214*Y#S|$UcH=Vb6$;i%=&{rX-3l*WE%yhY zAhHSX#DEwG-@TzjOXXd_J)MGXl9=t2ksz^HVrhpD2F#czc{Fmb zxnFIjn(RTyBT?47l4IUB#5npk0Z-(%EI~~U)jTGuwvkX`Nr2>yLpfh#&nX)kbB$CS zx;fI)D;_ycDePb>^PbJF?Fh?2JJX650CJJMkeBs01+>Xv+po6urb_VPeHUt}Pmu76 zduYN-7cQ^ARw88}^)P-KoLM)~Q*(%Exo$*t*qCE4N_%(YWVsf%$tHtT)?3LU>P5Vr zEZen*tL5neW+s%lec{w@s;mwt>>h&fR}901kS?!x2Y)^&oQP#z=k;oQu&1}BR502S zt6HWS%o4nx%wq9==?^iwK2u=(fS%;2$UK8iq>XDx3h;D0pT^TLy*bZMrsd_&z)euB z7F-bpMQD68y*|qq`Oy{~ozjhgM&%Rv)+8|_31px&Y!&+E;cL9J#CSJ}jHHy)8xFE4 z5^TzJznK#}C#%E5D-Cte>e+_;hn1+h{b#p`Yw9(hK+NDoIy(amf{ zBeF!@O7&LLT+7&Eg_s)Z`krNIpiwP>m_)*XqDhh<-193+DZ>x0{Vz3*OUHn}(C(r^ zy^$1*upppkr6a?%1}cME0vGMBdlMxYa!7))tyfY| zL)+9Er_1-zNjR((3Jz#umf`n)uE~<8eL4r4vu5Z?Vjny3M$z06w8HW>y=@)D3{xlR z@s3??xzS}l63X<`GzRj%5YT7)md?YIa&L>CRHvP*5v(h96QM8Hpi)-0fZ)gRQx~d4 z+`jf5Gz)Y-mXQ@iYa8=7Xyql3LTE4tv8JJ)Y21S(g;7npc27Mpw>ZHSLNy9>GSG6w zO2ftyH9iDI^Z_xhY>9w`79&r}x8etrOfdp5v&q$y$iZyI*`B<#c0AaqV;mya{Hg_r zi%JKg(@F}CYD$W0>XTf+hdeQ@Fs0?xHEB3C*8(!Cwv8-fjA8UgOZ}Tg!x^RC{77+R zs@;0xNY4C7lpA;*iD-yfRFNwZCSaocYD8N(mZSGf=wXWic4~OUt)vZQ;xg{Xa}~PW z{!9q2qx8R6n%$0zq>?;;iEOoM6CTs7)40XE6cvVcZ}Ar6$t2~(NhW>8BY*Ru@+DzG zq=300g_^@y^&5#d;o`k@h?v?cn}v$R0VEE08LJM*qV$QnwP)%M4)e{*=&$ypku)oB zW1FUtY%vO%bf;l@l8drxJGCKHEm5!b32HQ$KI)zS&3qH?Uv9)5`xloUt{~M7HP5ia zp)CYu6->Gc`Ty{Z57b#=~~rRcnP zoa_k6&aUw`JPnnY>7&{#7a<2jepOuojkPMLxllbMHUu20 z&XU?b0+^|>T?*)h1HyeW+gGQyYL9IxDlrmzpgrfE{rpprl?6zp+UMpCaF#j3`A9ZT zTS?G#w7UwVVa=HsUEp_mg1%^2#L|O%zP&fBsz)kyB7Bdoxq5=xz3AD*B*^%zDz7X^T<=#=(4OG zl1+gJ{{6gjy41)7js5PX%hqXQe`SsS)Qv4oYVCFjtqMe~I_v!T-CqJZMZ_>f0}|9t zo;gQ)rs5v=0h;&tXm1dJ!ss=+rg0 zLV*;)vuR#W-J-a6q}hSbtO&htAVxf;o6dWKH_9Fv#Bf9>_8Iu!>5UIo6@+A^~mpZhWRO&jbuE%k?LqomaUV zZclCW2&5>#T5qyAvKw;P>g+*#Z`W* z(Bl%Zb;>xqCKoPmgc0%klyUbhGmheb_SlcxpJW#3N&ZtGpg9TCqUygrCd;y!-1ZwA zI8KHpL?X?vmO2|PhfUs?X#7n1_sUGO2w-H&&5hYPQR|^W@4%?v+0h5A|MoQyU?E}# zJJP>&`tJL`kJc$RD$(qz9Hp3Pho1D5Vh=^rOdUv8a&Hz0=jl0Ks841v+A5bU%5Z~M zHmql#L$`)RWG~jRhRyVhMD`UYT zsDyH@JruZ9zLa>iAMc`|$EkoPHCJ*|rcwJU&)D!XS49kpw6}6;AudMpJ&+ha@{2T; zCZ1F2K@C!AnvNB49Cxo{fy5=Dr&~EJ-_(*=9@gP*lr*t6YmUM3aoU^|fQW-$GEp?A z;Kf|H$3PW6$>e^sF&9AhL~j&H%$@-B1BmAcfD7cW+d8mqc7M(BE1WipxBlBa#GH`H zaF9dLyR&D)yBMRUe}*hWD&K1z=?ZCY#&5lgUlQj~*H{T{ep(-RG{DI~4$oHZLj%qs zP@LIZ6+r;LFguGBE2LDdJ!^i%JR@RyUp0$|H(_kEf7aEd_ufAxQ;3Yuis@(WOAbU7 ztly;zj_nskLz!Z#gfgFP=u&o;Dq31k?lJXey@AS<>*K-8*&5f-VuSOCMFOVWJlCOU zJeWfWC>gLZVm3b=Dz!CtWTB5U(o1Tds4hSDc#t@G97C5x#i9q)_2}eJKo53uHTHLa>mX0;| zGfe%%hlGi$4lJHZ@7>+Ljmc@_@QN-t-jD!fV-4{l$1HKuY)ntKZ`mQ z%lW);GBzS1L@oZcBmGFEY4dkMT!=XvxbY;}-4UYS_&I8bpTj4Wqe%l^!0}#s(hNBI z&OrS?N56E)d*^8$`*BrqzNr*ORexLqJgDcZ#zhm*x+Dq%d1mT%RT3ZLqY*(^EO9b* z`m~Y@Sdv8;5>CE1ys(*w8^Og&&X6~7|wUBL0P=R z0?*v3E|i$z47)nnq~$ym=!qt0ZXlF!#_8KEctIF2pFzg zuEL>%lxh}5{e6S%FyE{BH}VBQ6q*lALhTRz2S|hEk_BjsLD_ zbd7UtyiX{4$?%_norpdAT|Zi|Afe(lSrNDsX$;kKx9A3WFxTx>0b`nCNk;YD@WGCPEY$fw7vj({m~p!HMQ zT~q>QYHP9eP+cttk*$CBtqWi{5zZIMAgv0LW8&NOPe2UNQYRVYndr&caTTuw%+vvv z5mt!4-6Qt@{S@fBP+>w?QsJ>RmZtX9O{Y67NOTb>?o164-B!!g8|u9g+?+ytK@^L1 z?Li|@ij;pIozv^*BLLVA`a&-+XU-;&wqsv1E`qd@=BGw)ekKnAT}(~qs)XM@{-#}T z%phdcA6D@l;N|peBDh~+PV_{kfNpBlvsk)KOL>z*@L=yM7I1_fR`}Qpygwbwn5@)X z{rf&5L2%D5l9>nKe^0rLhX+9wq=LFLbzxJzg;h4QXt}2Oc37Hl485XbvcFckGWh-d^de^>ep3DaB*QrYYBI5ABDU)j!G@nM>%4H>{X zXlusJ$Pix;Z!ICzTg?i%tpX^Mh8OFa)Wil~!wFyt^%*S`*EJ6Y69{g0Pal`zYCWD) zoM->PjQ@}KA2?Q|Zlshm4S%^lrZY{HTVWA#sLh*F5)mRNokk|Ksibh~5@NnNoxi?T zkXVAK)nedbfXdQX+$43OVsx!U=BkRsG9QeKIWbGX`0*|_3~wie$mq?tuDIw$HVTam@O(*_ohuyaJ&u2 z#$yM{s@hmIij~>HT<8h1SpRd4*5&UL`(hJgrhNqkxT^p|xf$jG>MN@nQ@a_}&g zmg~@oh?%Z9e3%vcY{||e^`?SHQ)fQ`ubmf3fgQ}7*+jut>D_oz_)14~{@?46rV{u! z9I^;Z#hQ8jsTP+c74>PvYPA2tX4nW`^4V8T=l9Rip~E^;{N|wUkwrqTx1@d7?&`c;XAiII$D{ixNeT{goFed zARN{na$v9gVdN< z7-p_4Lj8d+@G$JB#?upU72srcG^u6I(Dpz*Q&tDW$+QfLMe7he&Qj(CAMknNc5Vy-+#7a8X8jERtJH^ z>z07R{qL0jUeCsV*RzpYc?<{2uV(rXtgw?r1<;8l>DSxHC$nhL4bXWA_}+eNzIbNH z=d)X&8cgBP*w?VJvB_gs1ujF>QKf|tL7IB;gzGih#2Gq6v>-8)BGf1SIh>bXtN~s* zs_fW(#=1qD74ptgV3OqXfyn?pE zMVZEVum5%KNe%!1T*uRYuH(t}hiDBNel>@WG@AR{wo=#7(1|uS8kzxn+!PqH%{OW= zY0rd)Wye&B1Qx>NZvgjUjK`E9m-P*4U>0-oatTPEnAHCLgj)G2A9k>N31A6g+u851 zv`9opk@h5`A;5~%r(dY_aJ0{lw`^GjT{D=)Z6V*qYFlBr%VTe62@YyWY5h-JVnb;E z*NG-^*4ZDUlOfIZ^;$8s&xD2f+$H%PWeq6^v9XTFqAGXT%eKlwHM}LEO4N&CVOCef zuor~&aYurR&{kCbzGR?Kb;h16!!-G{IS1q>F)skWdpjfg|G9xMb`2RaDQOJOAl5{wrQ3=UAf$}fxOFZ z7eo8~-RL>KKKrA8UC}_L3E(oRVEA7^C7339z&Q(9Mb$~#fe8i5#pi%}SKy?RujpZz zCYpBG4ZmN{eAh(R>^%+q@ecXB$$v)qpHtgNu~|aoF!)|R<-%febgcPjdn7k;y)Rbr zW0C)uL~4RMfpxhwz@4}i=dGfWBW$g%V$zr|nOgN$`twDG##3vrik=m9+;6`yYOt#O zZ0(Ch9G~3*dEjEEfE1cI40x#*z1b58FGw{OMIdv{QbyNYs^n+rl-`N|1ng?GYs%+5 z8K?z2h{_$iKCI`9Dw5cBv^k8L>AWuw;ssr|`Y$QwFEW6^np`S}(cg@)TKojJw~CjC%Z$K; zJGo;1J24FRaxEbcB+)|M8+^>n3?zLZ#!0}yWfDim`X?YPlp4eAqIV5!K&Nl(*H8GK zKcWkOQ9gf#R@r}YKc?gDMk$l@1ztUrMKQU^U4tg@VRfcb#G~3q2+=TihmVN~+cU;mVmT z96}*6%;kt3JgDtwXLucjymx?(S^S*HhfP|I0SgZ@z1>X|7x0qM@0k+K>5RZ&@Pp@! z!9Ot$1p3bys0M#{;s%2*T1s-+bQ?hcz{1ZHuJMwdo^?@4S?>P zff}T@JKs?CNqLWj+b=6m8x|TW%9d`)#c6`*8By|*Rxh=7qZ{3u$$hExOf^>MyKkrY z>MQEafu2U+m0B&zS0wF!T(-)Gzwcn40XEK?!Jsg%@=zdC@lIW+GN4;4A_XHJpFoXa z$lSb@iIICmey&EN_R|l?T9*+b!DIkydG0HoF0HJb=rqH~{>z=>3VL^;CKK!he7p!6bLN?~d~vncwZy)sac( zvzrat9{_}kQS-OIO*OIQx=j)c{`@yw@9?TnQADT!hNdJyiC*E!;C2w0&2k;yS>s6k zb{7Pr0fq2-mk!4iBIH)e3<`W1dxk8gZa`ta(AULr*L5CotW*!NL`MWb6f;Re2ezod zme`PY_T`JV79+>p=8vs<3C%)U9EPo2k7N832S)!ne(+1AEGBQmb#@C=?b?SLJ4`(! zDzGFaxsA=C6jOR0yU}+|fUtnpe(qa5K$XRu4yedvH=jn6a?6Cd1%N07oIu1JhD^Y+ z=Ggxr7fh@=6`Ea0@PXiGD3M>*&DSIjEq#uUMoEBp^`N9AWkR-)M``lj^TyACVBO}n zi673aXEovzBf=5^Y_m-C51WtWNP)piUr?vD1lCznTpZOEuAi>gWvli;h$X4%H-js} zHCAWl-nK@S>~~{Oal{8g#Y-S_*qr~hJOFTevW1_0$o|i4zlh;GWgPqxbO)SA69M@x z8`=aSt}qn|r*RSh^x0tw3l2EI%M7|eA0(WjuRU)A039|T)}9Oi5Y?X-`Mf#k9<6ir zHoy>XZGC-nspU9%oV57DNrSyILsW>CRx*5YkGCutwCLPN;RRcDGRu_i3bvkMkPX!f~l)exV26=h;TC+YJHxZfLL`ukB(r@{zOaMNi zpr8;D^-4@-F}{L+py9&wxb`2H{s`)nuVuXY49r9OFr2Qr&mtjYPV;cPZbqDLWq0eN zV&*$>d1LL8-Z#KFM*+qnrSsby7XTD$n({c0NhvG)jV@j>;nSj&N(5urECz((?+1NM z^A4+e=W-opESVo6``S;6w>Q`KG5yj1bGlP_WaRj#?Ypb_wNxAvQ&Eo&$@#l_oFpLa zyVY=89sEYjlw0i{f#<51lk@uYn|=fJb8F|HT#}*WSCv+Yy7W~oAOX#^+eIVlG9hOh zer8=illgC^?2qRZaq!Fb8<4M9Vhh|d-{8YyDBhJ?FtM;GESlO(!I-_TjwXjZ>ushI z%9>3It$$I!i_2lwD6)De`Vok#2tcOI_r8^6l(t$7rtDRc0DX?#zVSj}@7JazgU31{ zid;lpqsT+?h=Kp2O`iKk?I~QzKaTx~Q9Rf_FB}xg2c#^f&sF0n02%L-AcRT8%o$Lp z{0qG|>~Fa7^G+_c*i3oM-e}%$hV7pK5H05OtB2 z0z^hfPx5TF9_!w|rB^Fp<^*t&#y{K4uZ8>{Tbs8|GV4Jv#=Fo0rrmise2j#|touDI z8rt+~Ze8n|Vz}OV$p%8FNA?36B+3P3Kc4-O+vl^OzbSR8wY?(qyr|dFl3y=1+E!ks zQtP&A2YOF3;p6JWLn5^MQ(3%H-I*Z41k_v zI=N`Gw130!KY2=LR5{yTnD&gf95nQHhcJ7V8i+~uw)4v@k>11@G_>8{ z>bH5hd2a1mZv2rA`Bg~~uaFNc!02X4p6|ERC|QuvxUKwtCz+qpUpW!10OK$eSf9R{ z`KPH*4C$g-sd(;oR?E%bJ+E-YNkAoENkR#7fg96WB!o@&20Na%07l*TuJwhfsOS=B z1FwytO!E~JXDCipvd-h(i*&x>s|3^}`_G#l#_^@e>lRA~O0!MtttPiG?Qnk-IzIiF zto7Z_t@WK|6ZA86takfz*$3PT+6Sc8WRVn5@Ci+gJ{kvoMFRw6aHP~yAJl8HkiTng9{y_wUrdg>86qdXBH9sSXImgbZAMNODY2-ER3UvkD57HR?28M?I zZnGZLu_5?Lb`mgkh+9*d|C3Zb%0PWvslg1GCgugGB<;K=R>Y#O5I5)dPPM~sEzMdzp=*jK$Zj0L|DvM;b9UU<5->et2fMTR7PA?Piu2js8jqNuMZZNIZ`H2S zHj3n){o-C1 z$%_};!~w~+?PEM!#oS{OFG?k%=MAW}VW~uCTb+}W16I9-(Hzb9CFV??YPc`w2=U?)X)CShrVfYeRIo?Nt6+h%i<4pV? zv_zlQT!-BiOA?NYe6C_W-iKpjVrtk{4kpJvZ6CWn`j0GQ&dJ*dsVgQNwCIIYTYfYr zlh%>RLpKw@`%s0e@~?3tZ{N=vKNpDN5gq}xl!AC+@)sh!gC z0k&|=$NCa|{7i#)FwXvts}8Qmid6ExVzv_VI5MmkruR@xIolkt{j>XPS-E|UV!}vi zCk@&?pjF-ORyH-3CyA=p;E-)@Zf&N#uTcI#WwX8>@+_Vlw!Ml%YT1mEYA_w-<%)JZhV-x<8N69d7DCL`|i6%D^m z7()gD?tNyn-WdG1zu=#B`gRvv8hiv!CH0zI3cI_z`vI+n4M2LzCVg$W&T?d6)^#?W zmHO@G0NI%PV^)@sx5{$t3rvq604x`cF1i-Jk;q}GRZst$Ro;~blJ{NAp_kYcbNRW*y#%Hqf=%EcdK$ zs|Va+1S#Zu2*~X`18_~DKO}n=%q-iN&?VS_hS$-{&!NrC@jniL8OfxihWNJ`@&P9| zubaYDWzz&XPhj_lxBs!L2Qwy904MRoDlo(e=R+!unJ7zSQYO3Lk}54EU{$prZ7(eS zG}C68SFCt=3=e`+R07l=3V~#J3}Cv4>u1=O{T$~Z_#BPiY7DLr&|BOubXvzwT0po~Wo;iN zCnqPg4^6$mQ_-!=W68bF;#Kd@<$D zXCwfAex;6-5xwP5u_#Zi$$L}hAg_!npOebx$cl;Ks|s2OI^7%to`TSK{@%Czdn%re z^IkSe{R<8aSNPVGB??`GHd37?J^%jhFXZmbRJmnqQsl0HP%-nf#GzTt zx%2=(4Ko#^P;kD}-7UOqu#x$pv$c)m^~b_AFLG?_Np5!CmtNWnK>MfaPRwIoONG7* z5=Rp;=|XNtWTc^yGL`vl(Nu~Sl!ljj^gsBlCYQ+nptDrvDpy0!$yGu;)@~VYavvmZ zx?U-@q)AWl&Hqys(w)|M+e-MbxP}(iuk@9i7>meYja)Q31Dt3*kO`S#X-T1qMQsub zaD&n?lGcY6uNuY!kY>*b7w`L>XN{}({bR)K$rf2s=EeFk-9_6H$yr5#WuYk7r&NZy zI>)4UQ5YJW->pU|wLZYI$V0nKW9Cgj4g}|FF?Q;@$=~!D8+U<^`$dn&Z8<_Jk*AVU-(W@I z*8-FQ@~N^X26YaG?ns29zRiP~qHU5fbkogAOax3}D4ltxb7I{rgDp&C9QJ2J9&!4L z&i1Wkld;|^VzQb>>nMAZfhf2WE$Do8zz(ofilJ?jQq@}&V6p0upoGvrMd?tRz1ngc z><-3@Gd5ACs%+u=e`^3YzG$f+z){7r{B;u|%&EARl*p_{*=HCx8H8Ea`o^I_3Yg!} ze76`H`^VIcz@RapzDqg-Z~-~oiceU4cx2QKKO>^I9?l0;93HV{dr(3j1^zdwAWNNR_|N ze_I;bz0k3+JhR50NxSub!lAYR@T#*7nXii<*3h+{FZ*vPxnk3oDPs{3Qe%8pQ-<(Z z9uhx-Y*R`8@nV6YI+XlSbTqufnnygU5pr7ojzmAXS~X6k-XyE?7RMu~iS(k_>8vxy z8(5ktiC5W7GWzfyeKX2!w419{O69YySpRJD)92%5OurW60El2To13z?;g{eTD^Mc3 zIxPj2RCT^Dll(rq@88r8s%7q(IM*&fxbeL|z2TQ*nkNmf23C3Lx*#+def;G zzM>|{a!9D}5l{;i`@aG92|w@oPep$RbOX&H>DCY*2d%{{yXu{SewPh}UA| z(;BcpqxG4unEXsIG7!5#maJQJ`JA@4*QV0@&!KF5TQ9C@P`Xv3a9jZ~-)lmK(JI~y zAy0L~V}rIh_8gEPllsr!G5p7tyUp~d7dfAI)I3s3yO!g?6pMUl4rOj0VBW}oA-A3% zuT+z$!8^mikx+#uC$lgsQRn{5z$6eUQwK2?iCU#P)`Pew`t<~{WsplrgPYyy(r+hv zx;+57PkgcOBl?Ktjw&3ltt;kOptLe0%p^CA=(f&F58q}l4G&Lv`X0i44Blb)%H>pco$rsHxzq}q z_=S2h7R?#GyD>YeL;i=0KsWG+05{+~wg&hJZIoC`r|*@3FxkYuZp5(ZUJyN4)Y;|R z+a)qaOxJ2Gv;=;}?bmyga-7;l*y?rfE$iyOievz8Qi5$8^ET7|g_PeD>hL^_7RPNC zrBv?GRJlZ<7YN6HUr6~sei;=gYEoFejq}(KIgJbZUt0#H8kxeSB3qAIz=vrdL!?;a zRHJ4ZYVD%LGC}UDdzq#0h8ompYs;soMj!$X;Tv*gT$5g+`SWNgykTl<=&Lq^h(9P} zG)ZCfRF<40fP7-f^VR=Pn1WtXMg0_Cm;XaWGOGLeT$rP^x@?oRo*scB@r0yay;~VW z=tqHnNcbkgUG_&OJ){IUBGJr27Az_hwaD_vP8qabF3M7HCRBd_|;6Kgpq} znXb@r02@OyU6JFNvZBhEiklRCe>b(!ioc~jJv&q-colg6s?;}bGW^CUN z5PPib*q9*?l0U7hd*~pDo|NM#TA5zwy~;!i_z45z>%|NWE6sKrWkfmyTdy%0nmCH(#A` zn|szdESQ0MpxWdYD$CeRqn7PVy>;$XQbWVi5QRBBJiM&HS@mG;=zY^Xn_+*fIg@B2 zqQW9i$TsfZae8TJvz(SMzu;R6Sv(_g9rrf*60L8EBy?82xL?Kb3R|vsOJ5)eh2D;; zn!v7|YNTQu(xk|#k?|(;Lqs_Y0P~(4bR6+~OCL256BIO79A05N)l z3KU~FOnd&7aa1V`KY}bQP<@`GZ>+La92fH##E=BT0SgZ+g!RHvffShYa$fiNA7@i7 zSKmALX{L9S!^Ob%mec7cj7A%<3leNU8sVx#OEinwcg9ZM$?e_-O2&t|HR;znJScJL zk;=b0D%3J3(@=#rQ>au}Hp*&BN(RQp`s?Qa>s>Z5QP2U#Qsn?ER}F$L^(}ZTu)NFf zTn|7Yvx-ej-x6Mw@lON;DY?xQJ2WkZGfd~>M>&0I!rsT}B^fE1zDK=vkSm=px0-ws zBg;mAQe0&vJB=e2lYYuEr@BaWb2qDnh1qVCHT`LW$X?Es*O@Z2-t}d7xb)llCH-2| z###ThkA;0-(~BCEDBE)YqMs%7Fn>lka=vT(ZMSCw-P5H7io?j(A@vJ{ zleWENpYE|;py3<8bFAOq1(#?#?7w%rOIiVT$R55QO955H<4^JUJx;uDy&e^^jF|f=L|BX6-MwtrtsrSEC)n; zJd0SERHr)3Z(0>unsh(DskLIrmrqh({chDqiGCytCV1ugHR`{f_4Rzfgoazj1wNno znwrYU@L654>O6e*5@bcuNaM9lweg+&4GdIePUFHoGPMJIjFJ1BdYg|GZ)$83$_#4M zM`&yLkq`>|NWY1Gk+x5x&2pFAF@RTi2%S+1eOPvqjI&xlWHIFb5umpFr9D z=I_Bs+cFK=?@1j)mV?>aaZHq}{7oGCyo? z_*LEbyY~EGT!%;S(FbBUL)4>z?9lgYVn)R0n!f$fHjvFWu$~Ufv_ie1e?y`-^ZPs! z@%NP~20q&>b=6 z_QLM169x|apcuS6%8;Pd?TV(uG2KGP)+=c+ZAnngrZ@d$HBl6^|EE>L_hHSWq3SXM zTpT8o6M#cie-aVhjsCIGP!iPru|WU;icP>B61#AeGBDOL+7`*!{{;vP3Hm&?JFh~z z=pO4(oFM-RPnVdYqF#Ci5^a46y2rZwD8g&RzR{OPb)jjXKY{7u<1qFG(L_>;=x}_i z&&}@D9x%O#wNOgiX>dG^?<;v80ffca=2Y>`c1u-;zSlc+4L&~)9vvHEUzWAnPFEy= zD|gScXZ!OnE$B<}$JD--ZY>BB@uOtFShF~`amit!TFZpVJnk{oP>+LhkG88=ty8){ z_ZlUI*FDv32gKa{u%}AnwU52N*iCB@H3Y;&TCcyeie<{zCF!(Y&9tgk6!#aaZEh58 zr%T_qSI=wSG5P{jSq=X8n&{iWKL6Ft6wdR(9);1@BAnc$P4AD!Fz-O1z;%oXcBgod zmOmjm|F~yk87V7M->=UE`OdlBkJct;pqH7UH8ZVR5!Thj`oZ6)ILsmbz8+)c%N2j& z#^=lWUCwkN=hN_^{z;^-LHKxRijpbFSJQ3x`Dq(}F1s1=3zaR}<}QdNcQ2;jPCK6{ z^D=jLH@Vtqv%AaaglgMP;i?E_6OtZwR&Ur;aKy31`0bu;MO+-5Gw?*2)`-8!(`Gt* z801)McDnaBifh8Q3g0LOa}E~}fm+)1Bri2lM>Hi37!R=3&AWVlM3!1_PZeldCrrY= zgyZqvd>MTHi?m$bfeF_EqRZ@z3v>PTo!l<(7hOf95`h9OiW0JKO}QE6vM@ZgVAgf7 zV{1;pDHY2WxDd!Ot zpS5nz?N+hf*pSt8AlZrk)OO~m*J|R~+UW;-m$Bi4MS*&LDtvdk$Hr)C{U{kNSMz>L zPc94T_Ecn%r=|5h`r5(?&=bvKdCnvNQEE$hHdJrK_c{fq@w*l@BD>s+Z^kivmk7Zq z`QsLr#;CZ`mxU9%EU z3hV!KvEZddkgI}-5N$xsKdGq!t}!ThsUeR%|3}k;Z=^xR9zX4$x=}4lY>GW&EN2u^v>W?vES@9agATm@ne6b8_N;% ze2o7~^UZgvNKGPGSU|B{#VYb~RIk_URuxMz^`u@*y8Ycr?60vu$bRYC1&!{z^5Euq zLAJO9)Lr?HHVP=@`lTtNL(MsDMl6h8F{W5n)&{`1Lo3{VxpriaS3D~eUX4%xmUg;N zd@$)GcuY9DAn?6Jo^8^~rg55KC{J6h=`^uYclAToQME{X<%M}!uNrXUq~!mWCFI5v zpPok-a(i6J=l)eyJ61J_xvzdK`}wa`E}w3ZZ7Ujp?}mR6+PPhD z&efWV$doH0zR)mlw(8nYk~=}!My9I%LDT@EzCV+HLuiRVe=T}T?}(Ei_>9p~6!bFySk|5y^nAKO zZ;6H16X>~4tEyW@uARWBGgb{r|5ZsGQ{=zL|I`@0dpFqsRlcnQBuG4I9Xq|I^1WCu zr?O=)csLoUUrV4wt_oJzq{p%hzp}QrZhw7Ht8zuYY&Nr&2ow8ibUm$bx?B2<5zZ)k z8al6=Z4QUE+@21|XF08%6#(hHZf@mm^eCqN(>7BE@OHUx48uJv;)vI(yD4rOEWc`S z9f7-z{BqIt8xDOnia2Ve1-5Rp=b6Q1UOi|3Z??fc>JdMfSbZgF_6E7ED?p$d-!+s; z*}1W`9_KWUx14Me$^ZC0tofzb3gS_^-ScpVx%6mJkO0f3@ZYb53*G-&Yplv*Xk3dF zJWq+P#8!z=J;2kZyT@1++n)kH4J=V2*F}xzVm~-JIpfz?<&z2E9qGf#o+3neAC{TE zC@d_u2TTSz0O6y2pws7ccIS>4HDu)ec$r<9SD(MNs(!{zWi?=wqdy({%(gKz^6BAb z>2cgqQj6;HdEogOd@3|c-GHh}J|_mEQs%RWMc?}(O^}F1JXmS2Huf`Yz=F>fUql7J z=pwC^&pk_efCE4e`p?0Z5J7+4i9@F)b*$Lle{Yv8WB?jkzqOd>%AxuyUtK0A@dE*rd6t}7+IAu4NYt%&s%Vjmw_|EZeQT_J}BQici zMgLq!BAbx&wwLihlbtE18?UaQ%_QH%1EtrNj=xG(MdNeqvlA>OlB{~ zHTMEuKwSb9NIvz*?AX{@*ttaWl<^4>+2YJRi6Rj$=-MX(u4XSe2cMG&Xs)G6M-mr9 z4`wVgC!Tm(myiD0Gip=?Y@O|BUT zGEn2!3q^dHigWf4!~R+E3&v`1nmW^2Fm4PBAY}3S%fi56idg_RUTF8p;6hmU`?LS| zV6HS4e}r%tC8#P7mUGy0X+)!@Yx!EV8TjIyf_xFdSlG(fkPVbe$zi?^old0R%gz4h zzKz*`viB_Uv(Dq}+!CX;wbtKkW#B3jJi)+L(*+ECi}hZeu~up6M&F-hJ`~XeEwS&B zDhrlrqcanC^l6B zhHThe9@>3>jZYZ*gc?y@{wG~tB=ZL$A|<^0juGLTbR2qS&aT&npYAU7HQ-$+^}8oV z^SBhaw`f(y1E*1 zbin(H&mrCWVo2wqNC{EI_~*0tz~qc)VeR{%E_uH0?ddL^JE03M{U#2dRQHzw>Wa^R z(U9Os`zaR%My}c=gS`dSb4M%^Pjta4O1oJo)8;rfUv!~uId=D%z46C|Co0|SWRJul z+H`L20J<{wLU-pCXB@a)Tfgw(9a+yW6ApFjlF1hiMmRZ?$g(JGgSMN+;9Ka zY2f8^Icz6eiPkl7mD-gKULjY9#vR)7v%tC@x=i1kMdDgLs6YB6tX&5UXSRn5x)`(c z3#jtbB}?$8yk(bIs}3;_J)bvAqew$PG*n1p)rvsZ+dPh=$*0MKxU^eHilbDw#-!vVuzv0qDMX)>}r+rkkVP2#szpgv6wwSKoW~i<8_Fw=rgQgdz84B;yLRXh<;l1 z=E?Bb8IuBZ&T^4ftJGd0iT6aezRo1;gv@l*Z;VZ;GCW=C)(1wYC;1NDo8NLhNoDpi zljvH7z$ut8E|g`Box$K{NHWyx-SD$lRCn|ox>yv7i_2oM$lXemk=Z~sHvSr)y-ePd z2j`t1!-vsYhlP@$eJ8puclI{l*qtb^D#jRLtka-^h-EEV{!|Vp3I#Grf=tWG6;fYQ z#d@t(-l7@<#eSFWUSbK=zA{-MqP>pV*f7}V0=|@&^vLlylvzZ)Bfljwf!3q^DIFTe zk~Hy053S+}$pmb?DfeksBOyA#Xi19CUyHb9)NQ~1EiqY#?W#>~OZ*l8Ln_QAw5I!D z>1{c14opbv63T`B^7_U58CFU0-&<`(TPov;EZ+JGpblv9)%jD$8Rx$Ia0BxM_@RI#rf`0(>ZYz z(zb|6Q20BN#~&jlFyJy63JQbiiijDW`-+21lLJ*yadBy)euo-E>q?9ps^jzOO|`Aw z#vFnrgFk!(WDw;aasL4KMbniTGNv z)eE)#-#h_=pFEDB(kV}!V-fDrAGwc&15|*un@7n&5MM{rn|K3iY=JhKiJGi6`YXy; zkoNoJtlZojo;@_NhiF>JU0=F?P{4CX&8gb;;>v-cMMrF85?rTBG?LF9le!>o6o4Kz~P!imm`>K1GARClwy8qD*nQc)qAoRU7fbO0b zkK?=`aZiCwmd9nd#GbEC@Kh%j8tYX&s(rS|f0y+|Z2puf=eUI0#8IeDRPeh@6x3aq zzt3Rq@_yx5w1|tZtsN^MOdEz0mHOoyw!^#MUCh!9&gaQPb~{HGI7=pewCAD4x;;x# z!!cupXYFJ6|r&O47`QUQ%L=3U{3Evi?^r#s?<8`r) z|Gq))u6aWa*?Hg*^v%b2}^bpwh?iMW# zBAxt(xnW0MXgD>_+Etb-~?;;h+}q(9}U5=Ey&oQNd%kTQYq@7{Bd-l2~8!0NDWFGXsJMVErD{pA8))C zbmN<||J@L$!hnroW$jm6e&mGKnXjN%?Y2R!SO3pYBkRtLUSjjI%Il`(>gHGb2jq*( zmj`F*L4I{h;_MZM3R&qjQ+AFl5Q-qbX{b#ag{p2={WgBM zgf-Dc(`@zKv}(xUx3$OB*;5$m^#A{^Pa=%pwtQw$A1aOtFNmm^{t#i~%pz>l@4TY$ zJ=G<1DkLr_S>-9~y3%4^Hpb?UQJ&H|4aj8C{iH+8YuCSljmIop72ZjMqP15@V8!_! znh%Z&q5r>Mg75+m5=l|wN2Av|lHPxikX>2()-IqIdW9C?FVvI9J<*=d+vwfUubKYB z@FVO6Ev*I4!1NCfhsUc4>>uR%woz0I@uQ$i&vg3Hnjil?8&GGZ3wvp!H36&UH??*Pl-40*x7ta+ z`#v;i%huFS(Y7ATS=-E;E!3QCaha{eR#ktvf;Ipit|Qvu;s!1RUFi#;XP#TW- z-yh?}Kx`G;BLT2kEK5!TFh$eU+%bWg7!rgAYJ$b()Eh#;g6Y)&_So|r#~3@KxgnZm zx@tftLIn~|Cqz{;1ZaR{?SECK_nxa=C@H{8(p&KAjb5#tUbbvJ$R4p(ijP$vt2!Xx zv60ohI!%~oO*1S$OcB?@*h~L5Oa(VlchDzf=NSI>WQc(v@!xxH*rqz~qEeEJ=G=Lk{9YID16kf+ElKd!a z9%wmX&M`yxt@Cd_MCBii9ogO$Sc{aidBr&*1HL4Y!(Y}d^5ylG2^KKDUhqB}D%r4| z{c#Mn8A|0j-Ozd4KbXP=hStN^wZb*Xctns0ba;Ox+~w2HxyCijAv}x6jkkeEu~{u7 z3<5gj6csf&xp&s&PUnu=teZ7_BkGGDCB*;HeE@oA%ok#wF-wJbjpNfFgeBSzC3>uB zd<~-q`2?)#F`y_@BA&Q$Ia%dbiK=P4K|uh-4rrc_;&00Te!$&Fo7?Hn53#_IwRbFE zVbVX)R#J@Uej5(z`tt+}VP8T*QnnSOg@#fv{jFv;r0+YE!*a{*6=+yHA^ ziRTm#&!-v0%iaY9bKZuna2o%7qE`>v*sG~DPC!spdbv^@qniWxl&2kMY_ii8hUWh- zhmVJoCh+OGD?>TeQv1{6H}e%j8d|&$f*Cf6h-m1P4=>JjZ<;c(!0)dDo9%(JK$I61 z7I2CHD=seS{vviTLr@3t1uIUWN+#>EBzb)-PTcxNl2PJthTwqhw>DZb<=e`iBUOBl zbcXC8*xIvcEBq3|$%Uq4`?-@sixt5Sy@2ed9a<;O?5+~Nzu5FANeA!xD_9LVmpW;e z=>;8ykJlWJ0t|B?h%o}Rwt*O_N6Xu4dVxlN@4HE2J@!vcX6A;Bst>!9hm`Jn)Sk_B zFF%WJ9=hG|UM4%*1xiY3y9<7uXi=ExZ}M^s?{S_(yAAN8QN3>;6`}cn4T!eTHKW}` z*gGG87)6oM)!T1)vZ@vDLkx7xIZ3Z6M2rY*uHdb)#YMtH;!pIZQIv> zLZYpF9FvL3R*GL}{{F%7(y2s)p!%O|6v;IoZWU`4KbTTfu|b(*hV*wSKHAjH*j z$jIxUqFHWbb>3(r<0G>Tu%hr<+kF3YL2V{N07?y44~f2=pU3`v-qeG)luw9FA6@ux zf7JGIsc_`sj6aSZpmG1q;*e#b)Q}RaV>@c(_~>JPo?|r>O}kqLqxpQx#c01>Kfi*v zCIW8ul^Q~6y6V%UQ}HPz$_fv4@qZ5|7k(RW|IYq#E~?O%UBQTpY;7_?~K#>-#SWZshU6ulgOd3UAJx$(Tn1B zZ*Y^R|C+Ev8b31d+gpN!sG@7dNj6;eQvDjnwX@xe(hdo9=L4FEH?tpTNYKjsL|tBx z3j5}b3%2V{XOEk$%bEN8$>fsSQLgMyXrse%f~bQ6*;piS=nWmdfG^*8ZeQd&TZ_xD z2o!^&RZ%FKRz-(W5`bzn04>oOp-{}~5U;rr89h%`JW*)-H(#o!8zxlLFp+)7TwPhI z>1kb;^VhY(&+D>gglHU(%X7XBSr$T*85|zzSFj{p@Q;;84x-ic-hOwnT(IWCdZ2cP zJ^RUt=il>Jzv#NgQ@&f$Q&5PNZuO6j#(Z?!qo5Xf<@R@8uY6dG_=^y$4uMXi(+Q-Z zzA?HfP4Gb&T)3t8%*@Q&83;$wL9(-VH_J=sLJw~t1-o4C0ncBu``MM*Z-dB~g66`9 zLoCA8Sp~wTA15Hr8pH$a&g9bZI4-6GUE(jfrN$qt)1%<=tFEmscoS1sZ&hIT0u%0! z*d=aL6bYk2hR~dHQNza$Uj05-P^?=-ch%m4G1005f|3O!zu#SDA=GK(DLhr!O9Zz_ zp}pJlQ}Sw(uiT#>*23ZW)$iXAGiEC&VPdygGBL28d=UuU*m?$o;m|+zqi za{TfrmcyaPgF!w0Z^#%2>dx$CP{L5~aB##yY;W^fXijLTDx6-(^ES}v^8-UF@GmY`!+>Z^O4PRbSI@f!U!|NTEA@lZjcCY)l_SX$QV0s+R_9b???@i?c zeX0vfGF`(gQF`)unv`rqo~ozX>|_O z<{yE7IU2#zn>zb^Rs_TPSYhOegYCKIqLJQS`qMH>9Q@z>e&m-wowsl918#uyMU(?B z&b

5Jhh+4nXcgZ1FyctuSipH}yx(S~mT>$&$>xF+Y$EB4X4P-Javp|LMw2rs_q= z(X|}T$exJTyjK*>A94%&m${^+TCoO4BtvsbF zC`__I->z;YOHcN>2z?XT>neB{eP<#?W1bH=Cn^e4N=9ZpqB6&YUTbY;W=0LL#d^m( zzMC?S{U-_^o?liSKz5T6+wke=V2 zzE{!Rr2IHuB_Ts=En0CE0}R`h?3?F~>!XX%$wbm!4mVXTv}u1WhMBvlkJr-TA|?g1 zD*Af+_%wW-0*h|7(h-FBhb9_$RS%8`R-i#mKy}0Y>Hbz6G6}5g2nEdOf&D-f4HLbU zHyG8BA)$~U;~~#<%3v11UGi0YsX(a;Za5MC30LEs?TMdZeG~UjZ&J+u9xY{JkO)XV zCWV6tgi`|;r z*2SxlycY{`Q(? zsL4i4VVBw^OTZ(UP7|3&(@5X!e2ICj$%{=I7kQoIBRTj9NtgRHh9HkWq2h_54d5(b zrT+X0WA#Ulh4}M$j+sTOQrK``xn(}Qm+yDtQn9?v*k$dgkB zz(PV|b&g9G{(0$qfylo*5YdX&a-|KbNPpowi^z^Aiq^)TB49us-_bvE11}(J_sJ5VYq)WAzShCNp9Q4zoJC5iehg#d7`H9Hlfqj(6hb*E8QOUC${P8N(M3FL`CC zux7+she5_pUrF$bgV~nPeF(D9UL^c~GktN z*YN@+ZmZwq_s`eY?huc-P$VRu1wPAT;$=N;Z=@=ni zxvc(&4a@yk-0yLDYytyEq|o8vDndQj^m18lO;oAAVLMYz?}n?t*JZ)fkop%oRo$}| zbGu&K3cT&xgE1UN+Dd#%JZWe`ENXt|=}CzxFPT*lV$VAmkB@_Y<0ar zg%XbdIN~>s7oMM$(_gobVqj6co~2-;7obh0RZi|z7n0q>%1Gw1#sAgQ!W~F~=q!wo&v*ZW&0-F}{rr!ZzF6^&q=bxL zSGweez(zmkI`YE$&T%gVR%16Bt6gT$PKyNIaE_=Z?<*HtMe;Y~?BtxG%geWbpEQ;y z7l%rzy80a`F4^C85j*)@>`s}=h$jpA8EAb|%kTx}?uoB_Z%2FV#7{tn3(A*Km~V7Q zd~$s6GeV^_?hV@}T(6Km9iXwZzMk0mf*4BX((I(vkG@{?<7DctIx^R3`+-5I}}C{9_ocWh>eiWVgKx^_AI|BA;Q-C^`xsLah-^ZC&N?SiV*d6H7c1=%S>NdUfX@^Iqj9j44&y4NbWT2eA^!6IseNAK*jM? z9m(9&j_n{mW)I?*hnS;!hz++VKmP8f;YaG<#;AsNA|>kp=6Q7&!05I2=aE7L*Ja)vMiV$JG$w#iuUw1Qg@p?e zfzUa^=Od7^Ixa)ztNAiRToVjfPDv=%t3a=P7dVU@L0a{2?aeXmdDaz|#O!*V*NBb**Y?sDT)!3EZT|Pp)?7 z>ex0mP<13%f|2AXkPj0|br#80qYcFiLT&xsa87@MiR5gntKe2L97Z9^$RZqpLKP_n z#m>`CfuWOfCqF|c?)tEH|AMcDeUq(azsU}Zo?P_^4YE<_^7CKU88cHXRKD)Ug!Qe7 zDlxeXUK)F}g2d*RYUI>t`sS_1XYcb>MD7D~ui0t*?OvYR*{3xu=MTA4V3wHPc<^KP zSEnZQYdH20ucCk6Z>p<#B~DsO1Br!Mqa+K>?_;rVc@+2I(lb|xNdhAz-Rcl2Zo7uN zySu*jc6h`K<3Yb(ocWFPjEugek<31*EMnQnJ9>mVJ}k;;tYoEsHJ7HH5MBGutYztK z^x;l5hsxj5P!tEB=x&OVfGT#(*L&?kkQkiTr|rfN)3uk$G%)2*Z_+_T47X&y3IuBO zQF=^F%tS0oGL#tH-g|P5fDF&1pRC^Azpu!_VHHVcL#DV6alal4O(Sj>4;-AEOY3{{ zQC~n1M|`z#9_EY15s*fLOodG11eb$WkOdQSSqv|yvq=6lE>lwv7&o$BLm=yY!lCmX z4bQ>fiC)qSS@{!cnQ=A38OBSSf3+9p9^#T^6-F)ntUlz;PA{OmuaPoWQpvvsHmf>g zib=xx@Wq{UyG&3or5;NNs%ci{6;~~ptXThg3{Cmj$#6v2E5nEe?=TAS2oKRyj+4lr z!K02uzH7)eC5@_om?b8NaGZ+pk1-6ZOc2{f-Zpz@^Ym}}C)E;6HU3gvtSmBYis{$j zs_vMPrKCL!td;0pt9fj}SboNaq|!h(Z&`UfA?ArHfv=PZ^8ik3r1%>vWv=d1rgT$- zg(v^fHrw$QA#DYKYK-Bqd)CiT7!Mz*%ViB7I&BXU8A(3`5iaiR9v`EMML!WI zaZWmBiMKe#bDH;aYW{faJf+Bbtrz?4+aqU26;GDueia5Xq!aSlSs%VHHx{h*t!Yw$ zci&r_QQ&c*Cc&MmM^Iapx;?1SG}G&$cYURee&DgBeZD1blqm_L%%X#k ztA1#2Y;aSZ(J)bG-!h{vG@5el>i2Bq_YH$fQse}plZ$>5<((6|<6r-^db8OVN5{Mu zFiyALqVvsQLx~b^V@j&F@YXDu$|zN623$W(y#c2Lo~nscj4EN^Mhq426QWm&o;-L7 zUeZpbujV4i$!ATPG;n{{xk3dIbFL@UdmUI$-pbGNQXj1iqq2oePO9AiD$!RABod2!9^tLWoUNl=6i_*mXx4w6qDXmCJua7j z1Vf4&xnH*h1qC@{>*JOC93R737kXW(AvkdPK)*Bvy2Rs#ruXWelVN=X{78(&={9&N zH!K|g`!CHc_qwEUM2vE^#BX*d(Hgydgc0A~!%xp@bPD2oe7Nd30B%?(NV`?a^0+Ou zJ{N&$GX!FMlvns-1&8kpS>In*1effFf5SgV8+}HZ3|Q83 z2LZr9hQA6bv=jL6T!F+^{o8NmpA*3c=X9$-0W4o^;#;T3Vsd+QO=ce-?bPi=eIwC- zs$zrnGz&jkXJCbqtSE`?Op0_uAO`6om;8G|n%nN*w8Cpv8Nerel3;G+b5AMmiJ#@)`lTvo)d0XgC2S&WZ7^m&9|vCpNIs}Fd=0}_R_ z!6~KJfEp0JU(RqELdJ8+VvXOqW$ZTwTeutAzmS1J{QBo;L0=rLtH?MAD6zWzcVe)& z(-DNmJQwhN4>XE1``Zq~hN!5hId#{~zqSH&5s{F_hu-itV6w5XHGVju+&b@2<7Huq z5E2$b`y6_?Gah0qO#;;LIrI7%6}hQF3DK&<`H=?}VM5#r2|n{*X6bm@w8!p{P6Q%o zR2W|&MBhY`JR9APK`T~D!(U&vQe%XNhehUh%G8Zl;d(JJs67Z+o_>{u_Cw%|TFmu| zaGsMG;pso4a{ZR@4Y*;WTXtwDGH^&lk;!n$TvrEZg4J|p z6TPr-i}<{b4{Jw{7xRFny)CcmjFSAW*8PHlY;3;VdHb6ygJ?RTCJow=93x?`IP(jg z_~c|5aL$NNPaj+-Q`geb(XpAYucD_vU_FJ9{RZu}RnIU9aRQI#5SN5u+`bH@DlvRF zkte9bjKbZDPlD!ksKXkG>&~uz90c3}ag8J{oR7ByCwRzo=SxfFK#S!&62X8&zs@1n zip1$k_ZMT~OMPF1a;|hFBeg0nogyCM+r8gFpk)nvKmZ*>zW9&b!-XcimoH_DERp%i zzAymwVaFquZp1t-=zI$0fBpJ+DdxfLs^5IJ)wHSI@aDWpaKrVz>1D-j)7Im-14=5a zYUf*evG(68Kn%H+_5d>ozc=m2vxb{I0?})bP&L zazsrcDOhnBU``|5IA@!~YoegrwI2d|7E_Z7@Cn0~>O-!{afY0gM`0tdg&dldwy4Zo^0>&PsM|c;ZicT z+paqcQ;PvwAe5C?@1U}Ap^ogcSS@eh{%=kMDJcp#*1r>BnolI>IoU_kU^JvyYqH=k zlfS(GN53Y4v&mjAE6eW>*(378$?tY-otS0i5GMA&L zBj@QEYz^faO@-;}LFw>MN%1$uy-4(9LO%6;-o+9YgOn_ONelQBdP;DDt$qSL)6aZ1R?~P=xe4 zu24ksg@754QogzPZ`mtfOnOVwd1laSRzfHpx(C;}Trp5ki3tn;{A(i>^e*3RzCK7v z#XwOx>z2`=tvqa!oN7G>n}4uMSD@y_JG0AX8is zw&yRtV~1y#tDxYR01K9QUB{liB3 zItD-?HmCS&JV`bBX}Ef2&A~}?0~};%md~Hd896g6Zldj{z@o_W&<-udhNNRqn>U^f z=IAN^oNE<5WJ|m)2NN3ySDMf1!|RuGvg_fEYMl)+H$8t@Xl2eUHy#FS*aJ<Z1`Uwn|iuM!hNMyfuFzj0A2^OQ60b6b#!gG&B2-1Po4 z91edJ|K~JVzO*Dy&<4_{_bzzO!Xrkd%i+(jcP&-Lv|e5JeJ~PfYZzQU&cLIn9RK}jQ+flA#Bxb z3wQEczFXn~8t&MRZ>PBns%`tf-kl!WV0(D~=~dFS$)gEsW#y5Fe+lcn5}S{>gD?1yrK}JIE7Rh5f_~iSB0|{+@8@!CD`hN;Gh2^ zL8pug3JRnXr`|XlaiBZ=h%F$LR6N6S9=P{%4WGdB(cX`H@3*qxdnybL^St#<6L-xd z<&<6~ z(Kkc)biWJk1ybLwskh9WY(*@qsdJnRWQx56M4sxo^8Sx}=va-tyu29d{7!yYAPbb) zRi(~wlp=#rCv=l8D1d-kIUfKEs^mD1NQ677fpYK9&57WCE_t;{{}3A&TU!q2h4##w zoI^xV0#76{zqjq)bUSR9-2$gyXk;Da=~B{Cl1n-$cYH9T3f?TE?t5@rpxmZeo1m$69%+$HoX_ zsD}3O&b*Z6{-X;LU7<-DMT{EF6E3y)vuBO#tnv6!T$7$58k68!{0R8$>lL&0wTXU| zUra3hyoE8at=GpkZyXA)i1FJmwCj@L|Hrb&$<}CQ?`05A90w@GT)elW@yYjdDSebQM8{12LHPt`FF)6zeLg9}W z-vuHu-5S>2Cvb-Km>5d&x4Bx}V*KgiE;D=;epg0D&_Mo6Yhl5Q_ve=~St2(0qm%^4 zboa$F&>=}6_jOkkG?+s&ySkaz2EToLNv;y%s*(=E$sVwu;dJ9gl>*4peG*5!HTsp) zYiPQdnt8Dnh&Y*xJ-d_*ZkF%!# zK3*~VC{DpnDpUYA080}n8+7NPgVdP4XPpdGve=0p2vpiEe#oQkyM(rdZUF_&)wHVi z19BXQ7~oizMf7v1mmimXBC^CLN0Vce)@;7VWg-BFOSVk5w!Yq{8xN*4H}<hpEHM9bH>d{ZUg?|4TG zNDWs9wC~=bd}#YDVifB zy19swl1jfOFg^2S>0c+HN=l-@RFWQw%6T3Q6kF_Q=@8q)dll%9-*?e3hzZI}=U?4s z)*?o~gNgp(a$Wv7LM9M!MJW()UQA$ixJ>qPw=#A_PYkeh(gwzWLzo&C|3kQ&}+K;`W?4DpI30J=?HNG%xF5BjTJ`;SEOz| zkzZ3!NSsv)M8;dDVz2Ek_2HFvJ+q7m!s-^cdTKZ0Qbjb)8TRt0d7 zx;bYMOv!MhCu(M9mQbtMdNQ>b-0@pBZ0S>`T-2i`C)M|rR}tj3Ej$bw4ZDyqd<`A4KObcNH}jbvXXHT!nm1O6>E$88LmSvhK2;Tq6=BjEQTO! zvZ_qgaxj@!{XB$n?FGvNg})aR9aO4+Z98P1UB&xR%lRFm@8r9>DCVc6q=I3s(iGyx z(8IQ*zo@o`g2VHYgv%Z5X;~pR-SRB~s=UboxO_&ZvbMFawG`B~{8e%R8A^A`sUs>_ zRNF?bq-JMZ1#mnv)ZQ&nSRdZugbr^5YKH)y)H8ohj_?5fDO)ht-nQ~LgOA<+cefC$(g{&`Y}vh6s7_gYg%<>v@d%YorTnA^qLwEx z4&ls)E3cycL(lF+=OVnD$I?JVVN^R0DdieUN?LNfX?x1>RV2g_>hfx)v11Om~^K~w=l%pJbc&beDA7$JKzLym>Dgm1aWXW-^b89Ma4hGvw78+i z%+zA^&lR7!LsM285$xTxq2T`~seWhSyNU6|-@cEf-2-Y=ItB#Ua7J{K1Lf9SNpjM> zw_WB1wOLTt;>X#HJ4??_$Olxg#J7VkgPhJtJ>;X= zqy|d5?KVbpN`Y;;{#4o8n!Rlt45hZPBv`0Qqr;cji@$$ow&)eYdPzBm95&}H5o8l=dM#u=B{uEkuGa9#Yl!!SPpTr!u^zc`JpJ!yoEUy)gnoR6E;kUCzSsoUy%R%%%^ zBudcj80E1pO(*dFU}JPct8Zz^{Os&yjg*-Y1$l($jr}@>(-R8QdlvVoeLp9SeB=N) z9Ssvk1c}blrYH6^6F5(a$JJ~?GT)iu>drU_FRlg(7tQgdvj{}Njx<{j@`Ky{qi z_nYge^!|YL3`znFZo&fJvg68_kr;;jh&42vU@_+;m69Mtq&2Ia1^S^Cq0+{QCR^O?7Esr$-9|9|R;Hku;o4_%tW=fG zGAMove;GmXot~#jz8F-&leBX2NiRK>W>RGN4^k zF9Q#!wIf<)@O4bvqn!XAU;_hF$6}7;1YfgVK=C1c0Urz#bSvP%iBEifZr-P0b9ZBO zESQu_GP{sV8i5!6Dh!3BOMod*JqLboKS7?4keU$Ug9yyf`AQB?Z07674^*lUPD~lf z<4py5nAmCNT*orgB;~kWQySB*L2AHyb|lRfMRz!B8M%3SuIt}O;#j*2m_JY%wx2ed zr%-e3Zc6q$eTtn-2m!f$6FsZBe{dC`t)XD%un*V>$A9tuGh`=#BUoSh{n4G#y=|LO zD{RPwl4@n_89QTLr7+Rm5o{s&{Et(1cu&|Iz{XBv|Q&$Ab&9)Ch zaTU>cdbl~k7Zs^#oR{NkwQA5*aZ}Ay=6uuuPQRbMH)qz|-Wq+HgsgiWm@Ws2RY%0m z7sbJIT{5`1b9Xrt^IQ|^R!(onVk;~bR(=G6c$>BIpP&k$jvPpV^l!Fr?xz6uG!zvG zSRcH?=O`P+?zej6T#PQ2b#P3=1kZ0#K7nM7vxV$qpN$UJC{vY;yKLxav_%wCq*9V< z*I$a;GT=;OwQ6H;&z5^00ffYw>7Um(HsFw^6af<>NA+kMGmelOYKE#Ra9yr7n;v`W zCJLl(1e`*TiDg?^K-?`>_YHK6b1i23%=l%e%|~u^=qFsDw*K_}{$R{qC=W9v8ifOc ze9_c`@|cj8Ej^1yG1dgI&@RQsYrzPGRCkofe9z36@g~btze}+ida$ZNZRjM`RMO%M z0q{7I3_hDdQ7dEH#p${B1NeM{U1A2x1b+GjP5LoNlMqYxKuFKwYb7lOm%2VB6rUVW zg&Al?Q^ASSTRblNAN=_n|&w0`eu90>s&oa9g z_e)OWq1!Z_B~EW$xhW!5Tbt8p>1*h^B_U*73XA0x@X3j&>~n9iLkVR?zVqaNbnN6< zL;%9Qi=DKF1Q<2ylm)$5Cnm`LY?yO0Ng=t^pOi^U zEe4zuZNxopb9ZrSuXLESJV3mEb5CFC_;3?jJgM1l)ms_Gcb@dWe?eB(2*nqFTjr8> zESfl<6*op|+b@}@s&oK$Cr>&^U?(d>wG3C~YnNbYEu~y`GA>T&mQ3Rckf@Em)gNJD z+y1T~h3+4+fAw9i@>Mv?|3ov-KtO?XJkx~1WV8LSl;)z~m1#+nyM1c-`avNH)mYoD zMarhjC*ujfHdN%Ke=Use2rv!~H#IzBAf=4x)&6GdY!EekZ}x3Y<^E)QyT9?(&$D=~ zSz=_OASbla!7w=|hw)`lK1VT!SniO0g_|KD_)s_yTjl&_L#NkYyDN@p=i7bnZNH4Q z){OEa{%M_>lcNPD753e)3)um*Pj4x0Uy4d$(FJ--9C!C>YmIOR5&qQ7p_g3h%}RE3>m08ZJWU{wev4t0&EV%L$|f$--qMLHtNWfi8ed|hDY&*{5n&d13V>?S^YRi|;?8s$5}>9)1%!H-ix^Zr602`$T0K)vWK!S} z9=EcxQr;!>iE8iYnDK?fdqE0$={uLB^_AKa@1P+0yhJ~>Slsc*oI)Vgc1OSk+f(RR zRV0bseE<@*oIq}xmLcJg*l9`Klww}>jGCMEuIrd?BaeQBkd7n0f z)drTql7bR~FT5KN2ZB`~q1tr)ke3#cK+ONJd#WW7|{rmj-!1j?4AjRFV ziKe-?(yuA0DY$Dw05{(c$C)Y1k58u)SKGeL&Sp(voyhmob&ijxOi9E(22sIRRH3+c z>YEx$s80n?+L%T3r*nF^aBfp1y|T>aY|+7=V)~I-WNP_~{pGc^ff-_}n2#0$Y-*5` z(aRSv)Ic}lk_mmgCdjP3o12@fUvii7&91EfA3k7gpjPaA-9$Z{AsjaA^VZ^RR@JvR zBEPd_?M389gG~($e*;*`=_$cO=0cO}u?rkq{eU6Dq%A03q35G|3$nZgL?JNHc54Xm z@;=*$?2c!Y_Fw^K2&MPaa3xW=I~Ef9uxGReh{NXrj#qSzdXr&Gsq}IKk?sR3m$7yn z=aWEe42t{gV{o#L33eL*wzLN8vu<_(_ke&|i>6|i;>Gd!S&v=)9q1<Nb@5Fjc9NdZ#%ze##&@nX(#itr+6>r|hpoVkh^5Byt=BoqEp7E#|Lz!9v zvo>)GeF`ozj1uwW(H?)R5^|Im8;Xki{wcRdzz^i^l>iQf;VERfYZfOj%zXKrLFLCq z%RyRht~Q8iWYZzhkr&_cLe#j3G>8notrD!^M9{yT*6hsc`uOpgyzwEDq5%}zO#qIK zTbMf;B8`^>$ea4gCxg|2EKc5_eJ_mng|a%7N8tJv)lP)AGG9|tWy7HUfJ;Oq#Y}(& zEi5do`RdV|$Ajz;=pC}It}5(7tcTszR_c#o8VY+j(t)!c34EBcCF94*5>I9a*kNM= zq%ZcOT>%ZI!iwsZk^W)ULPReehf6~DshDWP*WdS^k|jiK5LJnqedE0Np+$p&Q+v&X z;}hfa;%TYb)<9nPf-1!a9dpO=jA@!=p~W=42|as|I&T@(_yCF7KRV+7Nuqklk&rhK zE+WE{VwP}nzDpmxVe)$XOhZ$%vRS90y&gX*q)3@SF`o*c{10bze|A$d`WJUwpsA=q z4*>oeJow$o3IPZDY%O6aknsQTH!D#SxV{#>+JfYs^DR>`j9*IMcW&fzFpbq&y$&{V zY5a49>xoLqMEbhtb@w*NoW`PnS(xAb4p5P(6x%LnXr%+ zD7>NiTWA?$@+=gld-&@yos=89TBO&{Sw$av9#|TfNHlPT7sjmATBzVK^j?Jd^TXk6 zVPI%b@qzBL8Z=aXkTC#B)7;4@%3TsM?LE9e35UtD$;(=en?Q4AOzK)T73!X!CE;td zo|cJH?DpQMZ86TeGd#fn)3=F5&3vsKrN?vb=&uOSW>Oe5hnqd_6;JicU9*m54S|d_ z*p3uVHyb`W{@5z9)a9q*8@RN#rjjRcndaU^<{9dMEA>wNi6l6Z`#`FqquB6~iu?Q_gw9g~)>ld8_^wV<`vI0td zb8-*eoa4b5T#+f-9kt>f?m`t!;N?K^iYe35YiLQK}-`0zY8 zXociI109qU3KyIlKMUiy{wOUhOXwP@-QKs4XU)jD*cL?HS6^bMkIi24y}ADx_osFaLU1A=rDS7`9s2X#FHl0hP(CsO(}!rg6d zxGX4@ga)D3l2SrEBb}~U{U@*tB20xzoA_?=+P@$T+*h~3Zq(N8U}98ZZ2Lit{nxgJ zlf~L*txQ{Jrqi8`Y6Lcm z_5?-w4)hO_sYur#`@hRLJpVm>ys)%?^nNqmE(c2ce{z(+ zyV-My>42*nFNbe@QtuANw_sqdo_3yU%4{XbtE!@SS$g6Vf8RHTgt^3RP3+>kwn;(K zMmiQmtYRwv+J(@|%+3is1*RPVs@fyk*;ex)OnqyjRHr}*rx(smfG#u-4D-Z`L*{;$ zvxyi4YHluBzU&Fw{M@eD^?@}nTz1R7);=gawc=1hOlv^2!vWz68?M`M>DXgXHwB-~x zSkC$e_-B|&DG%Y^f=|0DjS?us zo5xUZq9R-I)I0f@Zg$%xO>|L`EI(O8+1Ogw`!lmUM>8y!e629xy{n)o^1Q)$R-Zw?Gub4+;bN`7>0KU~V~$dG_a_Lq%=}MX(y&()LOtEHFh^?j zKMqqB!%}mGi}@sR&xowkLKW@PUy@18NE!ZbdyYkIdc}N@d+kH4Z(l??B}wmnzur3r z)rYzP#UDK6|8hIul1uc0y8d-epG;HVya?rEbKOua{4{)2Z!;f#vAAKcY>@s^0M3TC zQ@eRrm?|lQ!WS{8d~;42wp(GNV)_$lsGB#<@_W-XBI)jL2W?4B<|CE=M(pRa2@<~t zO|H-R9ajeM_Mu^U*K&61jQ&Kk!a)C;&+|^d99Np#@;{i_H62W)qrTL8ufo!RvJMQO zVeHTMR8=m^y#)4eT)$<2(ng2@0-V<2fk;kX?s}`Tau&`qyV2ak^?E`3nFk*Uzf_H# zW9SoDmGot$OJpTA1!H%GXan8cz^`Yrq2Fp#EX26X{A5gj;3mNt$!aBj&;R}G&q?@O z&5VX=!*8|5)Dhgt?je@5KpnKg#UJ~aZ3{bwfW14Zi;Fr#r5Nwo6BWNS>K~F+dCP_a zF1of@L{#0eN?*%{BMYZbB-kanvf$o)Yig`hWJF_6RN|6v5Rc)JH>bJK}vX$LX%d8!-aIG=jRhZxaW(<6tnl<`6OQ0;HFNbHC>9s zGE{+J)iJy zSfR(xu#)ZS1{spwiR}eeOMDXEly{|0P*4zG4Q7=xC}?>FCrjLu?&ERL$#1;pBO%6$ zU!#0OQ3azZBiMH!4@0<#hD7t5k8|bxr@I z^HvmE0Z-&7b}*y42y``XZOz*6VojKysHL@a|1Xxkk|}I}XQxj`;cssXXKNCt@w2+$ z6*jw)UIIbsK3Qtu9dL!nv6^6I>}VkrYJ|*M(B?}mZ=9ewppLJB zjrrNoYYm4qB;9?Ncs}+vv?2u6rH<1Q<671IUQynBC)5~wG8mgy+I8thSK!3QZ_n;% zQ;c%(y;kOzWOoUxWEsa+3Z-)i{`;4@%}M&JH)6kLH9IxhSXm_n74GEZ^aF&5@(uPt z6gZ?K9j3ThbV1e~wc50z52+ZIZ>!r8eyWg@zD2@-8P}$4k|<*RgCvX1fVWEUKq-&B zcg8j$y-1~Nf>05bul`h=7Hl%U=GR87`XpfR%hZmUvRAXxvh~&Tb7g~+xWuagEi*pw z7bs}(3%BWf^*&8Z`+TYX63tdqWkjf}oWIf%a##f#TqDr`_w4wwq=R+;j8p)xkrh^7M5l9S^=EEcpB64BpA<+aO;LVJuo~tE`Tqbm=>g2*D zaCD*pcHw2KJ#Y0bVkYSU@y=)!0IKTWFs}6iZfY9}R92Ab`7Obv<$ZWPL=BCo?-lmk z{!+3a=YOwvI21C&;w03Gh3S@bUC}4yQio|slhn(}$jBsQg-mwIAyzvNU!BY)GDvmn z()KZkbzptqfM$&-zg$HXlInIJot>UPeL}Mbd=*To8f)pu3Osja8f=pexAY>NPUES8 zdk@LVq?_cX1M3Z*bvfjoOFrM5Zw*Az|M*Xw{g0ObTn2c)tjYDh!vw)0xI>0i$7t0 z4x%>=5XY2!;oxpR@M>4uyrH7_IWZK5^dspmO2UDZgW3Q}NMsftAOhD9SPF2tQH<0A z`SYcMsSwio6?e7y7xEqF`heBEL8gMQE8Uy~7LjQJNZ$yiHbi!3WM-C2Jmukql$3i= z(&I&=RBvC|w3{i-H2kx3Cc)Y<*Ur6y>kUGa3d`ARg{?>;od%=Ff#xcf1I#$UW{MP! zF9*lvGsufeOC}C3HjD*sNQsq|j02)nKm|^F04%AGL65g84DpGp701ge6^>{Rxcn(a zA#8|CpMvTVCM6DvMkAT=$*NO1b?WtNPv(#Z*vYR6$;$Qv9>+7*y*5&tK7Kav+V3Du zru&02X<5emj}*jGZ~pm+WgZ~9!C>=AOS^eY2mlvLy{yEOA6ID)c_<`QFB+RxVR6aBrqd#UWk!_KVQ|{6BBgY={He$KqNMGiCVj z_@lId>FLRDs~x<3xyb_MHS7(g{ir!!hA=#mvvb8gF+Sdg`=w<<)`Gni1w?DZLZF|PUT~N|y}veG zkmq}T3+e&)23$gCAmH~oF?Qy~#_8eGGWKQ1i)+(`kPcwiL`Kg!?D)yb5U|Peo3OO$ z7{P68za*BVRB6xcq}R`Qr_yKN1~aP#^H|;zN21Mfm5EfwOy7}4gsAq`_=M5oWyuq0 ztXuY+$gBkz^j?|YB6+JE@ui*PxUTV(hF&nrZ+K?2Ms?znr!>>^Se$yiBXA>$;{{tH zHz%|fdEmA9_Mv4g>tfRB!h7#LcR@0w;Oo#g1GFAx8=ZPqPFq0S2Nrz-cvf1iIDsZq z9U2#Rqb(T;`J$IYPx~|b0Ze8i5I2*IRUcpD>Dp(OM

Ya};p}qqM@qr;1m$HK|%HyY#+(@S~wWhYJdLBykJwK81qo=c?D%3 zjMnbqQft|vKkGgCG;a3${X&PIJAZqAKFhTrUajjNr;dj*uxib_rrKStjj{zA7GdN^d zx8Fw#AuFPzsXuW6mBp|p$!Ly@ny|O!<3E)%?X@OtPxz^o?<4fT;yG#sIWsUms*m@* zHTtdsuK%?P0-^NQPkpj3ydHi7hS*A4o(ZER%KcS-@0H#G)v$ffF1?RC2hf0>HvV2- z#B-Wuk;akGz4n~U6qT}UH)WvuKJ<-)`YB(E?verm>aUTSG`jqfn&9R9{G%{S6osuM z(UwGhR#)Yv!zLcy&6!Z0|5$Vua z^}I55U*N^@^4!}-{b~ByGA&pbu=*ikwG=-b=lO1}L;sEpRb(wozd-x-b-}7(vHpP< z4r6}|G;rC}QCW`+YrEN@k;tOkaVv~x0F!YO{ytW8`#=^%s;JhOzAvPgc;YJ0u>U1) zW=rOcE&ai(T$geN4z>dhHB+mV1bTPrdKWpe8yF=MX~dSi+aXLXjzcq}bFAR#ef36r zUx9L@$ROrOexIjqu-0??{O&9BjJmr1+{3!`MZ26_;D8g95*C9xq0 z`!}KwS6VuPujDD}z%=egyuQ%kzj=F_NUlaRSAeXn{d!OH-R)?110yIKFs~PDRvDFX zqLVaK=N23r9Yfb8@u$QcXXBndQt zSfMDfgP%oND4Pg^%q1QGZ~NXN{44>RY_y*+w2c~u&`X}G5Oo$z=57->D>JS0Jl$l zZ;eDiUa%c#!w^*6HY=_w#*~QNLI#zTz(cULX3g<(#iS=sj=6}R0?90 zR!*)(K4EwybU@v&9C|4A;H8}&0MnH4x>L^CU$C;VZ4Nc@irTNR?>qsnZw6$jD7;<< zL3kT!#fK8(7q2Yq=<9o?IufO&MSq=7k4GMHnJgG#;y;ilepQ#*uXIWOLFrvueKF}# zQ~M)-;T9zeYpXYa=gzkJGFkN20g_Jp8{|ud3zxE;f2F`8E#pot-8;5cQi6E?`kKDe zJ6Ev#JXno(llEGIjt_&GlaIRT)k$HOGsP;5W=o~6zxw!`---<}nWvm~2<{z2X16}~yB&i2TwV26^M_Xn;nPj4b zyU1kL0AcukbuYPRvd&6vhj=UxO!C5YT^w2ywpuM^!VBv+zfAL#PYdBAd(IdK&pK(2 z3C5~HiRitWg$|$JF=u94S>nstY4I;5+!kHUFSeJ=pc+0l5F8nQm7^y041#I0_XokjfEc!iQ-=cNAU!jK^;P-&D*65)L#*FNd|VS{a597$h&kyRm_ zUVPxR?S*?Hm7wXT_sl@P#J_Ph3`UKE{WlkN5s}10@p0Yqzbu4Sz6Y}0?*X^Te z*(skS2-dJ>tyEqu-4B~rZ_oI=bURGY(WaEXJ^JU*pVMyC{cLLqqcbjD(Q`@#moN1w~|+roA>yna)g4q>7n?JDZ)kld&h?$!N=zo>c96M4tlK8?Vr}#OrJazFkbNUgDdRK!Dgk3S!I>W^A-k#49_Yi1Z`qBw7IGeb`n=!s8*WW7_&HjV$j9ym5ra??D3Um_@%(${ z>F~9k_VnzezHJgzN=a!Z;!s+%`CwT-pb{MZRH^yA%H!8~y4WyRRt#E>icB5%*JL?` zpnNS$&&T_&f?hG*V(kOLT?Df2A&S)NgD-}!PJUXOF%HQ>oYCS#)I>)q$ytP{lU{jSKPTzR-DoO}so}>|2R4EGdP}Dw+F_w|- zuN6>JW^aAE`T;>krdV!Y&ezd%D{+IKKAT$5q&VXT4F}qX<9ANh%NHMY^)ISBT_Jjw zl@_i7vsD8|AQ`RO5vrn5Ck`hfD*4%lBbJB$L`^S&E?7E^q>cCw-`7`%;BcWY+%7Lp z`PQ+06-9(nLXaJp#*vx9%os}3`zNgPb?$f+YqiQ>!y?@n4%JyJYRFzG)gX z^Kj;ZJe2YWPzn2x34ziXp!MzP<@}?~r>d$ZTn?!VYoi#zH}!E{=wiJgD5&m(Y6b1e z)OBUHUDj|Jjq^Xhtj43lb$&`JzD{KI!6Rlc=VT85{oAdVfK+f9I2P+XxwlGDrHSg4 zUT@Nx&dNd;N;Igol~fO%CcbOQ;)Ne-xb1^H9TFlMp?v)p%xhsRkfZYti@qB)#0suQ zU7>jF>~|uh(_Q-#v5gZ}WFG z;$vm$XluIx%QX74Q`fO5tf^9 zkGDv@BiD2s`Qs&EOrk~sDRCUr)2cQk!=ykRa9(;2*cwCWq~C7Dhke+=A@3c3baz$^ z1SG-qQ!Ms>#Q7W4dpnb%s_%gW4X0dk5BaiMs;oDE2UdX_8mos5Z2?Eb-8A2fQSykc zM~&ch2zvZd?^Q9J10hHh#LcTHZ3wT?8a76K2Ed%f0yl3)1by41-!SCfIFDnN$jxzG zD#rPytMZccR8fJ2)b(lTg%+4%WNntu2?h&tvA182CAowr`2&5W(b?IX1Dj`yvrUe- z6tT)@nj9YhatRU@jrggXtcgF4<9|>6M2N%yI#g7L2hykJ$#ViwTj*#V3h+32At5;9 z@D`*xFS$XUyuZUq>HR;3_N3rM?oq6@oy=21OUA|bv&ZF^IKFrain2473)pvV`}W3q zY_0qpZ3(_kBVgc@I>OgDAx{m9Q{a{g{SVwgu%dY|-6MF`4_R12+6&j@FFJgg3;~oa zo`!i6?hn5@^i&B(h!EtltUE;|FswW#Dz-BcIN#1e(Eq<=Loc^wSAs+0j1;9_{gL8RU31x z3{j3f&7VK(yp*mIg5az`r5KCDSmZzRE1SYt@yL&tqnPXEgHQL@8N_Wi(ufUxFxl+4` z3r@INxh21xmx!Mb<%K3Yroa#EqOJVME*`eJg=DFh|ucB($2VUMlT{m zsLbJJG7I@>GbxU;2O7jE`H+urr8Ck_KaX68-Liu*wC z=F@s4g9$oL%j}zD-%X5=MRMW~!W8D;E|bD2?F0d$K&?Fa95#)z*7(z&Cb(D6fIB3a zR^CaH+E(R~d_OmLL!lIOs!xu7KffG zV=sqthEIrO+!#R;WVj^vI=LE7s$MZ|$P@LC z4gcWKfG$qE3$$KQKVUH8LW1C{zY^bmSz6TS+?zfDy2uwcE=m4f^?zR1ython`AO&~ zgUj$FpW=2n-bNSe)gp3*Yk7vTEZp*}Ty>^g>l z=fYD}TxH+%UDt&$E-Sj;S10kbG8gGxX zt@rVc$=T6cGb!?j_uI3iW7lfS>!h{>MAzdRy385Gnz^!>O4U(&G_~7l*Qi|C-1Yyi z3m@%w^k{(~;rdt?B6t-}Ov)T+%hgV;0TF{H6%a9)o08^N?!>{N_CGn~ZB$7Nv=HCb zLiv)EVZ*$j!|lsm0rYFfuvNyb@_Y^m-P%)5ot(9Hq>2#fx{_&1;S)^t>cGME-91q2 zybqv|P(YG%Pdx(}Fyv+R0XODNu%~7FhD%UXCXX~U@u|ql($lj2TKGyi`&0fQcG8)! zngjxbaM1eZY6&E3HfPPR3#7r?1F3;wBr^=`_O+WTmn5)RFj;H*>U z&T4LSkCZE!0HkOvKS}t65n*`#VQX-{i9q2PL@pQ*Cnq6#1p>J3AH<1K&eZ(3t)_*9 zJc2~Ho6DksAH^$2xthqSHTi9Em8zHn&v2I1GuuxYT}SC@^n}#rxqF}@x_&$F-yNaz zjV(3+Z+Y3!xc)TYR?(#0-Y%;%Vo^92C}c{jh=jC8qe^1^&_Mt7FXu9Y%qPR>hWL+0G=~B>!kgDm zB_DI=hn@M;7MCAuuid7n!@<<4L|PX2xjs9WLE|6hMBfkU!HS+u;+hnolXPWK%SEig zdkMtyMsj#r%#L>!t)Hcup~C;F{;{~YxV4pYPoJMIo(7m?q*#Cs zTs<@38weo=qG)oyoQYdKs(+g*t}e>=M{lmYH)s~1zvR3Oy}B9=R$ZPa53cs8J_$I? z&=z&@z&l&{DA-JuCNE-zSTOxxNR!n^``LHDDixnZwAHqI-Mi;QiEV%YFUYm;JGX&{ zjx|^z*x^eBB}G}Q=eAo6q0@-s{0#_BA2x#c`cPin%bv=qmmBY2A3j>HNS{9XA*VT2 zGhalchj5>8Q2XDFQU!GjddP~dP)}pmxmvg(At3=_x-)o3lkWGiX@}p_96#bG4Rir(S zWF3vK9%(u2+8xDe`~j*F-E0~64~Vn99vml+gChyE14{;+vL@$PrIl-I%L`}2KVL+E z&`R35jr+IHT5N$=QtN$k4E~SW5H=Kvf7t6Th7TLac+}0yoLPGtpnuEnx&HFWp2d%GF#JBp;*SqxMZ7 zJ&Td)^2`^h?~ads`_uTJ=T03O%bG8NQS$C{FLZL8ea`jmmhIXg(4D>)vQWnej=m5y zBlq%qZFIl1Ux&%P5f2I>piwbL$P7%E_xhfW9uSQ9AMJMMH;&dKlvVm?KV(%+{Z<=` z9mQVlxV6N%|L4!E|FH69nt;tP&Fhq3oxZ1DZ@iqgK?pxv?)$>!TYRLSK~Rf+?_biX z$IdpxQ-*q%hQK5&YBPk``dURl-~2QTgm>r^NGKU3avD?r`uU5ds6UZu!tu8PSQG1* zkocysq`eNEU$dHMsdf>J$Ol<}k6lF#D#2wxp9j^(%NJVhf{o5Tr9MCj3E&lOptFQe zp5`CSdfi+3eMVZQ9ppY;S@fuhj%UN|bXJhz_rT~>t}NXet12gp&UOCF@%Q$A$q#*f zW0?2ELgoCxw5Wtb!3N??ynIa#Cg_5gU-HB_Y>lbO9ut&dir(IxF0@L#C&Y?=NER9T z;1An>L^~aeIVmLKDT18`gy?|BShei9IQzg_b6@{D@%uV+{P9?*={ui04{A)-RqwQU z{R)qq%Ov6fsWs(QI!4zCG!fyEXgxjhZ}qFsXxOo#ydba`r1advaR0iczH~8ruG>${ zI5;wr0BUkO|AwWZ{LKQGj&~PJ!4&&67_?M00`%Q9;Am^+8Hj!9uMPvM^5Pwnw$!2` zHi}XZ1-*ZPROm9qU@-c?T@Iw#u(u_m@|^Mlxz~xnx2eA;XJy^B8_g;CHujdlMJAd_ z{_12kF>|jVXlj`)YIEkm?q&iWB+}%m8M9e^{hLA%IPIQSn*rD?G!tyvaB7u-YZx|p zmkcv&nP&g>wNWB23DYOCM39#8TZUNA*o1C@7I3A9{a+P+fV#_TD$JzSca)POb_tW$BdcDWz`v-h3mruW*%h5Ua$9dcz_uF+!g}cM0+}FBm65;&w?J>DOzdRK- zZ|c2~WdMBgmb&^st>M?^kGk=3bBIp6@_D_I4_p`j;2S4&hM|i@|Dl=9`jzflPW%qXx&c z+WX}l^V_!}@fv^cq$@xeecg6hSZ8s^v&!8Q;VJX1Y#@jDTJUWlkfhrTZWWMA+RUO9 z$7xb7i{DkV2#G!zDdk6^AKq8`Fn=CQ?ZSD989*D3d{(ohNS1SyjiM6_6t98jej|L% zA!!@7@E*UNlf_RF2M6a|)&yYt;Sn)BZ|rnLZS-R5HEq_VDs&~k0(m{JX-`+QNR?eI z)jua0>gzZ`lt?0?*^1*{`D8&KLbZ3$I|_4KK!q$`ZH@_)&(6)=f5Y8z?Nau6^EX{$ zeZcI!^|+{7)}E?U{y?Z(12n&%hD*x(fPT#Ozg*$oAF=tmU^=}=1GG2Z%qz5%JapFn zNtM-=G}n+FciSTATW;9=j^5Mrj5ndxhChdX`IE{`Pco- z=LoXGPr($CvdF`Jvja_@^~}uYf?5>xf)DS)xA7z+tEiOL{Y=f`(qDMmKM4dviTiNK z@Qo3CZVE|MVWfJ7CXD)Rs}h2wG)r;;O!Rk=lu#j&o3~b zOe|D1AdQlSg)kHRv1%b_G+J<8NT!y#drB9K1A(mj6-$89Bfl`dji(?z*cd%pf z?G+IBO6k=9Ecmh4FGF)YUs=Q34O$vGCP3)qi z4l9y+X!o7P*d6M0HTh+KU1Auki{tT&l$Q((uCGUUZ9${JV}q7(EDi=j{Yzry7iD_p zZLW3ni%T};}`6tk~%qo#5 zBt!J4P#X0l8CLWGZ#I-^|65xtm zk0F09LwTv8JQ0ea-?AKD!VOwwf9LWOtQI`@dW(#1EcNHf@Ozfg^NwvbgVMS|z7GmP zgd6@6FMcqphm~b!lxeFt*q$a^*;AL5m6W?Kn3^xqgoBn>a{I653{}N@oT_IR%TXCa3nhh556iO%1gerbS=%-&eOxK##$99Ttg|fE9h9++`2wh(Ovh$1XH| z!;!qL+0V#G%7!`-XSD*VdR`Wnr!(J#ZYjtAD1M;9;&%XS-RhD1EfjJ)eF`(NzeQ~a zZ~QncG2vn?4Sy?(1i}87uf@RDKHQ1T^;|YjSf+>JFsxUnl4jj)ev^`-1Uy0mUlHbmX~MV^)1U+s(Cw?W?}W}&F1*GSEzmoVX4&DmyTTDGhOZfSF}%@-qUwN zi+L-c<*6=(>LU+qPx@a6^L`whItZ+TYA-xT^PO&%IR?~j*!9L)KGwVzT+0E4F95X+ zrkbVyYYV&M!cN@6%!<5w+4i1w$LjduGT-Yi1cK4IyQ6!?gf&S}p5C2j&oDndfH4|) za$eHyeQY#9`@r=&0=R9NJM8F%y{^_=1l(wCYoocA2Q<@}vn?TI#AhvHumjhWZUu!^IIp%42x@_tI+ikz&&nVtS|d`jkj-!=E0FkUeFt*`ZnSJMS!D z_(dPrK8b$4MQrBqkX+-fVzbzFE9RyCd|7;w(b3`go6(B-!hA@6-v1W%0~$|k@_O|Y zZZQ9+@Pzbn*wPQy_kWg_WtSgDR%w~`8vQ-lC5htRqRLa;yTMzExTnO^DUjt;bEL|t z6FY)vhn(KYJA4j8-NM7eEz)GddCq?S78VRx*aBW3UM|cqzH<;M%eCg`WMM0fQ#6nB zhWnJ6*dCAt;)q-Cw%t1}>$r#Cc}Ps~b0CIp4Qgg{ZwRo~ zN;!k|n=^yh3d(K8?|b!!j*-)n{o%yJ$)bgCJD_aGr_{_Jz zEG4ghb-Q<5+fGMifiHfEKW<7Ft?J?dGthRGsd2k4tR-Q;7Y2 zB%eMD$xh1mbqa%<97VEZbHs9T=R0_$c_I2!h4u$0-J$+xAbIAV{Oq#QGvcpD&JeQ- z-F~HHC=^#w9wNdNFiL|e;d%FzN(V_4`P2p3=I=DM{odT3qlg&Sourubhm0bDoIbQ@ zz3N!dP%;O4LPz|_G)tbnsHkvs;g&WFE>H=x0C2{#R@TK;Gw=N`fxX>i#V0v;qXw+r zA0o(y&2-wPf3{n7PmflcM7cfBW<$X7F^xp384q!?3k$zzS{wf&&U8RO#eG5e<9dY= z170FkFiLXTp}zL@;9PdZYQi-dPAv!o2eHGAGtu7?u$QkUPffdw=DV%my^XtjlbDXa zBI|B=*xrgra<^t`L03$n!V7UpRQfq5(OQcMFC!_#yz3v{<4MD6gP`yHN`Jq19(BMi z0+{;K`*LoKuMcf*aG1<4kzrzb9+AjU#0|_b@y?AgJoz%h_IC9swab z@Y$#20=RJwMWt9>MH7bFG`*DwV2YIjaX|tC61_Gj3^DWDKbu>2q-F7Ty{)VkBM`IC zH6gZ7ZFE(v2&flmQO`R=-x>wQ3yFL(q#@o`>nJR(So=UDsng46U~nQ-mcZ@($0)x! zSKr8I=!Z$cPV3~vM~RZ}Fd}vQ;gous$8vvHF9n3|$|EkBJmrn~@jMlO{`|>*)*giy z&iHKXQvf-mbmKZ)^n(JX=Z}Ldziz~3S65eM_vzX6k0HSDswj)%XFX%Xmu1Dhb*oJs zl0GNzIm({_ZKl0mm#N@VTUXneTS!6c_cAhk!3TAlP(}Y!w28fnezk}DY-c*~%;`8h z+={TwOww>~*fQzQuB}Zat%plX9;x!m_!U2lC>Mjomc~U_P>z2+$`z6c}_2vY`Pa*}A)H|xHtDPE>HdR=VEPSj5 znK!JY5IfJIdDl?kc#AS$da3yMdKaHQCZk3n7v+w>SJe_u?3ul`j510FN&IAt60vd- zpC+!Pa>$s<$nw#AZd;xL7IthPqsz{Wg4e+cr5t&H$VR+*EuFiCE-L^EBH`tz$S0s! zBjHVYZWtvKjZt3U%a?2-TV~qYk}jlMMCrjNOz9CPMZrM?f*~O`S9rhs3&N4{*ruhW-j{rFg$c(^{1Fl`TsU|?8zfSebZo&NHLte~Fuol(E@NaZ^tZa*XLX9<3Z zg7eeMA0*uVc$#=-tVvF(O46Q3{VCWCo&GFygf~xf^PS$C;wK{T;kScB+k+H;e&cP9 z*DQYtzthO|P4xUSeVV7Lq^;cNmu+_QDw9>@8oQG{Y_xBxSNooIe`W7?Z#HvUsgf#B zQECF)aHB9rrN4LE^J`%FPft$z1Gi~V;Hb+5W|oG7&xWs_h8%Xpv6ZiwMIbrEa+(&P zbAh_k$Oaiu{Y+6LMRdBj?TV6_ zcDkbzS}Ewd%}&Sn?=RxF9b+K92@ccPFhSPHV3WPf>TWvMQ<3$)7ucv=&y+OJk`fxVVb)`Q* z&U$EQXw~VP)#XOhvfvLdr4x}hr`0|7uD(9U{wlE_s>$@#A4(j1vxw#LV)IHMT)fQ@ zhU~AqsEHwn0~Uot7RpxK@K#jf?SeBMVh$B~g~yo`VeBdx1!IGkMckPkbDEyrfU~&= z@kR!dZ)wo-Km3qRbMY$L>|7mpv?C>w8#farYLl!NQjAT|bu9-Ay( z{4No*szqPlcnMl%7@70BxFme%Bra-Ctij>HE1u78qWw!9kEf^CbJx02$`>P2Eo4zI zINcvUItHys!!_=XvW`WxhxH6?A@`gj$?SCNY+HH=!h;`$Fjm$jR|#9SXr!f;Hs z0iJ%1is~Ldwqvaw_j!mtzlw>45sBrUyXqWzW!#hmL{|gXE$uo4Tj;%}z4ysZf1hIc zqm{2@4WDN|+LC^5qF(M&F>I%CeQGSvHLWH5<*%la z-{VSer!qq_w9xMJz0@ zWZAcel|;BWl$jD4BKPXN8FN- z#dH>4jI@yo_n=p2ToA-e&kAeI==h!51uh2%b?f7jZ&`8A-rt<4ir<>5(;x^Y`0z3N zB5Q`I?R|2Z%CrZ$w}9#dU6!A;qouu)Uhdt+^*SI2a>GI7>_}I>hP^>gvpkh;PpMb6 zYB<)AfAA~Tdd{atv&0$Bo_$&yyEU@D&LHv#f|O8&BA^l;rY$modI?9!s89XeQ`Uus zB^K?T8%duxAm%lOVjpd+$LUPp=h$hU0KxG5_z5W~3|5aH_w)1k>Zk{#?EWUX zdFCZ@>VBMiQdvM8{yO<)qKdgut%qOpcyoPd3}X(cWbNO$S9V%PoG$9&H**~I={${o+BO|em_J)zB^hK;ZGw-=a&pG02Mb8IvVgtC{ce=xppyhqFZsxeOLiVjeh zG2PnLIot?-|O4tpXmwZ-o4Q0eu zQkv2HOjPlCKZ`?kKg*L3j+51`y)edtpJzeP{Ha)$fCbkB0)(Nleus})l~-`Xxxp5# zLo(wAQ^)A&8*B}YI6BY9#^RY`dko7m13!j3%Zb&29MLEcUAEl2SL7ALEYtC$?X`6q ze^YBptw!p<%NZfp_R22yZ^qT9-KI%sHw#Pp~sCeY{r_B3IHs6FVb3zw-IL z%kN(MWEp3%&@=O>TZ#Tcf`YSstP*QuTbB0RoKHcR`}uW(xQO-l*HB6RLINTlh>V@8 zmz@ajhDW7$Ow%|=4iDC(qYJfJmy3#Nd@QAwFBfSadl&H&NIL#KR=q%=O8h}%%{+=2 z64xB4*lVD8qON4o%iyGXGx3JiL~P@|I~p1<*mOcnf+%Ojehe0WVd_(4z0SCPA3HIorvTCLvbDJ(eOdQy-u5>)8R-)J1EB%^ zjdKm??tYd%4nDJb&AZWe&wm%o)F8z;UR%GjTcc>s^J>u=I@z6<9L15OhKlwjet2@D zbYAtY>$3>~MESL{S&*2?SRuZh;^~bS<_rq9ugd-D?AMkEU=qh{rAN=l-m7qNi`78^NJp~4{k&{k`fJ&h(sQNhqR#CJ=G1ecQB3mwC=vlrqN8O>~*zF;Da>D?ZWUD{&y0Ax^$K;tSZ^lgVY#TdZ83^!|>0ZS@`1NH8-!Y61 zvo(ApBbr%8T{ws5g3GL9h2qq3C|Aq+dEgCa*lb>Y14!PeKk#bSVqh#ctw{)1SNFI< z!>D@gR{aKLv~G$&@Qf3M8Y*reSil7gcZOeuuySz;z67{X>&qdjW;ZltY%)IV$t>j@ zM10-$hhDFek_{`jKnMmSK#}0q<|$hK*vHwYrNdM7CJqcKe|`UqAYo3POZ*i$1beW^ zN6rZ2SU||O9Kn-!ytVZM9?7joLLofC-0|EUWSGhTyXcqfi9u1zJNI)KfFLdF!*^>T62_?tWwtd_v5Z@H^|QdGi-)($QmfNC+ddE zY3G?3$;@AMFyVz-OZJV4s(rZ9)iGa2MOv)1k;jp?bC1G^c{bwYnOci-%-39TZTi_4 zU0s9Wfl37?`8#a0-kXgSR9ZHb?;b_nVGSFqulqeXAKC5#TDQ4x@>tPAV{&tHbit-% z$M~y;Z)Tv?hgGTsIjCb)LsJtRkqe^{ea02%jwa65_~jLD%;2F%0^Au(%r^F)>cE(t;!{b8pyg8#VF zFU7>t?_{ocDt2rFG%dRM3g`h1l49ADo;hg0WS z{*J#(eG8e;;)B-n&!TGhb15}h>l&ZNT^aG4eB=GE&aYDz|8BVH0~P+-vaswj<1mOptv9}X%;NWl2rFIky*#zo*<}-NwcHr7pxZn|MT57bfmu?AYYK=6ptcJ2kD~V~&@=P>hhT z!%%`L%AKb6yNvTGmx!Ry9oies;svBg2o+}K z{OFgXv|VHaLbvywj7C^)=*~FEtL@WeCBikUPU3P1!y~{6-jT)NBOaqc9GmD-D2ds~ z0~;IUvfLv=*t~yGh#=cTmU*=f<{w>(HgIltbA*L{+O2E8m?dw)P=elMQ9F+xjN*)R z{M?)tA@teFQkAY5S`Wl_HQi8s@cLc6{#1g zz2~pQy?pB|VRvu*UOA|p?ebvVJ=)V56u~bSIrQL|Dp_B4W&TuD6c%w?+2p~H2wwbq z5fU6o-Wn@$raLk+D7Fe{0iADWtIKDeF{hp#+5ajF!Z__i-YCOxoJMy%!@|SRQL$wy z002)9I${Vr~cN#e^3mbvM_e4|gsoE2JXXuIS^E%+40f8?i9#`7`iD)^Ws| zN^|P?Z;#ea8BXFprIIRP&7fagM5ci=KNy!ZGsYo#Z`1o0VkyE92}LS#1enTMfJXYC zfr>_n^QHRaao&Kp2IJdcaaeXBhlr?^GnKo3s^G2p5)xL%)Pm5>O`NA6dy1^zTM8!hMj$6Q0tYOJCZ-5i*E4Uie$5A5 zYT3?mUPb6mlKz6v-ehGiIKpI@kWX56Xow|yA9oXaCr_v=bG*x%CVVjiLf%|709`sh zG%|C(eT$gnem{SO+*>zF{w4om-xz{az+!V=?* z#sD35$CgP2WBTxW@FnKrs}(A}+WL=S%L0!=t}Qwt&~V+@g(#?$eqRf%sKrs-LDaGU zPdNYdPmrk}L2G^k$}>h=OmQt;hn=L#fQ~3S3MCb$P-&FcANqMVg1KG3m0zDksrZ_& zngnUMZ3dDELv7sHsYvX+;4veS6xV7ANzq{+VEcocc!tIc%%X$5Rd5pPWmzA2^q|rN zlnLok_74vm;3t$WY1iQvBo)S~D1US)$n1PhA0V%VOQJ}eK44tLH0uIILEb7EKjisk zXreRRG8x1ZQ~cYWsb{hRUWII>mAS?&No&{C`|V?#Bb zfzLAuKBvdi2K&3oT?QvEE>Y_=s8Mk=x&ED($^^BAc8CiS?7jGuwVxLsRu45G*D+1Z zhhFMT3OTPe$Z~nZnLXn!WSNu8NiNXn&M4?rB~}Omnpd zk~pSyP{$ND>yf7)c1IY(gxSo@yp~^J?nwwi7R`sEpXo2QT%t#H()f?>@x93ottq~S zLcfPsTz=kRF;VvcL02ZhoZa(9@C%$%T5f0FAuz{jf z>VS(wu)8onuLQb^XJ7q9i=f|=x8wiG za^iHls&kiNi{MhFcL4;)RDB}bxb|JV)2B$cZLrN^&Q~+T;lPxtgjn} zQtpIt!tOUY5g0pM3AXwK6=lr%2{0CJ{VrtyVkLYut_tKcsf4qx9X=#s>G=^tD0ZHF z?8IWU{Q-yMwn*ao*?`FvNgeDJl*faF`9HmKV_X>m2{=}}GW$Zto3i(V-ZH%S;e+CEk!@GP#MAP|Y zme88wk3hq^Z$d)y4vq72h2fX%iSui!zaYD}4ZgCxkByOL8xOv$5Uyz*hqH>1sp-(I zzVj!83ha^F)9S4iZ$|Ig@tPnFg#j#ku$r5`_Xnn&;deH(ab`HaJjRqk%2;S4w*B)Bfv4mMk1h>NfoVF_wuC@$PitgVAE%hX!uexV2Z5Nwrr2i zCD>o)VXr7#4gGTwKs&?7#bvwmQeHXJ`{?-3`NPgQp{&hcz5T)a?g@us38l`1viEUs zCR$vFqHg^R_ZMD%{2~v~ZG;Dt@c&=u^J2@a^yMXgk6ml6S1~Fq#Kkiy(-^ zPuJxy8H^j}YZLXQr6%{5IUl}_lT#V*)Ag`V6|(ZlqdC#;%q}>~KQm2CjT!e_#94G) zlw4FbJk>hYNiH4ZuF%=Ew6)FS9lV9Xec?3#Aobi<4Vs1l#R8z+6WRXdDPp(uOued5 zo3@um?_n|QicLSMf|!>=Mr{^`la$ZQG+UGAHnFp(?cak$ax(aIIUIcLRA9lrKmho$ zcOS)^rYdeDSM*ZP_#`2m8Yh^*jXLYp&xBc{>u0~Ts{Dzl2-jL`~JjcS4pKwnJOP|y{>z1RyyB)2M$cQ zq1)_}r~C8xaO<1co5Q0I{BG)&N#UWixQZDLxXAU5dC}l%_muYw3U6y^O&*#iFy=gx z!3HssB8%7#a(X}-AA$aMtT}In#d`*3!jqi#go|W~0FrEcU(E@(YL`PJE2g6Mp5<=~ zntvbXEKnRu6ia{Wt4xxBm_+|Hd!y{~ees12x|4@c7m8skoG%uwPBbW2gULci|GX|D z9MG;{LW82DoqctO3ipu8qqyke;N%XYX{51gr@M}{;@z7H z$SHYZ^~{91SJq_XDixHt*zGF7T|VO;36e*HhkaZs79~k)6#_8oX`Je_n&BTmH~?p+ zuUl~A{9)bMAbCYa#c*3TS+&6 zt9ESp*Tl()E0m0kY;9{gjh3kOmiS0Ya1{X{0FWbX++0#p4~Sa^vlo7K9-Mr}q>C(6 ze-|^r*VIbw`1<|o+JabQ`4_TeScC*hEBpcZ5lO&u9}m~^J?_@F?FBZ%S^lgzU zfq(D0;9KO)JZw5>-QZ8q{dB9`Vv*!*f93a`a#0bH5Bv=$ulDx#S4OED>JO5GGFE=o z@JA$5aju)S2IF%(;HpZTVNz&;<@8RPozOwP%nmsln$p-4g zOR1`Un`}&G57biXFU}s&L>S*F_i8y4$2+6QV0bQmVh$P@NaW}*4Lxr}`Cc2GB~obY z{SkUz-K+mlNhXo_RW^r6 zzu9&7^ps?285q#v@{x|%+uO4_Odd%{f75444gwbQ@{rY@>l~7jw+1eVj=4Zn3~-*F zY!gwi^f%(*<%1BHjX?9Fw{PFNEOt=~2@A973G0Q|Gi#lloz*sSG79ouTbmkBYfnK7 z2h6~mS_9#j2O-W1ECAm;XPn#9&`|fBI6XZre|9Rgw|Bq_WVT+fPBDa`5Z1d!-EUU>Kcn} z`*`Qwo`ncI8yh`4c9w#O-^xx-D+7;J*i%i*B{2NL>x375e+&wCHsiUy9bFMt%U;C# z`U%_-iqqLdwni(_y2*cL-y;`aXsW*+CA};54`%3U|N9s<+P2-+zOKTputy0N(7w)=RnVY%q?jT1c_5JkDorhKR;hpRfR>C zODjG<$!#BOeg#@36o@F8KG^Ez>P*Y}U}+^9RamN;)P`cU(P^c7*Dy!ss+8AgLIjc`w^q->Q5EobRqT~|0_ZMAkN7D6~0_V%MmLp zYZ9q4`?s0=_*|PY&XxpGPa%&7ki5H0QlI@5gZu+|AUZ!A%=~02L@JE|H~~5eZhamL zwi5@%H_5X0x3EnV&2GU16a342r+j~`svt;LZ%eVJB-F!+pQU)NE==n7sV{g+wA2>s ziVyF`JbPTE5TH=Ku%bJ5dmSWu=er-S)E40#-YxkR+ArC>&z;!6JXor@nCnjx@oune zf=Inwo&<>9Gt^>0m>HLnBsCY^F2Y3n`g?E=U?IX@+QLj60^|F!`A$P6(KmITKjPzJ zWqa`4q2Mcsz0>vgyTT;pt9ftIG~RKlFI7fKtvRNDp+Sp7&P?uSwLOm%5*;v;2a-hV z+S+HM;i0z2?>*E4(NQn)yBwW ztsVItX}q8&glB{sc?#W4<}JQ4!pug&rx(|I=Tn6|Nj3o)fGtH0A|%VgVDNuluh3&M zdo7g1al_x@n%l>kI3+tE1V*#r;n^)7KF-wCUlBh;Cb+uhmxYoaqLlU3pYmSpG|27z zK6G^=q12W38n!vZlG6^wg7(vH$$Y5|?`w}1+bRe6$B2dz6GvTLLRPHAxCODJ0zv*= zJ=*i_AO4XB_@7Y%ZrU@6PA@Ex{T|ZnY;DZK`P~2)d%~8jC=LNa_{GvQkQhl3`cpyv z)aB3|vG={CVJAvzFU-Z&MY5dCu1dLGS+AWuTxq{+9x4HtF00YCgYhBW8_;(|){j9i zr%5+Z%W1s1X+^m={-MwK(K8_2RY|F-kxB!3CufJ7x!p*b!MWCm{H)Tzbfw#HW!$4s zf5qEyMur8*tRR?x+)SY1Ms>4xk}2frzse{$iX zrl!SWS7SPgeo+60;`79`asUM@;ystq-Oe2hnHJx%4MboJH=>N)dP?6t(a~7!r9QO~ zcbr-<5sF*~G)lJ?IA&z>`r6q%dSA*f|p$JO4S{-8PqVZK>zEL&Ez+V!?7)o9j>|67 zR*5fmc4@8c_iSZHm0|!fgqHEE`)$X1b1BcY8|e>RaR@k|s7IK7nS=PaxF!oC@+9tM z$gB&xwtzG*(He&#sAvCZrnuXlP@n@>r{N0b`K$;q*Wq#zufE{a*^jGr8a z4^u{(2qgL^i#)$3zHLK@|7&tP_`=x=1e zlduhN?JRze1B8Hm!zyPY*7P;G1JgtO*_NIW5SA{D{bY4wFfU`H z>-wnR*J{NR!{UdhtQZPKY76)7JV9@ANRV=(;zF@GFZu+{Oihg@{y-d%!|tC=x#I;Q zjl{HzXJnqw%;3-{frO#=L=?=(Q|{Aq=JwxIR3uQXldH`iaGkCJx5=I5zI58%;qo_k zqdD;ZS?7!a2+##=9Ju5M#UZZe+ma6lHM6Rtn%77!o!O!oGx?Xin-;6%LK7}43x;`8 zX%{%HtgLmq>R_OZd2ZZdVy9(S((3?el%e!?(}<j@L>)-iC-XEw97u!q8s$sxyW0sJ7He7OCau{T<7Q2P+Hy)AF@Xpj` zt3=T3_8?3?rb)REu6-%&92rRrA!qLLyx-@P{&^&9WJH zWO)!Ok{2$^XNKXQee0I)C9oVZc<2oMitd4jG$j9~mNEuXfVn_wkE@c_9Rsg8@TX{T z!txv@t3Q{8qYZ1LcQ5sb^8(Ltg-}kn0o6uDesV59#xrF!CtX*Mf5F=w8wo9Qn8!u%ILst?tyZLl6WEg{rklvZb2WR3qHwGsXu3%h%VgA z&iV8p2(3^lNJvO%vc+g_3-KwekLBAdgdf0vqRfkPYDvP75M&trMvQ10dnE} zrY*eo285NjoV1<9+zagw?ZQm&O7!Ao`d6NkZcui#2aPp(%-%)(;o=YrY5%xJJFgb4 zzc6yU*eOiiukx+(?Q8g`!(f-V?~BXZ?WZWn3lSj^i-C|=TNNLlRm~7G$|gvub->$l zs+q@#$_^o}%F!@<2B+~x%7!3-L+!{C&$z~nA|LL{PI^nklmaoE^myPg@xzv+ScP81 zRJuoBqffJS7-qAs4=gLq-wgslQAha6lebWgF&|c6$Y(-%(}N8~^<~PA?*MAmH)fnWAhppOL79m>#Lj)jx;I1=s_^X6MJfCXp~lz=0tK{ zPk8(G-nw9FYHEl4wEtH21RKz|2?>k7Kf!o(_4fL%KYyE&5)UBostWgPnw$4`svb7@ zro|W-f$~zG)z!6uk&Z%rUERGO^LZcfB76E9V~b52RIwr(v7v#1%wp5(S0LJO(hr?} zaCoqKXNgK(Q`6hpGA{ZrPW7J!Zv1r(d{pjABLN@5-=xi2XW@xFD}C?%>Tl-KA8*f1 z2E&!%=22gv|k8K4q8H`Q8FABB?V zXyZsY6^V~#lvH%zp8WXb58$pIBfdWE;qL0}^elWnT<%QzRN=&LZ6Tlc)elGGKZB6DjDO%NG&a`WLei+uB(H= zA|j$lAg(d|cwdD*n|)zvNi~{0`tc(hd2k<%<5W8>l1$b+r+`tr100nN3=MU`xW3ZV zLIk^hy)!yG{yh~o%?^aK8snP)Z4G!jwBcqrrq^SqYv3lLU0q$>b%S076CXA@2573J z-+_|IAr`62nRZuK*Qyw>m+c)K+yDl``gfov9mYhZrK+lm-Fr(bDmWxGq1Vxg4pX(A z%b8C>9uUCnc)PxTe?MYpsOS4o*7F63&A1IAt;YKG#%6auHBO^csxRi*a$3d^liJ*ihRB4$%)w*c}R0XTorK>MmE&7=U`hecD&gO~0 zcXEEc^J`k%hHiM(1x&HoF)L$g10@31)%VCgB9EX7`k4EVr6*9K~4umeC2>tBw3Ma1^w1DV2wieg77VP`{ z_`@KmkV^m_J5wa@0$S}+x=z>9P;|KpBOoH51DOJ7`9l_qO*xe(?@+iPQv zBioxboXEX1$AtJ7f3khCCL)|m9G8s~!TMFzlZm%w)+@J9C}~jWC~Qb?rUg}-)}zdW z-#OE^9Ww}Cf!F+Iy{C7kefMut)6xzd+KqBa+(-)U@6_WYo()h?P~de5505~8QiEd! zU0W0NO0u%D8~*YhGdWWr5kkyi>Nc==TP)iF!`;I2x|;p3^6VYC7Gg^k2iOFggc&*> ztqBRi2C#VCukvCief3bG&KM&QT49CMGxprx|B+I34llN()pf}Mz zfDn;q-eTq*b6}s!0!6Q{fPl&VSAA4!T55ixJ&!ujhEiz9(7$ce3i>-0S2O#h1;fvj z-X&I4h}r+KluBn|Y?#S-MB@z2X$c{>BG9CTyLN7Jw1FwYjN=Lp-E}ke;#mu?}WTya0?W?U+09a{E3=lmMCPe#XR`E>iw;O zOKQ2nQ?^ne9Z1>^SBFCPU%@lHtIs|qHTN8XVgM{42)gSjBCbj(p+%22r`sJMPVuk# z9&{m|nDtzQD{X@5?)k~sYkLE73X{S<4xcvhW5|UbF%pxdfRvYqACqL|19`sKvx!FW z%Mudgf;_?D)j4z}(1JZ`TcN#ZZ*Bj~hWGi?i#~^5ioo0c4=~R8Q|KZwKKnf9)jE|49CPe5(pYl zQAwzBaB!5ntm;lVtR>`&tPSb#Hbb)@pTqhU!$yn<^nrq2%JqKo{S6O<31u^-BE|2y z`-r~6+N{)j^Yo3`LgK>KpRs1XFWw)!GXArZ;DBufK>APEGthc+z~)=O8XAFF+S{Kf9%7?ceSzVmr8#;$0?{H>=R4;&sHp!l%n=_OyuH%cp|rKNdaOznPFyYi57 zU_L9Xz^8Mo{R1ucJ7cvwI>y4%m6$2<8TWt_1_cEEAF!<%P*gm}U5inYOqaX@BAj4c zy1d7u4UT?pEGSI{;$mdzBsti%6vn zmgfIwj&pw9+|pv(j3H)J4BL@_AU_3s!Z3tDRbl004SsUeZ(#c;f?GiK=44Kc>1JnV z|GbVlzp>yj@(q919r~_AOVOAKU!Nz%uzWMdKGl*C2O|q_eo8W{6m1No<2FMa>= zct9%mf;B>$ELc>49Fmu2E+x4D^$QIS{tQt$RQNx(J*>r{;9;z9ttJmx*#S2vyYpZBOzpAVlD$DU|^w zv4qgY*^wTJd4ki(EmIj&8r!@bq5$1=|G+UiP-o-mr2-Losi`#%PJLR2h9zgWGMTxji-Gs=Pq!W0KWWM{fsYt=EP%ug~whCe$ zDk%dG*aIm3QAVPI2Y8K5P@OMTCx>EHVR`*1yfD4+}zxAwX9+6j8qM5F~qDO z1tUF80F#DD@Ay6x(b5mE?I&tJ_{8#!Seb(R5^+Q4P#kmrkdP86$D9>b7=O4ZRx*b3 z$)hr!Nu<(CJccsoK_TrVi5szw!Si$O($TSU7~DBQftLBanNY$CNbt95^mk~MF%^}Z zpPX#Ncgi9JFdDy>oo>wle= z{#kP`D)7qGsG2J0|9<8FrT5qv;uaVb8e;d+s>;fK;2F&?Eqw(NK#|rm>hI6kx}3}hk0~J3 zJ|DcKS)u{Lkd{$)DXA0}S641@pUL>1pL~g7llmVw50F|YWZWap<1?-9@9wjH^{$v1 z7!-g(j(q&rFTE#EI4(Z#Q-nuF`LMNtf3OLjBz;zVi?!xf(hvFhcR;Kg@Mvh&9B*(a zDJeNQyS|tCeT1^Gu&8Sfq)QRAqPLRn%$OECH{3oNyRu9}Pp^LIdzrA$8w${*j&EZa zq4Xz%acpWUTXi*}M?s%`PFvpq%bHq4@yQp3cCX2(`zX@E9|Jj#4VSNjf4<7rc({f; z=Z5WMXWJr_+uB&r1TtKf|#FXY{US1i}+`8|DO``F+l74!L8o zUJJ|mw;%Zb`y~ii(#|l%S+qgjNz>};yuDs^U-$OS{)|dSvMwIST?=n?#zSgUos!oO z=>J$*=!|FN7JJlqg`1l|KSKjbC4523c|&B-9{TAR9RL-8*Lc@#^5;Av1PB zEFMaA|DS0}>hW@Ky5;kr2824i7w(Te$KSty-}~ybpndl$4CUE6dAjr@^N#YYSZTN1Xz)EYC=BM)EB=X;3y$lppg& zd}L%~o?pYIH?s};^WAjW9*g0zjwGfsgi=WVf$fm(d`=W^^m8V3{dd26e7Nd2r!$3+ z|LsbMXybk%$-At`7Af>0mZaCuTYUTHZH|`RD_5@Uj71HAlX;`x!RcZx=K&9NDh60W zb|pNe($5u{8gKYNdz_r__j%TY@lf)r!`}g%y3-=V2*rO3`0`AvoPxnz_T53(MN{Nv zYC|mOJJiNb^^}8MIn=9btAoJNG+uXI(&9!$I5rp&l%aJ1UlKWsY&YPWl6xMAIVT~G zT4mqMGPUC3MU#I6GMnYqh5UF%AvzdL8MuM0k2Oc2ufythf(VJQNkwTN9(;AsF-kA~ zLV~OmLx1%>Usdx;K1{%xq~{167$%q{+;ehrM4mS*%Q~hiHTYiSRaPeLIqi=T{4DpR z3ME!F9{*BdNl7#Sp?x-9YPA9NxfmJcjc#Q6?J!mupsTBE)4d{ieiHqOX}I3>g&Q8? z0A6|ZfNV@h$rr{bws9;m|J)#HI(?S+90WZmf@4hvuqX{?Cxv*K9LKS zZ3v{jAeQ{Z$OAK7-R9!&0Aw*9K5;qB1o;Y@DH!DItx3~sG%RsFe+qdr3obx!0_>uTfsC|d=1lP9WQJ78FxUpP61x= z+_nW_4-G?y%}g;$yTo#XfofHrnuhw{3?fsegY4uA&;;;3!HEl-qX46;x5=lLq}gW4 z8VR5gp}BcB%1UmBX~c9$`7DVvoeB#IW`dheSLZq9&7=nZKFW(GtKY7RB?*a%2JiiW zHr^w5N;?~C`5F4=Ie%{H_^nj<^`S__{fcLu1W2giOKHAHQU~VLoB!*?3~XH@QaA=O zn37w{m>Gm}Z{KK)EKcfuL9+{mS*}uX3g_^}XU-kj<8P!VtoP z$*2{WMejLT)3>@!uu?X6FzQ z!|-b`1BYa~q+2)$REbjg4Vtk@1)1NHo$qRD3SzDSDa3ndPCS}M$)bYJ>*wPRLE=(P zBHQ<`KpDpokLJ~@5%P+kn-)f`M0`s)!}U9S0QsL;OpH1x?lOwi3%~Ya3U!Ln3G7-0 z%rYJ=h$PGB!U|i6XTfy|5f+`rA3l>=sEb(uyHvVurdy+iUTT;M-mLKaW+)h(Nr6VF zI6Df{%-ZiP=Sp0-i;~?t-Mptmf{ic><_W~gx(`XyibzCoWc>MARt!rPZ4nv~F&Iin zsV`=JJ#-zaLcjxm*uiGomCVaDQWMbP;~9jggDtgOf`IRTSh8l4_83Mb((j~941}yF z6Qa_1sr(=8_FnKTK%OM$iZo~ZI#TxZ&6_t;@_PuG0$94tBCEfS41j}Wj|o9TNP}^L zR3NEbd&k6eG_#(N+G;lh%}M-sngzr|T@~&|!^8DyC#wZH?TvT%*~n9JuSBrey-!L^ zb<4DK%PbEc1mf_6fGi!|t@;Gt5Nwi;-KIw>Wm0J8g7UW%Ou!ffFk*!;K{Qf|r|F&3UC9srr@*J(ru&W7fXy>b$|sB&j}S~tkfn+Gp8jWd#)fJCQFKBQl6 zrvc!eO4YDmXk$e7Njx?*E?cJWiRj{iN-ffm5LLV^yak7*V7#=KRW|S0Xr!I1Jy$

Ib!>r^vDr{!6<49_mrrm@{h0{sOl}E_TMqot-P% zu|?w+)Yqyqt#bwP>mN0?$kh?71v{hO4HS`hi|4CE_%TOPTJaj0FwwpvQHim*q|+#$ zg-1Xix|TJ4T5b+~pV7Z7;*Rwm2N-uIY>&{l)d8 z6Q>x7nOC99wM#S#7$%zJZD?Dtg}J#zF-dcI$aQ?rc;_bw1w_yyel~u4AbTiNc}@-o zP0R-tHZDN#7Tjf5-Wbc}qGm6(wZ2`KDGaC(a?N2YyDZ8gJUrAt>lS>X5SXx${LQW1 z89yFj;Hai4=6Ign`lQ19t&~59IP0#~55D|=v;eLP1oR3f27Z1QA~eidLQk;uAk0z` z4&cs~oZL+wO7=13t_(U3pV_Ek_u4V%zr`JV;BOXW5E)CK(TDUm_V;v0CxCn8FJ}9% z%?UnXp#f#5x^u&%N)`c5^E5gh%U(L`L-5}U_Wypd3_mto=A%g;g9q20@nk}qSSN^B zL;m=s1tRv=K{36*Ui$E!{=e@y7%F{N0fjr{i!BuK4KMWd396N{9iM| z@0DPj5}w=|VLj8eKOc=O4n(75YmYi3k25d;?Py@cu-xme`hPx}%msJ}GWGUc?EjE= z|9J`*VY%FS@}}GD&qw?J_=5V2adm(d$+y0tfw4~k>8)*L#W^=Si`LPhf>v>DTiYdj z)2#h{hl$3E2G<;XFDlZO<^a%_Qln5xgx^J#ia%Gw2qFB0*o-vnw#auPg+#H zeE#g!?kWbKt(ADSz6QK@iI2$-FlRs3($eOUl6nMqfXTftpFgM7*VlJIHUTC!5!kXN z@D#fX&6%Vxj6%>j#!66>pjVKSGh1ySW{Ll|jJ4}b*V5)(w-96_-;xFV+0Ph8{408T zsfkY9gAm!rI8Y5iD90eiacK<1h|zAIy_C_(#wsWi>t5y2F+TnTM2|~x>37f$4lDLJ zWo3N;Hzr2B!grjgNC$Oa)33*m9t}*;{!FiQt6X~1gj40*B!^+_6psuQXa!9s9bq+U zX(`nWOm&D{?(}RaDHOyO^SJbauJ1*$S>~zMs^lHVqftvpN4(U)odB(kg+m z7&o=R@$n0Ozv7aTC773B8P@fTyU=idQ3VUY%E7+oLp(=l8OxQi*e zH~jQs{u_I#0@~aANAzeSSiCWz!2c4RUeyIwOWN_g(-%uD&!Z-P6?Ju3q@@oJfOr#g zeuJ_bEoIJl92SO0+% zJSEVd;(8I5n%Z9B^&nfk`dB%~e^dM3Om&>5AzCd=@`sMW;Y9zr8pX1^>Eo;ow$Qog zzC6TBL-p%m=UHW?sLbiliEzQW@qFBKnJFmGkJr-H`jK}L2(IrdU*z9%y{-Q44bmH$ z^E-t3P+XN%`%p3^CAH#4P!ZM8*+q~LHJ>7XvM@RM#++>Wiug5ZE*qYM&iD8s9%oR+Kg-+lT(sqa<}--{5$y_V4`Pc&#`t zaS-t8YKMJl@(zB5YlbOu`EkzK4aqf`{iZg^CNN3c$f&2;u0K=OeZGetc09kG8+I|HM?1pU!6E&A@6`#zUas>D)^xH)%iKYe#x>Cz#a3h4;WeC}w~&FXk$ za_9uPKJQKY+4;$DQ3mu+qH=R{H7#{Gzz!8%x9?t}jl$a{%7Y=O@&mo7B>}6@^5L%o z{`JYvO<|8gY)RYz`<6DWKM%-5X=1$(e|k%SYAh=Q0&r( z^A4X;027P-qqel0D7^?g_VfOb$-TQ?_eZvOb4^()vDF zL1q|zLQ|_NZ55RcFUX||8naPSQtrc5hPCWfhsi*Q6Nem zw5Gf?ZBo;UnA}jmFSmyEZ^Jvj!7~efx_d zS~2|O0@_B}(7ui-_K!^!Rpql7U1Q+Wa3j8Z_|41LrUeE@MsB`;ZgoB7E8eZih;Il| zf}0>;%gvr)s5~%2%ub;}^xfFcnul7s$lN4eO zy2+hc7_qMV+K?dnF<{@X1o*JTp{&a@G+w?1o6m?Gt&;z^qC$jLuA{d*7t%rNshW3} z&)gV#_Ei4_a)Ah&ki>lUmG^|^<(ntNA|kd95kv$Dj1bI?$wa_88RW9kDk8{+g}=Q0 zaL`EE{PPDF{(Dfmpe9lwBE(F=>YhKInw+Kpag3JMJ-x#*U&H`}w-3RlLohPd?MA0tk`4z3IXmX(c7 z2OCb7`5C=3Kl0{?fq8-{`5-(3I?^d^B5qVa(g|NK_nc;UG_H%N!v|qf7ilQ2jb;nY ziD42+6f~*pP~-!d4;!&LPBTueGJ*~ABKB3OWeCE-1+NB1>20WhF-=x#-}vcQw12sR zaF$bcHfsR6nSI*&!hHXQ$SMrv=S0b+HtFxo1TUx+ljpHj6n^S-a~mVXJoibe6MO-| zA5Q4^skQQQVVX3=b?20Ly|B+vx8S1*TO94e386^TtXh^tYrWIF3w-ZP(9kD<$D-u= z>sV|YoUf-~=d44%COjAt7p(FApMwU&d@ns_Oo%GMv;DdxHmq;*RRYhTHhD`psY|tx zNG`L%P$AnVf~O)TjX5+_R&46Bsh&n^+X*8kDpi!NlLTLh!*$L1BsD=cGaF0&XMw?P zA##KS5g~5cT1L&(-^yQ*ZDf}(LV|x9llo%L_Z$2K4J zlDgiR>=_#7CJmmEt7jD=n8yF1NFbnEn<@v4S!i|kWHQX~P2q(e`41w{uO871D2m`a zgxGnLr?Q(6@T5cr2Lb^ow}AW!;5LY2g0aH{CXuW{guw(0cg5mYTr7EwwdiR1x3XJDv52N<0nK02n#wDvkv-V$y+ zj<+mrTC+)pPihL|C`zFd8}hye`!k4^dL7v%vHyYF^Zi~t&Nt+PVAfm-Uy%c;@IMkZ zjOiDNt&$ZSFI96!pM_9&{r~t91=R2+1B#S<-|08(&&rY*WgUb=R zCNSAqdTi5%PyCEvEATfOz5echR*9ELnAWMf>@Cb7p6*j4k5jBiB4$8{o;G*P+kKDL zVj}J}KUFrbkB5tk?xqp&e*XpX0Y^`tRw;f8qx|5f?>l4osm#Zw;?Orrmx06JDZVbY zwe_%mq#x${u7O$s3~!F9tsW}Su7YI1)DdmFLi-qzhYugUFBC7j-J7>e!Z}poIh3%@ zULQXA^#$f;@~Rllfbkb)qZeN|2pzu?WzPK`PU*$NX#+5ASf;AL-x!wPz!ExrFqoY| z5S#q+<#|vgX}Y+G2;KJmIwHU)B*YCqE)2j#Oib*vzK$i7cd+=L-PP4S8zU)2v|v~2 z`_T}0KzzUEr*n>-9UYOhd2PW}^z`(W*4FKD5c0aI&voP|ijh*$K2eS#%dp)O#oO;Y1&qf6k|i4D+e55ePVHZ@z<|nQyB#Gu@^> zw3Z+~2l|y6B4T1*psEfk%d0Xy9X}1t8m`;Xe)zMD5Z-&+c&rg%sdd92qtFmJBK;EGBRZP;5w^~$Gl;2Nq?+m;%mac8xU)KnK&Lf{e8cXUv z)r?1DJkfiw8Wu=INXiN-JDEN#oRjexWh@taeYy^^6b`^@as@nhc*g4OSFT$nTDnw9TPxI@>Cg5&kO^_i0Y>gA zi0%eJP599alhg(@-t#xXNjR(?n?qNKR;krMZRGxQ&xM8yP^ujNgWZbj)y`9Lq%1%0 zT>5p%)!jJ<+OhjV&-4MWIlz9jBsVgWfB{CjT|)Acvd26zk7-EiECKNCV88)Y-Wd z7RVrF$@9tt!zyAJI!J1Kbq!%nFcE?XX$h61r|V9_j_n0P=#;{Rj*u~Q8;OY`{Kd#77|C^}p8 z-;23dJE_2!dM=H&3P8(uR&u2juJ<7MPou=JIvbo$1K*@yO16?=gJv25ASnXf@pS(n z__g-jmWpQx4E(n1nN9qOKyBT=%9+`p(G|4JJn%TIH$&-&Lr6$y`EOi$DgxqIoh0uMI|&U3bMD8`5|hWM=QQX{ zEG3&muPw=9A%6aQUub9XFfua!9s!9sj7RcbOpIjc8i2`0x0obV5qcd4;8!emb)(wN z(U(01m%y(@-jOkqr6}osaU49QaS#CUCqCjH_-c${t6M5aPaAVbK?S zM;u7+9DI8@Q9}n7>o2u*I11FCF-)B1{){_~9yn8u1lY24N@v(8{105KFY0^%NOYG_ zQn$m@E*~ds{X!_wyyZgqQ)(#U_E$?6tv105W211+L>#83q_mlZ5Og2-53c$7K{Jct z(I=l}^8TlVk*!^(lIOc$J{G#z>8HhsOH8ylnH7?gTcQYT0TmYKTEKDOYW?JItx^r9 z2$WU>l_IUXwYfb#Fb#zjMgnE94a_btcSVn}wG)`1G;{56lB;#qz2ch8py8YR)>{R^ zzQ|(LqZ!M@2Cm4I?wqGJI4Vi(8OybsEOL07C4GPp#o?{e>|a)9-48z-Do0H~9r|Dp z(iRo@(U+O2f1&5d+Jul}>s|o?GI~QHNue|HKUtN|0C^wBXdzF)+DaK4r0A_l#2m(l zcogY&x#M)Ih@Kp!?yp9lG$5iIn^cuO|1x7^-ps$>*(|?RX);f`xvcq)zt=7!@PPeJFEUSL}%Kzu9GLj0tTm9z>=^<`@P)2RjjJB4&1e` zUL{3U5Q0@UeAFL|Cdba<&_Evy24QP^f%{_bLKlPLvPi$4lBVJN;JaOhXe)952#*Iy zbouxoC?^UkvlCCAqC+{3Rb@R@irZvR2PrN_AyJkKlMnZyG#mw7S08lPa~sFhciD@Z z#c*tg>|}$;`YJ{t4d&(L&$VYUIH{Z;HjG{I!09{Tzn`C)YH9(W^NgZf1q)3>+nB2U z9?ezSk&xg(OsEwAYggpuLtNe5yemxkZ@;o!q5vJaA8V%J3!Bcbl4h)vfT7;iqSsBL0$+5?JS`^fAsL14i@zu3Qi0GS&EVnrUApai%5S+kLR z$deAzQYGF3-kh>B4Tud${=tzH92|VRV(gpr3vCq@q8R#%;b~0i`aSnxD-0u|X$t$$ zBv*srfa4_RSyB7Dp?Fq#oBgUKOqCSTC+f=>;f7+EBEPv0-cB}vUyY(;Qe(gG(1vHN zdvYLKa-+GHfIT#@xds}DRal!?-Jz1sL}54Dz!Z^I4jC3nFy|5HW8}8VEo^npeZbG_ zBNY`C(2(O5-~bie+r6U`=o@_~!%9oSh!fNGND8l&&+l)SQ7s))U|~jTyKC58;J}9B zO%3DvS7-xo`d7jLGtbp?)u=ta98(BE+M(~>z75AP-85W-c~>o(uv6PX@7sKI$~s6K z?(1U}7WnBGz3!hRVNig69qWZIfEurr8K{FF0VCIcXN*zl5(^Sk<{p`-)cMq5^ns!; zg}K#9)Kx(D0-8rHPtS9hH110;K(i<2(WO%<{4x{5Trs$uXBtLV4&0m{RK+L&w$09N z!A`2)DuI04uV-k;0HEc@4EQN#IR*c?gi-7iAd@08Li&6*`hq#;S58BXXFn{RHb+gr zX6Kf4pW*20>nKAl{X83Kh&;ab`HWlJ!rxd~UlPd`T6Nzy%EZP)#tT{9x_K3lSd5fE z8DO(>KO2LDdNBm{Gs0~9{x`DQ99x9a`p(&tV!w{Qr7!5|deMZx4Bjq8dZH?5b&B=b zKL}T3z75uFRZgTAJFC2@k?&HOZSeMy%?chrfq7Fnv!d)H6>%CJDr)qjmRDzY4Y51Z zBkehhch|#+7oKcM_tmnjJ|Az`l`UZD+JKAk882nH@esx$VEgEDx{5@uC?5HBszzm1 zoHEnbrJP}l3C(W^lWA^RS}GfOdrTOLhwOYEExPT(jzxj|*GIlI_Kr#>^89zt?_=%T zD<98N6DHT0mA6JEaf>lS$UywJPcQ91G7T`Py^g+jZ-_#MV5stlfAClQPgExOoQo)M zc;z)T6}`i_E*VRCB~)#TMa7j}jd~k9%+ZBJ%bpEs9URo2o}RXZ#6T@H8VMY#b`a4o zF&#d>GdWdNIu6B-RZNK7fNH{IFAb+F%F5^<(evl^g6z=0+!ohaHyEd$QG|UhDJ~wu zCE+!?d^5b*ut5d*j^)*)X7+>=gz^9S7%HP^>*1QYdb*Az%;TGrMm~M|^d8G46UE$`8vXi)29H+BP@FsMV$!E32~k+#K5R4BJ`_g6JSh z9No-pGKH_@_&L`X3qZBB?gVj&y_cfbP^SJeR9z=nef7_e17TA*9d1!<{PG&bRc>X4 z+o3fzbseC^b|*rfKXxG@gX%cZWw=B(>qui|!u<^NX888WTrWDVpeR4OGqL28biA>m zuZvgsnY$*3->Jfox^oc@(lLwBU62Io$A!7dY)bQ0$hBp5Tk(3| z3&Zab;;aqLt(Ue35D-np8&OhXoT8Jh!j?4yhT zow;ZG@3+kFzf%&TGu3e0z9-*Ow`Pb$7KsgZGij8OWe;%BQC*b7q|BkEt zMgg0Ohf>z%j~_1A1Li3&L+%mQI z`eaa){Fd%<1J%(o9JFL>ZSsyuNlBdS?D8hb4SOvizCJ!GTQ_@mSYO{ptY>`e z>FsIj?pE~y%>~)IfS}<0sw%PC`#$O{NTW|9rN+{>^CN1d6E%W(czAc_hqxN>M?YkH z6F0nt<+^_3v~x>K?1r$#g`t|_dNq)y-57AGr95_>*K1%-oSvQ@OOQ~CPKS`2-@L^o z^yD>ySe3Zq`STBFmyL$Ti+(D{o0J*QPFTIRk4Ny&&NttDGu7 z=Po%eQe0eDt2Fl~S8ovAJz~0NN6{DQ&cv}?qPZtW7>T3kSU+3cLt|h6eu9^b6QU<3(knZVnc`2>l z2Qy1lOi7FyX&OIG48J9)9H)oJRT*@SIMS~6@1A0kK$}WSNGJoFbFTn_i|`{_rB#Mb z(cE>jEF<+sK7bXH!9Cas5CSIajZYqcmBiFkZZO)>a2SA-!F6eX^NNNBEv;Va`4QsA znybp@Cbyo9lhw0~Qj4Bf7%a;;_cJHk3c<5`Zgl|4m|fi5lE7b5bjrLpz9cP;DP=}B z1fRaF`(vlUXps2SY1dwn)2m41nfvB9&2~x`ZVRp5IV?s$k?Bb7+JxfV z$sE+u%%D>4l=-ycjQ56O16URfRU(8ASp4@AR65~_R`slbQ>s$PCu95Dpq}D_Nz=t& z*lSBu1=!E6KG6t7FkZB)5_9NBQ*zid+N&o|$F-?iS>-+mZb{xlhoV1nqXJV1@+`WM|J`c?Et&-2%tB(1%~PiZ62sd25~zM z7=m>YWF^zg9!psj?j0O-Lnrerkl#Lnkr7Wro)WeF{IRuTNF!uiyri@Dx@AY)^49WD zK^NcAwMI=*jVDqCo#QRd5tS}8mknxeWOa4DoCqk&zPt6M##=tOz;q&@j>VGT>%G|6 zYX)iOJRZ3;Opbo2t>7@Znw(dvz7xPC>7zh@vX5}Y_xZ*`%;g{$=OWbTF3GOpSiTz` z)&;6*dwl2Hr%@ie=w{jiwBWCs1w@P-o2V766y;mTPR#XE@SO9cp8JxkzDsUyH8E@4 zw_!4VggkGssXd*$y;Rhg*C%%=>At}At}d2qT`*o$({_1fmZ?yjQ@(7an5*_cST->+ z(Gs?QQD_~ePJX(f)CNO5((FBcwiP4*iCT=EQZMyf`OF#+5U`)?590GN4Zv~=T|?aH zeTw4QWHXUBoKn+RHSb17#us++QImA2tEt^%l5$suR&W7jy=x^Vxvt`O#{D+dx6zui z(^`Yxtm-VW2Wa1vfS(&mygpjnCl_8kakN_bP%%BV$_lLKqrBYQX9YGW6dR4=&r(#@ zkTYES10~nbr?A)eb$+};SqKAV*8>rYs`^_HGt|XUK<#rMdAH^e%#!q1ulO z)zTz10np=wDaDblNGt@*sd8#w014)iu2I7ID6Z9j|V)oy;pI4_*EPcr9eY|~o z(nB2^Pd1pxalFpJo1Hn0T1*w9)7;Vc?~Za7qw*Z$MRUv{W@;?tWT!wG&yi|CY%R~f zMw_75v_U?cXhd9=d_)(3tW^%Z<+Wli31zPj4U^dXew>CrJ(p(Yt_IIRU>OqH<@wYSK7>1G!Q*(A1Fih9xw+Y0zh!f6f`4ku zZXvhfqH;$3N!b8M`=}nZ0SVri>dqIhVeFks71gbNZCvt=ot=`6dgYYb$Xb zGXM0&YCuxIbQGWy1l4LotbdRu6s8=5$WHYiN$eNIn0hCzeS0( zh%WPvO>dnz9~a_;GFfaC?%VRI4_FpB?hbHgE51K#b_08sH1{^$WFYxC;^7y~SK`AD zkCyzHO)7kP?Y=SZPj3mitMl~A59khhi#UUa8$n;Z<`_hj&Qj=aDsl2{jRoHn$XZ)p zMZ8p4jT&M_DCajCe2?j(#@+In_m~rgycYF)o~z@$U+O%Qpv%BEgSv}p;bh;iXgg+E z5YDnBr3g#BSWo(hJ|~L8#G*sN*}2^ErseV^A%K4~K22l|S~d8}waDtlpt>}7gJbSOxdedaIDV08ZMNak5pRLY8L{Se&9Y0giw3ss$Uysbi zWtT?vdPJRsMmo#WMUxkw`)b+d`Z5^eVmJspi3ZsGqAnLaB)y&g1TmgvfiHRIu>Srl zRs%WP(_PI-a-LmS!Q;53#0jbPv(U>CQO-O{Ma6|fLcf$PVM5xjHgogz9QhR=33V!k zMp?-{^$E44?kFd6+-@YA$9HS#+f&Y9Wm6&+YP(M6abm)X7cWXT>h+~La=3z;4SjZ2 za%~e}+rJ>_!+(>cS0@=-dQ+O9s8L5*xREFY zzQ620e|!S>kcIQn&gK^DJk7{3oB1``g@P>$oomTg6b-KI(5)AqkBxB^OLmeCM@k~F za`+Yn7A{unP4#Q42Q zDq(xlXLD=IDt--_a-oB2gy_A3U3m4Bp^)+61l`v1>(`CrFIjS;$N9ExvvSA=fAmv^ z;~%l*ie6Bro#|q=j^~bK840tOi(RU(f|Kp2ip^x%>DA{gmjU_3HWnG@v1_p+>PM>2Rx3O#{d8T delta 119374 zcmaI8bzD?i`v>X_poGMb0@5iR(k&9w-JR0iF`#rRgMxGkNJvXbND2r@gM@^XbR!_% zHG0nb-gDl2fBfNRX7=7|ul1}azu#v$(T3BJi2aNNM;86&jT<+#SFBKOKtgh#Gbj>q zv2*ciaSJiGVneGU-V6_q#RXv^n9H? zl-ZrQ4b;_59sbb5HXa@hE)FgP4+KFdBp~?b1}ncm*ZA&p)qH`hKoHsZ zL@6D&A@m3}cGVPjRmdHLU(5j=ub>bIpMcPR|NcDv*TX#bc{%tH2PS4T|Mz{|LV^f9 z*dv7GEeGU#AiRhwAvc6L0S4msZQK+yEyw|)D;Wmm6-3~2TH$g3wa&j^lH~p<<+Cj$ z6cKeV4rVSS#KXY*z}nf<-A@lJfjgzx9wLV@BN#>qu)z?7tX4$4LPGzXC9ePn4+7Ul z6VXnBjtEII26qxbId~Bxu%Q%HX$VmbJ3IOv7&revp1Lo9p#I!Yb1mbF`e(i=)cz1$ zgt+`N@qIzC4alANYlipl10e|_xI4{>xc?IGKd*dH`iL->a{tc=H2*UKQp|!}|NTQi z)Xi!5{yp0N+!dNpgHTZ=s&P_CLqnL!w-fO3@cmPG`1P{^M_kUK*@jE3<)YRUUsEjOoYP+<4}Ask?gh@~n9gxOhAjn#Y)E{yjdm-+7V z|8*HrbGowx0ekc3ZbZ$q{c=c3u@wXddY>Bs`Pxu}c2o`dm+0^F3H|5(kdu->_oMJ} z@gaY$F*~z@5OO_sadL5VvGoI%;bCp=Zu!{SBZYW_?M>g@#NE7SR(yH4K8i~1n~+JcmyG&cY*!SE+)CZ1*-l}@$>Ks)r`PQph&es z0bh{z2;=>iD*)?rFGt*$`E$nzY^BDSI1qyWKYApDfH7I%@&41Ac@UgQZ}Iv5cA5X{ zDRLejLF9{TTxsGVMEv~!y2u4~=s)j8MhEpc2oXYZ${nFhkC9S&3*t$_D|nxS`=5_- z3v&H2Fz`fuf+++6q2m?c;rMR=0iXV%m)u}|f+@KakV53vzq5c8K}ooHIsP?O)_Qx+##Bfe>{!6tq$_%j)7Nr^`YJnG-wKEJj5g=5dy(WIrfFbfdGbg_aZF8K=*0>`R!oZ;D?Y3|L{oy`eG@!Xz<<>ujL z5({u$T!dKJ+KPDI-BH3|Mn0Mxg3kW_0cF(kHn>t-45FB z+`ga9;>ZQyID={Y1hA7eLzkGS>bk~;2D9;x>b~aRv5F1rISvmGD+A7)VYYP+Q?f-H zD(A;Fst4UErqjYN#$UX4xsM9>xnnCV}_KEhXntiE^1=hcM56>lRrs*-!XfaHhPw3|Xmx9fO0|`uh4wj!)u$t?k8LH*aj&Fu?~PY|Csjk<1jD(@x#c)JS{p9>$L!Kf*OJ(9xegnCk263*SRATq7m4I^Mfj zzh>Ww-6L1URgI25K0ao<5;69y`)z_ss9KVnEjg%gFzxfFl5Vv9>dVocwGOj56e6Bo zSuFvT9#bl$uh4>L>s>f?%8rjcQE%N^`S_epUP+-iVS%U~swOxry zs#zaLk|i6}qy7D7O}dRjje=em$zsNvjoQy{>?uw@GHeSL8!{E-L=Ai&^tlZVhpDvv zx)!O1SF<+^TQI#lBsHhxyiMOEUv6PU2Wr(GcZV-2y6D)ONiP>cbsm3V3YH) zU0uE&NGaP@R0XP^a|jek+iyyz0_h#d(dSt54C?KIG&|4(=1+H9)=iXxEvmbe2d~T+ z7onn8pF8hxncWb7b%}AhJ%h5FM!foq;Kqrib=vrxn>4q)F_Qmyvg}nxo7oaHBTfV(OLdFTXChX-%>C}^K(Nw-f$r3t z`a}WIIk9KYUxnI#m~b5yJNwvq@3Llm!iVL7@~=B->)}OZ+;9K3UIqRKd`8~_+DD^EGg~DLW}>vW#|Jo zb}{U?9B=oJ&Ir`I1Ng)SE4#>kx!s1t;Z;` z=}0|qe<+ETynuzYoKQpb)Q8Cx6>!cbuPw&OyG?B~tXYDeJ~u~NeEUiFJUJtb3Vx+t zCf&J5jzfI$bN6!3_57o`AJ<$PT!NOCw!oPveDJkmO2Ip=x@Y!}zrJmW%I=KB9q6P* z!$yVGVroWI_s*f~1#qzzu#|{pu<+UWob$R%~3# zPnmvv=Pi>-uUw5de(YA%65Z$Apj)4QxwX#ew)meuIH~+jd*QRQ5Dwn^QC7N0CG+!J zWf9nMt)M!gnCOzZknO^IKfN}%xma5LV;oZ|Q#SqThY4GwjaXmaIkE2Lo;w`qH9MET zS3PenR9n7eT55WZZZl}Vsfm`{NHjk6A_Sir_(p%K-!AR?$qlMQlhg|+6a|HX5y){~ zdRI1>y-ZPaFL9hJT}w^CoDD=YbaA9@@RcH_H``o`?>jS5udku z*fBlT8-oR?3bgI;OWP~G&VI|e`da>xcoz^I+`n?_Ro>`~z$P%4PSJEm-KWfwfQxTD z5I7i#Mz4S+IL+2aZxpF5?>*MP_`X$9>9c17;uua`jsMD^z~@Fq+nGwULM7*$AZ~*R zQZx|5mBWa-Ga`#39q;a+d5CAxrIXu-jt1xfll4G_Ye~MME2sdGQf(ETYt8Qak!JUy z6{h^Jo%*l)Bio7FL+|7uGNE0!x|(k;Ff2QcaK?6(crGo6KY$SVu;9d#2|UerVaZc!y;NBMjWBC#FLB&rvC zCtSA`vgg)=+0@*a?@<4?+V7x;+{Ns3jucBynX#If35J#wNLWf1^R^g}!V?Y6TaK0+ zIrLcVCki|o6gXcV;G`Ex5gVk>A^z2LXQD)#03yQ*v&@gdWEN*0rR}=aY(0*l>yLrSfwGJU!OT|#3* zYYTnh0F{_u8u#6LvpeEB;7aaN0uN5xJb!4w5(p8YkkD{wj3`L<956nYT%KDd#-c@Y z1bXJLw2T)#B8VdwQDLjTg*z{gDFrVbSz}Ls{xZ$xMTq}di_gKdWn#3X|Ihd->${pO zYtWU-$J)v&k9wS1TnNnhEqiX?zAN}S7ur8ubK}E;-~J(UmPW!7BtUGQ=*^74!pavA z&Rm0$nH)UcGXgleWLT(SH{#w}^pmh%Zzh$D_MUk19ej00h)gL#IF?~KJj%d=n}sE3 zvfMP1zkD{wZKbEdcQX&2riYbVDD08sM1`(tb&8vFoZ<422o?XUzz#x`?o{R`?h)g? zlr2Yz<>mdWqv6<6j`{1G^Q(ebTB)X#269$(Jc1+(HgMGOH>f__EsN3T zo&rpU_c9Yepn%ID;{?g?3=8B2%Y9n(Q%0H0iZmJ-@-M~cl!fLOr=eFU-HIPKjb{r141E}TQav(mnxVUeN*3;=Ar`W& zk(B=OW#y$hV z1va66ie?(yNWfQr9l-ac8L5V!)}~tweI7BW9CSUzb&S(rE`+QcH@s@Mf;D^ZCbiY* zkR`(Nr@8r@<+<`H1#YJLhIhoJW7HCQ@aD)25GPG$6BiDoa4O15*kl*T=MPZNYjB#p zXC}w5mE{h_Vkh+V8#2L$z8z9f@ z&kn&XJ~Tzi3kuS@?lb z9O8b>H%batw!VBcS8yMyjIBjT|Z*M;>`MiZhrHGf6s~=_0Ia#J7dGXw5bMl zOPE^V-j`A442!c~1)TwuOI&IJ&Z(EqY~pfg$=$`rk_>7~=*1x=u8M^OI$%k}P^F}H zYR|$g9CY-M$jHcsc{umY4vhiL5jRrbq3d08?`*3NmUq#%#sbfN`h*|cefLRoSRJac z2uDp5aNu}z_9{k#j!v5Actm^hOV306vdeEERh5Y)HlN(qPx5?yKOM7pw4gMIHS~}n z2U-&NV`~HHgCG!_`Ly$mhjD&papNW|UwC+Vt3cKJ{%BVOK26d*oHwHu5>LrTw$d9< z;_$MfqM}Q@${Bxa=IY%y0k74Xr~*n5s~OwRW@Ob`G`~e7S^i0Xfnh7lRL;L zb5B(j2I+Q(j+tgAoS*xqyK$E;9K)4*L3S)72@>tX!61G2JWOUq=1p#e91CQA&ON1U zTMkM}BGiVhPfyYb70dj(mTN&;MkpX&2$#49j?nTi`VdtSS9qzKAP`nRFm ziL@;I{KQnj$GLiW8{P%VAYqtAR}*a!+>wE)R$L!x&kFxoVe3&*hFc4pT(_lw3HVw@ z82@}FqK}p=^x&cizqkQup>uU(uRx46qY0Jr!R}c6@U+TfUFBZZqk3w_V0g|%6q?i5 zz%6Aqy@vLVh3mjO=N?_IzS^3>O>BH78jZZ%O*R_V;k=5Lz=+g6pUMjcU zIiY4jwTH@S_c4qun_|-8BevOa#vmp}#xD~US!JmGu^~l<5a|0u*rM?b_;8?a0Ac$) zGv?NshM85z<%%j+OxG9QUoe8}geBazZ9FV3bqL=t1l)XK4nE7vSF_H4dr;=7+Q3$a z9R+iZJh1-7`#=OjgBly4*u`gdg0=*l^n^t+QuFZOuMcOZz?+nn?6#R%)q#)voF9Pe zhBa4dj-rw+Wc($i;3q3261LBWs!?o8`5^EJ_RR^MDW*sQH94CmC$L{i((BFzrtlo^ z{1ViD>dEGG-DmL>$;i-O`E*2T35+E_?}d&6*rVil+I|>VfgmZwX3^5X${>%!v^B1w zLC{F`{qASDKF=(PAwVajFDHhQL-H{iNGItqUvt5f{fc+96TkO-mBxk;J-%8PXlWmj zHki-GWG;QaiNq3sRS~IK-_VZ)QvZ!CEV7}B-ky3-cI{Ve$WBuC^zhprp;~JenyZ)F zOZ|n^NKEE+1KiAb;GzV8PZ$-E0{X-N2au8^5%r=idO=i((L=J1M_rn(SzY`>>5o~O z2tk6?9{(n#7?lN?>=U_;c=m&LO-r_SqnQMo?C&2`FImm1^bPh)c^fbu9vwPuNe*j3 zaj4tRU&~9gP2l#ymb7@eRF_||JOIuQrt@gM7`22F45jTdcoOqM@&!G7kzhOWS<})! z_917@)^Oh^*^bQt`PS`e!BU>)8x+M42y$FwO1z~7`M7vG!f$;GeRaa30mgc+qqtsz zT7rWd%eMSP0gT0nhp&$UW8dl+_*(Qyknr33N`KW+gXcNhp}YHlqY*{1JpmU+(M>{? z2@i&zH+dLLW0)l&@RBcQ)g~?Yt~$u1$;;Z)1i4koH_BCE?6)DSK6HvE`gP};DQ*Xg zc2-3WI<3_fCfdEZh9;RgM)ymyp7xjvgb@(PYJDq&me5vY4H;u-nSha4_P%_70e}Xz zFeElY!NLmh(WMg#gH4UsM$ly3KIm!_k;ZhF+t-0ZR?%a=L?Yt`BY?<89@a@9QJf_3 zSdkHNW^yL#Rs1LaTEV*V%y562Ml)_Bxp!qoBn39jw0uWB*_%QfIRf@3p;+!mKUO}D zT+dBKVZeL$BTc=1Y36@qUtt}sT{mknef=CI;((WCd2q)-e#%12KL?-L2k9GwFyNj@ z?MX&M>XMXxFEcBYZK^A;zMQUlr04vNIenCwxlw>&0!>U3uaPU?AT~yiszo8IMs-}a z1gcg9O-^+zdhw_Vm(WhQov+)MSR9y|v$J#Rb-(BPD>&!M32B984>|94xAyXnpJ`T$ zI4ixQh0uNt4gU-zjEO`s;n;s*%)hmnnE*Y)1l+zTBSwnJ3H|k8;|42 zgfLA^Oe}&fHrKPuc6-wZv?TSV6%)0gM6n?Mh^aFXHUxX?NS|Z($qBWnNC+DuK?h%W zV;AozEoz;cU>@pTKVqHe?60e71WQ>A)x_AP&B5n2dcHTke=pM20pKN!s#$T(k#q5J zakslNV`}szWhSG;5)*NWe^#T{*bH`5t@PgI>0K69hiI4h)_E|fy+vjkticpU0P(1% zHM65J7Qhwoz3}gxAIR@|Q}tF8t^*AB4xWb6Kpzk0CYjx*x@XS%If@~ZnF|1g$-U5N>Z==a(%T^TKzy|iGRIW$HJz<=F(Oqs8N#-HN5y1}`_o->ICzE{}p zWBQDnD?7;C_PHHDaX?;04(SNuRas62V@e&RtyWXK3tCznj{}Ps)drcUcH-FrMxM{0 z^Wn?s%_!Q8zfdMry2atQ3{R!WVJ^tv>gr2VEOG5Mpv_To zNOED@i^&()^O`H_FfLNj!_bles#>yKvc-oib=iQe`g;hXgmOuXdc8>xe^*ysuApgR zCi=X_dn{Lxa@ykq);eOYV6psNH4~m?8I_Dyi0;Ai6nzdpcIbexn#k~rZ40AQ-CR15 z*uwX9!HhS9azSjzXMTXJdvI2jLtZk82aX+Tjdq@u9~BU-YcSOG|@?FU_GW-F{ zLAO~>{Ak~}@68!ZDI}{$UPH6n(#}`IE+zZJ_;o@VY6&CKFMG_NlmMKFIRfbe`T7x( z#e3KTngP4O_Vm%y-aU#Blbg2n6;u*I z{tn`lELe`y+%X%lI+lBoI!W!(aIeFGYHd+&3cqbMTt|tmczEnRXS3VM{NNU(0E~g4 z6;dirkr^WzGI&2>*JDV1TfTgcyOStjD8QmupVLTrZ2FagE9(XWh@Waw$o!fb>-Crx zP$Wn2g@t@bxE~uzo0S)rS|mJE+2h11K1kFY7w6PG6hI?|klEPzO)xNP{`P-_^*G49 zss>>AtJ|zvervYPzV4m;o!@IOGnkU%l2?%>JuYD?@OqF}%tu zcMhl(<<{?V?mR45ev>qJ+n*FHT}u!Q&N7#@8HU^dW#p<5Z%>t>xxqg+2wz3x@Hgx@ z6|MH+slwynzwfjMq4e&^Yr#VYsEpt}hy{tH4)eqt?i&%mpc1= z!+Ph(s4$UJw`VIdKC*x>uVd=_y-nuvRJmoJ&v{3z-fJHa9v~P? z=AAXscAIsWtR?k7`5sQjXHPA`Np%aCf=~*NkTC4jX?A7$z!KE-oc0qX;j0h#&>O5|IW7;Nuas2%$0Hr&+yV2G;GkVkb99)m5)r%TA zpFVxcy}e+*K{$_5q?#=o_&_x9XQ28s-H^JiDT}>@mW~hI=ZRB>W`x(6pu9^@Peil0sak}R+D@CccpFDgzAAvm$hnq=I-u-1(RiSP)3iY}=cQa~rClzZt!^65$ z4(LL^=~l~TSKF=0I&xv}ZDwf9%4Z($ABU@lvPCPbp*+YUSIKS545)~p2vCus)(1om z!NFN*j|rZKJD=33bxI5`_n8WG5rh+e~V3fCS3fuDz<>6O)m# zKv`*Mq)3u6y zPG=-8BIULD=p}J*P5`H_#}&W2;j`e!v`gxb>%FtUF!i<2i#$UFj$<151MTALNZa&O z9&P@^RCRg=Bh_}4d@4MFzD_wDMXD>2%SF+iL<-KMJl~i$;zah@cTsgp>G;J8z*}p- zo|tsd25A;DT-m@jA;kFmo~)7sX+j$_)F^dN%Lr6G0V@eFv6pdw_0bHFE+c$}rFzQV zdkgU>w{WS&#ZB0#=;RU}%Gcq+2pp!$yYozmqqTd+FHUxDU2o6T5yX@6<1g_!46)HO zF@3GIubo!%_r#x@o97BRbu@}?vso&6xoiIZREOPgshO&^bA(QAq}j(MIx?@$cJK?I zg)++{$$2+;FGUX4w{_FNgnYR~i|I=9jhXx+^lD(qM3-bM-Ad3Hi_L^!AG4FRiSwZD zy$*!L)nzyq377Y_p0A&Oba$*}24ED_+Fx=|>=^U8<}d!;OMoGWi&TgjVqpT{?lM9^ z3}n&^Q{Y02DPlAi+8CiGQ2)>v(GBmcj?G(rQxd~OkP869`CR)*6Q-twF16&b9wJg@v5L8*G3P*-+Ks_H6j?D6 zpX5RPizBl+2SAAUa__AA@JSW_|F+)-y`kwwe93x_YQEBV>C61(z6g{z$!x zX_sg8hf!qS&OR3BB+pHktf`=wXRp zcRj>k(M>YPFo|E`-W9}?OSlWeBEEEgDLd6!W!V#hwE~z3*VXStVG&_S#IEDV(K?cH zc>0_6*qR{uker&zoCoLO6i_vyjKrZ}>7#DFzHk84=DkVL9{(WF((o8iWWX~#dqJOb zbDmEgl~Bgm%q*ut@^o`du4|z~2CE8ip|(=RM`}=OkUD9-GMkhF(Ylx!ssG6@+=u)$ zL`Rm#OU)V6#y^HeI)us)41RL^z(KcW7nI@i!ef>d6CU;I?6GpL%r}6~;txEO`jASA zNr?&V*}N<4iH<9}f5Y9wy}P5!_pVJ&J{Gz0*HqEaJ`!QijgT7KAqxPy9X5-%Tt+dA z5cy!XU7wZ~+$y?EyVsc^6RRC_iQpKJJn0Kq0clnxr>eJ?^WQfvCN*UA$I_uaF z2kT#9y>t4i6Qfn@z_1ty)+zKgZ6gkuI>~`}-VY0AxFbA@r`mEa>-L!#r_Vr_rvE!E z9Fxmm+SEBpXl0 zK)T4xkt!K_phh)#QOv~=ql0gtdS@8%GA-hbW>F05A_!ZqOUr1FyoRQRGK4HI3n&5^ zHN)!dCotI!Ym*6AKXH5%%m)!;al5hRuo+pU0G>68@M(wp)LY}ELs2$1C5;>bXZW0+ zjSMSc#R+I@U}z=QXVMrF3OFr8=QY^V^N3mJ{eWspM;G%sfkIdzLYz|0?VZb>xEn)l zVIw&?)^#@=W@2WMh01W z%u^Z%$uG8^prnXa?S=_(6vDNf4!Y`Wd&Dp!C(8{LzdKL*=DC_VeO`a$zJU_LPpJ5T%_#iZ9zhQ*jGao4YHwwV=4V5{|;=6UU zZ*{F2NB7tWE2Un*D7h`c2~E`lj`u2G7@={v+f(%U4@su(b{^I)7voOeJ$q+&TkH`= zGBVeH{Ydu`)YLg0psvzhdRaEd^l7<)EooMWWx(Q~D^eOj>sdp0mis=&<|t!mkwOVo zeeOuCec6kEhvY?p1T-Ta_(xkqSY{$J5mf-T)VN&X`t&kjEnhm*5vGo zb-@v+3UZV%)FdbDCT^Uh@jtX0`~oi~H9!ds4&A>$dSydC)08xpg>AduK>p zDY$n%{S#z&JOYm4zL{v=v+o>OPqLK)Q|pFb$}SN$OArsdcwysAS3NAIG#ly;OAd|7qX_qa=%_0&Mm z!qW0!$LFiU!Bjt-I!|vUI^1)w_0rkO%4!kh58`M;95u0OMpvwh=A2bco+kH)> zkg!oE(V-}X_OlrIR}vswn8IT7G&O)sw=vX2y+Gp{U3~3FrJ(J-EvA{)bIA_E4yD9T z#~pzmi*+j~zdxLW>88{^6m(o5;iXEqFu3@wUrtIwSomtcxRD}t z6)v0Obq5)sQ*Lx?);rJdPZ#ulgT$VwB3WY*?_FjcNg1EGmudIQ3~Q_hh~J3S*o`p~ zsSh=O@*e&)TfC*!TYNOmLQ6gvPuuuGmV#W-Fffh$UB)R-%&I>v-|WQbSjm+ef>EU- zcFd&f;$-UpG`jT%B)u6dcOpdNHnxE8kH&K50aaP*UbygA82M}(L%qksEYR- z>lV?8i!ZDRO};fIAN$2k0g@DAabAg3I}g(}6)lK$!n4-hZKOqCVs2gmi>{0biQS?S zx#Zw*T}nJ~xxP1eEb;^KD+1=a16mHhEcf*?=6e(k(ME`0Uh$UcHRi#)BAQ!vP`fWf;R3@(Z&c3m9chm3JwlNf`FpDtE4b=mtX5nhRyD2M@oSo*Cybt z_EYcgh{%X@{`xx50gu#a8np-IX>8f4p$=Hd3nDuTIRKvY;&CM>M%FV0LW7o;R?u{P zbQ_uu)@2uKUqdmOpIL5&P4%Fh#dT!{T>x&qd5)alxpG4)zWEE=C!v!cZ#E`d$cfpn zU>@RuxVCZ28HGAV-rv)M;R90LnU?rZ(7NyL7nzc`as7lR_>UMBZW~^*pPtkHT=377 zp%`P^P0YL+vCW-kcW$-Ob5rt|(jSfA*wne;eSaCs-Tl^kAP(9EBvW(a+gx%ek z;-xcelBt|UNlv!wL(}P&wSaFVkNRzNx~|+pWGrH7pBL89mQNqC5gEv6J!3d*w=maw z@|g$Qt|>_(`T=5M|0UW_)&o7yp#(fN@rV$2A9hK2h)*bb@eR7!EVs*H?P?XO?P;cR#C;RYyS2wJ* z^N*bk@Aovff2@9$Z#EGcr5AZpbDE3l$ne>GgLT7o*fSEV*chwOa_YgQ(SlZbBoVED)hG> z!&Q1a!m*sPetpiBX1D>c-!*>;hgW#E9TOtOaI5Q?l(~i36 zgd-&bN5Ja|*;6V;;^LVao7j+QCN!>^Lv=}If2%U{(6of<*L%s6wRNTpYS!pl-l1lD z4|Kf6?L({2a7K<4ky!I6ZRtY9<*T#3MDymg7;y2(%LSF5T((A^v&zu6UvS9C+K^T1Y|v@5ls@LWZJQ6JLaU zLxA7(c;suQH)wz1`c%*t#?;e_b=61e5bHgPy!8}fWQHK4bBQbm8KKv*)#wJ$$S8tb3+G7x5Qy6?K2cew21tKEtiPiR{6Q&3Wu;SmPaSN54Re=r zS0{Uz;P6O4u(-Qm2ti8z^m`az7r_PnNQa!uJx}Qis?J-?Eop5Wj1~Cx&0lQ?+r$ zcfPsWOqEzWlOUF1IV5Yl8gPpo%)KkT*FUarO@e;BMToAc0 zfsrrKbeX-MqrHQBqW*-MXRxD;t6BRap}JZ43Snpf;xmi;TYYo&xEmJ(UYF9F^Ids& zKJ-||5cy5c5yEBdS7-bi=rH0^_kG~E;pf(oB%_(X`r5c$cH1a~Cfqkm|Lb2~y$Ruo zmB(dfLRJ?T?{UDw+OsjVVlfFIt2=v!v{}amT1LX2H>1~AT=7oj7qP!2!n zk}{o)WFx%kh0*2aVU-7M&Y?%ldEcwO)@eHCye80e{0PBFQ>b_w^iNfuPj%BqsALL< zgT`?hKK2RMkt4ZGVK21`6ET?b(_7fGR#x=y9Os|t*Vve^^u)7dZR_0<)5e%nFar&h z7NG`h!7X}`KHGDlwYB$4rA7=so$(txiFkddpM*2TB*ZMpg2WBVA|yn51kd2i)_?Qn zjgINd`uciDdeHxhX(c0GTb!Z;SYHFHy06v9NK+=iC5Q~43~Dug$>_`^ydT&+f!vdn z0&wgZ7036NI!BsyT~Ke{Tn6P|IFyB*4?Fle7<{N!DG;qKGp;8-TL0?nsy+ZmJSCQH zR)@mz76YC`bJg=@4=HEiW^IGeTI1j)eRCK(xqwsIenvWIf#LY*gWE_>N=mvc z?pPYKY@7CneUs-8OQ4X@u?HdJ1)0H4LmUz*rTZ)#2g;c>kR2cN8s+;gC+3sO5V9-G zz1h7p4+|8012`)918HfgGpgOdF_|?GmIOV|@w|7Z?@bof*bbNYx*uCShAXsFRq;U3vUlmVyk%Sy;!MY#;GX4zJk_nSAsC8o`*N5;dtxG?OC*%7T z8U|&dr!P>GWT2;y3IX8fowr7f8YQ@-+E?D62XvtiLTCZUrgIovU_DIeH=^?w&Z3$9qpaoZ^EuJeK(h3L|;TtCnjc` zr1#mFk&9Z*=m#9I449GTJ>ovNKBQmA#t6h0#Cfvw>{4ecN=r-W^Y_orBGXIL8-~BF zuXm1h5`;>He(dX$a@XL&jQjQ1X&#i&3mT!Qgnrp5)!?cW#IvOs3R_XY;Hr~y=vU+> zA3`K@O*l%F}# zrZGFG?%1!t&76K0lpdAeT7M$baf150)Au&{@cBbx04#O`kb%KQ8ZCKnpoj_C*9NyjodcvC zpS0+Y03yyDjT$4jc|7&T2nL08d{=&Crs^c)DXx%%Nl>XP8UG$$Fm5C4ejG*bZ&o#1 z{E?Z>u=O?;2|Hmd4COP*6-w730`%!`Z*M@ z35itK>6OQDzJV*>`;${c-*3<5zU{H?&QzL9-zMk3ox*9_^Zv&X{Xq6#Odep&+J%4- zx+i#j)(1SD8J=rci2ZCKd+X)3$ad^}BwoOF{^BdiACy}(~aqEk<%<#CyZ z4rvyt<^pkTj5Q|~Kng>Ik!PFa*zy!1GHyp8rY!m|d)3;_-?Z$Bd!GEURJVekHRmyD zGra33?Sh^FIIhr^bUU~nv-p`kjb4-EJpfrVO}^cOpPQ@!fC$XyH`(Y)3B(J}ux0mT zRLJQk0P~C4y~7?(tzUF6&jyuI^u844@=%r-xpX8B5d=w1xaA{BrgwYO9O16LiBySi zxs8!5im$J44)E2>;&QN~Z{L{IYiyL?|GK`!?;Wp_mit`p0vz)6d%qePXbBj}LfsdEd6zK9|ebLQ=CjJ2@N}(4;yejVGW1#-#frbEV0dWVrGM$;Ga9X z(GKA3+3SWiZWR_mK|P~$%AjNQD*2JYebBv5>_Ac>{eMi9?|3&iN`#2%t>&NoNdNIsNHiKE=Pp+V%7a#Kv_ZqW@x*Ob9=&v_-d=_5b-#{=bC)FYX>Sk!R%OMYS`(cyR^K`DV1an)EaUe4GbO>28{}@*^6PfCxdJ z6I=*{A1Obgh3SYXCG1U>lQ;Qnagbi7zTjkMFUR~!dYr?`4ShWJFJfwHdxlz*CY8Hrp#9RU-<4euD9QzFC;5s&CYJ4(kauUf!g$sX~qz# z8C8PqxA>StV`j|MBhQVXLP)Z~0h3iAZ0p^g*O4%D&0I3gNud4tyEOoUr^F@`mxX|; zc7eE9HC%z2D^tgI&0Y% zFA50a({q9@PtmcliKadnH+w(AM<*WoS?X6MOK6#s$`P~@5x+5AwW4hKP(~&f;3@dG zx*w9rVC{bwL*`Z><|AVm6^LQ0y(V%XhAk3<*)bc2j$$W5OCH}&E)UxD`y<-l{dcQ? z7B6wHs4&UIFIGeCSCr(Qd}Da#tYPFc)}#EfuSH!__rJ=HOjMC9I>zsaC+Ua5Tm;D> z0BtOPZ&eo9sv=X`%pCk;VPL@8nRSz_pz`{CSxp_;0aeUebfg1)@eEXoaD|3q|3`m6 z%YwwNUpPvgxF<`>ASd;wrR|T$lANFhzu1@6n7qa>#zT%AWj*=`1d-pG4AcVrvbj1H zs|y_h1@n0P2|E9Gg^8o17ZR}!(LzFk*5eh+*?)C2fXRebBl|1h+j?8_;Ps#l2KxPG zj`s(t4FVmC4_R5=SjF3pWsUyW0r-gHR)8}8wY+}hta-x9rIForJYc#4=W3!SZ(ug3 z-ufP^B~i?OdbyF4lAJ+?Uhgt5dhqQN=+Su%RrpH(_dpq$Ku)%BQ<}_d{Ne|w86X*M zZlwEEd4tb+{EoW1I@tX1loXQc>S{bHszhGfp+!LBH`wcixF5Mwv_7y8br;mwQTyAQ zcOqlW9fciv@IwGHk;-$8iHmudGGvSgnCroG0TKuiK0egN#l_0Th89WMj^|28R$9W> zaU201XFHtXRs2i?)XSL#>wh~3W-$;T6cWo%%Yah7KvvIUu*owdnR-*Yc;E(Bkg^II~WPV-H0qP^C$x6*9tpuuUbyn+O|RQ$?UA+t!@ml2=hlmJcs;Wr+4~Zt-UQK75$H|upnW{Yh4-W`ttX72DTEq> z?h_y5%Z-tiTZQ5hM8NdBLD3fWit5wDxi^9yMg3dOJ&E<{taHZ3F2drB_fd_pLv7yL zk$HH89ee%Oxm8BBW^X6YnBA~_vZZB66*vTZSI8f={;RhL=;UC0yB6Wx5&HC2SLqY- zU~tqF{!l>ylllE}MYrYbsS`Les_%7Bp%7IYoWR>lOKbM{Y67u%iZ)9Y=#@4;$8yKh2f)rW{dZ2LD*vE{Z{WXC!(K(s zR#a0{(@)taH1B?%vSJ<{eo_hT7>3U_L>QuJDZ6k!ywe=EmWIjv-<8lUE-gV^I9;sP ztt>wqKLvQ%t-IoZ)92!zogbucxgCCcq?F19F*i3~rubkb?+9p(W|}f^nwt~Lp!iFz z-P6owZw0mr{sEwB)fq1+-vEoeM{v#xS#?1#+8<}rQXQ76gU1>9uC0nV5<~bL@(?sx zv=&WEPZPzF5mqyA*2E13oKjeMb#Hj_c_h_`AMsmNTcw=O0Poegrf(z%tb?Ko=T9JI zff9=D%+|@NsNkG#&p!pZ(FizBA9&rpE*Nmi=I-u+2bxhv@@477X(+!2pOwECfBcy` zLlEu;#U{J|Y|9V6cA3OxH=2)l2F`xYc-V+ccnM!x{f(?k649z@Nj8%0`hSQpi&5wY zP37+zc7>WI_(os&1iL~xtk=yhB#JDTI>PyJxji@3lz#PkhLLml8A_zb)+EMRS#{y%U7;wqkc~p1Gw7zNgy|#yTDKKhx7qyE0QAMbn+ZY` z8nQzCj@F`IM9)~C?as!o4W{KA&w5X{*b3N>7t@f`gKkU5x!RFNj`2QQ00#Au$QKo% zmf%AWAr@r>>?+v!Jb3n3o6ZgUFJv9;tVZ%f4657^=*2DYfU_^6O0(~rWF$}GHv#SN zd6a?w%z85#95rHi-TUXiE4UTP#3XT#BTZ+qKqdj2KzCfc^n4i!Su> z_!8D#zlkTMOC?K0QX}1i;+VzLC{-IXEk#{2Eg&70S5_>)8`8hQ0xD$KV+vW|Cy!8- zpzxTt{_)FW-Co0Wp0x<2!j-Q*Q6U$PPXXu0U_H+(5MZ;KACh*wj|x+3Zu>Pfch9@+ zEwtY`>V+=0m~s6spcZQi<2|U|sy?JEq8huxNw|jv0be z=b)`YYS!Cn^^%pA7V5^dLXX%sM%iPD3uW;~lNl~rwV#H!bh<8epk382GBvlNfP~^= zE#E;8hxTWTPqRe&TFjuXm@EJM#W&TVYBEraSu=Q20tKSZ^CJ|NkatDN~0hQ-QC?tcSv^# zN_Y7?eDC{t*YmuKweT+&aLu{S-uqKqPI>UZ;{{49%TSG)CDrGgp9bMvg<;kv!X|z=Bb3KTZuW3?Rxvd5#{1Bq`yMhqgZYA!u;^XG;9>-M6SYE}@0iUzpr4 zkw5q_$Mmqg_u8zBO2s#ZO28X|$a;S(Sro>kewzDXe#toKF^N=lbn14J^=E1f^Ff>6 zHUFo^+TTa+KOR5ho1A|P#1rZgc0iQ4)$_YP2A~Q|V=xESZFMTY3FH3?4oh=tsqo!xwZW|v$492#Jp!di zYt*8G=jX5e8@C}J9A zuTJnp$;G~OxJUMU!@kfO|9?y~x994B82c`cM6V|IFfdukm?}~T=}{ZpoG|?Qa$Rzn zK^}b%T$EDEVzDR}$~qv$ z3B2{^#B9AE4F$bn^Ayh5$;nepb(hCa@6qz(my|3`KY7y-%ZSos_oOK!Qb-cg;U~8H zpF)^v_OO&CK6W`r>XbhI;!555?=+d{^%bebK%%wl^Isyy;}Y`UZ8^|=#3Q#paAjCb zC=!T`_yBr(E4l=_cqa06b<Ng6^78lu3?gAv4tdwp@z7LV20ML}e1R}T!32LZ7FFP)qPv`v7L(tpsAx-RVhB45qhzDymC zjEv+r8rnnn;m;e`o}*m0K2BQJe^y%eF>w*CN&?Kx|G8Y6-@+8GuiLjsjMGiHbLf<0 zo?wn?B=&Zd*S~utr7G^SJ5P5zg7HQJFVC$YFX)!Wm66rfv_3ukf8o*C)BwFo`<8s2 zvSM(z*L9F)<$^R6xjecDP!wnyILM&7$?WLOw*#wEv2Jn*tmeT%!cn7dpDiHin_ z{M7XbC%fx;1c^Y!`#IgdWYGntShmsQTmGyEWn*;vvikMBYGwh>P{Zquy!S@GIRCjx zh$G;LxwAjumHoq-{XacvKI71Hc4bWU>^RF{%Z~f^Wez@`CDL+wRQ*v7Vb_he3fQ8Y zQ^7?{5dQkkRHW_~tbd;k2^+X3*raf?U6OC`Xhbfn=CtcH z2PICa*PD_n2%_h*D{c|;XvokfSdZfp-aeNO@^3>DJ9+kY1|wR28_A@RCV98pQT(0Z3Y*DBOUIb}PlLNj(T|EGClK^mx!5-I$B^>#Q@mwxMV4P=`vi%1`znR6~1Qg2dp;`BKwJy1JV+)gj*-z%T(|=Eq1Utx=zjQEw(yxHEq9OQ` zZ>?%AEVrIQ*5Iti6YlJF`$K$JurWOlXf-RIi-|r-R9vzCWBjhbIAgogj2z$6fbT1V zPJHh*?et#$9F*|96_w+XTG7z5z;eG+;f*DUaw!MD*GB6v!i?_CBsVpQmX?)8L`M40 z>RJiIZ6mt8Phj6gS1tfq2cj)JQSdZYu zOO#~_X0DLm2}1c3p3&DpsbH@6Do83m#Ujf76zEkE zu&pRQ_o9pEsFXM)zfCG?`GCc})ilRo!e5Qa_st4hw3dH|&!qK#;L$%!i}nb9O9S*( zy_cDpMb}2^(@TJe&r!sL4EckNvPnIa$-=nP&zCd!Ii#pEB>p%UF z-HxqLJDspgD-B}30`H~CLjIXl7>ecoE@2gKaEy` z@jI-i7O^L&>--C7U#=o$_{P4xc`N<^`9X_ozXR%aYVm+oa(QyW zrvtRz!*!Mpw;JVbf&4%)5$)6;DyC;;nf@-#$OhVR59e*BZdReYcq(9A0r$-voJrE5 z0$YLG#0|``VNbx*-Ta^@ExoR08$oYNl$YO4*kLnamm|ADwSHr~n$7 zE84Yl=d8@*ygjcKSNk3Nw~aZ^1mW}nNqM1pw7ucnmUiB)@AmW$@x^IMls*>=eJcbZ zB=7R80V1W1^&{XmcIbJ4deq}}pH2!389TjtCUn@3LFcyd_UVa61JB#S`hIsR{9JT^ zSEvE_<|zf<)85u20PrrMmT(AH{x_XsXl)mep+Aq?x81h?Rr%=h>F zRvuVzepa%w_VVNMgBx9;6VQc&340~@XYyCXrjhqG*~!iv=Q5B#E5kItAS-Ofhk)1? zB!ehE)vbMHe_K0?IW+eC8Yww+!1`Nb6Ir=$m^rJR3SjbZ08!9tQ9u4ZjP zOczjvGKvKzsMfvTjl5w8HeDWr+K*s@4Mr|+@XJXw?@`az*k-w5?9EOT=q?mrWwG>H zAW+)}E`J=;>M5&O~#I6|0cVfMVZn@3G^3gj;@Tz%d8BK_XQ1H@5KgnT%cKmvv`yCdZ z?+5bDN@?mg<0P~z-=L^KY#HZ^iagDt*5NcMbE(P6Ns9j~Aw+F$QU?lRqvU;IHn-7T zM-fsPk74lUKZ0c?Hsc(15{S!8$7KsJbN_9>05HDIEj|H=1;hMT&U!r4R&=M1r(aFk zedWX9ZBpU1tgJM#yhfji+oTv@u<7Y1KR}_YWsZ86jfm!$s0(JwJ0tE%DGTKCkVKJ` z015@4@*$)^w~|B)+38ctaKppQEG9zpn?oT*e4F&k;p44;DnWb%$8knJE#9YdY?|fX zo~J2*6uXV?(e40rv+-s=%*@jKUNd`Rqs3q9H#3#}zwLAUDoXh(B<|=^%G{$xfw@8O zd$!#*xckEsZoSmLA8?%c-an=34cX0@o9i!tqjJ;J)2Jlf@zqAdzeY9_#=%d6 zPe?-=y)R9+-l=tXv`N|y0a{r$Y~8ME?!HNgeV4>cyA)gNOfOB@dLKvF+Uj=V|Oy@QgS8V0p)GF+2DiQ`|Wo$JO+&{Nif9QFjf~nj=uLHJ1ni+ zx>rw}*wW>kRUBZJ)!hETG^O;kzd22T*1^IoWMy*>Rad1QGVi@>K>Px{V6}6@8fP_@$*+(| z-)Qq+z?@QPH}`41KS55#v6gX^yT)pzQ>8bSdw<5EseAuVpr*5K+HKCsEvXxc6Oy&t z_>Vv(LWS9$9c^Xc;W+!@KNqvR@vD^h!YUB74Tl9zf*Nm zPa6|+Djw%6mW#|;Yp3*8tkP1@%DipcM+mT3)GVRXx$@=$r8AW`yzgWo25uw1k5<=mluR^p#3YN#y3^cP4|wlS(R&W zhUnsPYyiN)8*F~7dfgYRd|cZ)JaP@R7VX@h!%e~_*l^23DR@nQ>U6KIp22=Ai7(Oh zsNoff;=_R`8EUH^SueK=y zNVt{CeTHj0o~eG8;iN3M9vHijT3vL1g{;WYd~4UD{z{0uVbRY$tHE)Z5E^5Qo$d7q zW;2a9y0=_TLC1qOOx~}Zy?64|@C7`*CKJ{}%Y^baxJ4{pL;dU#%97+FN-MpM>6IPUGx3j}|Pyx{Tge7zO}y_s0srSq5Vb-9iW?*r*(G z=41hG-EyMPr%}ItF{iSdfqa^lSD_xml6e9J8(}T0l1jspE$u8Mz#1 zF8}5wMbn*Qn%B3%BTmuYY7h5a<zVl48bC<29WrpuSt{_3$NW$2${fNxHPQL3{&$gloM+(w%xbn;0UO{p zr%8GL-M1I*yVMfpPuX-9NLCw7)o^w2o^A9OekyjdEM48t{TK(r%jb??ay%)K! zSga|oQvWc4er^c_ZL-_Dy+OR1Jua-fUJ2(QXDL3KE43>FN?JiCsdOw0aRP#}N zrkUK64rEM6{AXHIlNI@rW=zTwW!4IZ<$5G%Uq@J^z+H#GA7rGhZiP6)^C0VkI1Z6K z@ib&9o(IKutMg)^A}LVn!?kwE8-e9|p6)C+8O2bGno$Z3lz*3aG_1~g$D=?NErbK6 zn_21EaO6_tcWm|*wn0gX#)EoEy8p5um1&`R;$sEN!8;ydk3iUDm^Ml=Nhp|AF7EYd z;aosPJeLkYBU3Aa$WQdjPIX2C3w{o9NT9aKOx5j5I;Wl&rVV&bOEbPCUebB{j928@ zGBm-#T5Nm0cgvEYb~%7y}2JoYH}A;TcD z-(I-vnY=f9X@n(qmrw5=6iSC?y(>UqdRzl@{S#9^-P1sL&jp8a3@pm<<1KIzkMe!O zyj9dbdh&OMH3egC4=@ZJ*Wo!oD`M#UHo}){5X9>t8;y3r-Nir|2$#HYU)EZb9pPKpU}R4Ra^z+YJDFNh(1P*DZS)*Zx&l}HL4FA@*^Z)dx-I!Cqld?Q>{DjACO;B{TLWYx?MiwF}-*Ubb z$dQ`pJ?HEb6AF2>$9JdafAwm&&SnWV<;WoZ;H;&#KvWvY&5Jbu63gbNLE zok#e9nD0~I0iCzNL|~w;y>;`**)4UR1A&e*EJJ*;NIW{NgA{MQ^3AX+|8JGe;>wY! zca|mIsGslkYurv2eG6izUW#w4@xBs|89TsF_L=b2 zbwD(X|99PF|G^9iVV_7c<Mbruhr76bozSQBhc7jKwwd=j7S%S6I~wr|wq^q@;$p7%VRY>4O- z?kRHPT>ti1EBN0ZLD3 zeWOTYS6>g8Sx>6h)mIBToSXR($tI(uUbiQDr9)Ta&#qt~>@vzdGb1yHw;XT-Ayho`wPeI9DqZpGr|xmuXgD}e6BBKz8`qFYzqPs9Q!s9 zG{^O4YBd<&Wt+-n_s;`Mo}lE-4T0@}<~T9T3?sj{z4m>+`_J8l*?Rq)xpx05CE@ev zvB@+6LQ?HLSD-nOx;*(SfgH}&Ak)>TLQ4EESS?4SYzYFl&U3K_GWFtG=f3SQLS$?@sR+3=9PgiV%!>ud~umReYYiWwc za2r34j`5+wTkHCLnDdM$$UJ_1ls&B4oG{3G(g{Y!-C-$uxLWL}O}ks}z>nv0%nTOy z2#Tmsp8->pvB}*;^T3~+*Dwq;R&+pffZ6aqzTGz=4)Ho)V!Tp8kISGEp+GGV@duSR znJ>98bIB~iWXRv35uiDK8)s{3m^Kaz(fDEn;Io({rQvpxu)1e%d*~2$tOAd!!nuba zBJs_1Jq|ztBj}0&d9zZ*+7zsvGL?*44v1ylG1+eu6cC^QWAd`a*Mum}SdJYpr;R`V z*KCRjzthq-iUHznySfL6PoWG(X#@WIErSedrvE_PAituljjS>8b4{(Rc)Ta7Hjn5F z!Ju31b-wI$cy(IFV*`~Dp#lyC(vf$FT5=G&R9&D1UCgIc(+(3>X^aXYp8HWy5t zSkcK!OMeyDLI|H&%p~j$VeVraL&rUiw`H>L80^IReUtFkv2vNKbz#mtqpRPG3jlU> znXY%H-=4gx14*hKC~Ir}>!j4+5{+_g74;n7i^Kj4Egr8z9?6lr3=Y+l|IE$irPo_l zX|IGQzYnINyuyF(?OyEzNoyXH!Gjv&{tZ?&(-b#cvqni*Ig!J?z@PE@ym`$t+3Gwh zaCJUy2L-O-JS|=_6@ps00xTpo`eqD3dF?=pSp0TP)@i{OWU+cNmc`Elup?vosq38f zq97w9QXe0T=b=CejP|f{Ay#%4e%+_*?EF{}mH3RQ=#gVgowZP41%aCow;`Ydk%RQY z#D@21fzF(H=nURM=mElIo#xUYD}i`(_uc;ZOW>K&Ps;}S#+78sGY^jw6|z?3*VB+^ z_X9@KlBrw^fINwxnvF6zESJ=4Tfu09vprE%RyV}8l6*^y0tatMD1E@l-sLBh+%>!I z5BBV*4)bBF(ProL8Bzlvp{qky0@`s`DkNVcPpkX1NCvs@8P)o`xdnqc9k|6Pl=C;l zoq8;zfi{R%@x|FIIG}t_hBaHc%MQ0PiLCt!OWw zaXOz334QW*Rr;ekUX6-puk=U>7~q-BHr1|INtI;pi)*%9NVwDv4lDSY1o`{|xulVV zyAF-#lJA%AD<-l}qywcwa472>)|s)GE0)hQT2Xh}aO%gADBoX2P3P2LC~ZO4-3fkH zXCV~kx;2Ia^4;2gz7vhTkq?KK-nyQ&J~}jB-T`sa#WyC$_@UTtYDg*`FTNBeR{CV? zd{J&?y^#DU35?B`{{gYM;^_3}N}eY>5#f@$L00F2ZkyNA8?6Bzv z!h>0L)avOwz+FUq`w<+yibw=0+ev=+<>MJ?l4&L!kOk8fj$mfGI_Tw%oE3vd|8@Dw zF8-TdCZ%W3@7=~A&Mp|eF;DRuu*x?2I{gFi#Px-b7GFurDMdX=YhFIYPeyI_m+`(& z7D^+ibEVvxEaDyH@p^AkkhL8=NlZq?y0MR{2{Ew&VYAB}Z3?C#k3>X=bz`EU?rnFY z4$tHMWd?q&q`~|{qO8GJf|)X0`;xeWj%KF^2pu4BMA@+;R)E7aAU>3|M_9k4nuDWd z@;Qmm^izEC`Fe(1>RyT`y{k!UZeXw;ohVU~_qxu5kM)jAVot9X3GU93SbbY27(n@5 zcM`-e6)D(`0{I^ccHGP2{D}d=AAy~sAAt>b3Fzd>m{=X~U!f~NHumY4YNOBRU?g^U z3oYLXIktL-=@oXI3C|6sg2vca(+$BYxR+{cc#i)^3lYHbS|&nYF66K7+kI!O6;}>! zGQbNQp!?(5Jj{kjKP-H3MwxH8E(73a?KCVQ)i+h^&X+0R4fKjj(7&V<)c{CLYg?Ps z@s`Fz&2NOK@=sOBz8CN*10~_t_(z}k-^`a%P3cT6XB@V#D50gk(EqBlEPBH=iu=rG z_Dl>a@v>dP3U`Xnv@-}|SZ)>3o<2OjJ31&5({HSs&0b4&fus;YiG&WT1v=RrL2wBO z$c+`_T5f^mM7jcXubYs3(qP;D$@Y+pC42&evIchv<}hC!e&_^{K8kazX!laZRcAx4;zE43sr!=4Pg6{)@m6?HB&Gcz_hB8HsN zIX!eF;b@-e-?f38N!X)r@9(RXRweT>y&Dj@%r}_0DBu9#U|d9+WY2)X3-jg=7^}cO zPd-ccerGo4^>t3`|Ire}a?8)q$^ukN7OqU=gPKCeau5*1^r2=Cam$}sIXpWOO}?rD z$sDZBRpoJ5Vt)0I?8uG`mEkaMWwt(JkzTV!y*E&G#?^{RU0iM4c5L_+RB0fu8lDPc zy^E_%&Er*h{R7qlV@FyBp3sePeEf-D7^O|Y@aG@eV-)gLHkK>rHC8X$)Nk1%(i7r{ z@;Kcp;8P_PNka3)bm_IfE7EIHG#%)T;m!&!1y1wKw}=CTPUikeOJo4vMrFZ_`Y^t> zn0;+duYwI1Qe^p6*GD@Z9dxOQ2@W~o1TJ1GY(a9Xb@K430&`1d!**faV?IoBR(;G7 zJ+yRMh^F~*?)bs**%!=7mF9le1ymMl2ZY5QeWk~xrRF=|j6P1c{VHAuFtjbR!DHzDb-CY9B-mk2 ziyxoAUTcl)s$dzo&7pp5?hVj`V?pg&Y5{95+sXWKB0OKRl2tQBivbnc84IX7(w5QN z1P9X{iP-VdPJL)Tvz`F6MwRWo6wp7kZ@PZp|6X8!H2Cez->U150K!cFm>@(R$kB;` z!4>A%Fs$}!I=RC0=r!5aVWiJAIcB-`**gEZ?uw`>%#Gc=io%Fz#y|@|0tojOQTUp1 z$dDVmU&TwR9GX#vbWJ8Rt*`v9f~Uq|<7lf0d?qkeCHE zr-0xNebhf57eE=wG}fP||HHV}IB8W2)MhukA3dvGjr{>#kGDX~W1Ik2h894vkPRS& zo+HS#^NzR*I{{ILNhM>CfXseI!$A+HOB<8ke@t7x;B_dlZ-SLyz za~rOhvM{APXrDLg`>PNC8uyH#;t@74go9dB#qN;JEGNTU za4AOGu>usr;YE-h?Q&_4;@0?OyF*1kvHLApF6gZbh;pEU9WNw0Bg;S`78VODk%ijO zRyr&gM$s)92IR$z?)f97=?*92<-YC7|7Ac?A9SfH39*9Seylfmfwj{7jHXQ^#R{dj zuEMAdk-$gvm^G7b`?}pDaroNxVZB;jvrt_prUR%BR5FAM-Q0S==FVD=o#YzSRjL;m z8bAg4G$8CIgC*?710M1)X?XV@dU7^LT&RigqwD|zs3u5 z{vfte=39^`nzVZnq3|ic{~l@8dJn2T*y&uR&}^yKhEExiO=0NvvaBrwONG`$JC!Tw zt$T)~v8yjz{{6=3{Dw<}_|*CjSinDsvWpia`xQj>)nJTNeqFwS(tl4t28tw|G7UCJ z>9U^w6xb2ZXKL1-O$_ZOZa;|>!jZ$6;6_pXLWpL#k;)N5l2d^J|%|SH7w{zH%Rf|8+kiGqg zk?#0Ij99K-jxj`Ae<2v9(&CT2FEBg*n}R5ShLSPlPs?u_WT5;TZBB?X3qP7tc;_f7 zN$IJ67Ll!PTuJ#giPC7blc|EYLXWKtE^(+tFtJg%liDhM+?96tLJtR|bw=MJ=ClgZ zGik4Q*uVx^yDP4;hc7p_)b>>K9)twT+4_By8~S4j#K7r1!^9DV`8v`)2>%JDv1)Jv za@xDo_vA6e@y7h+P*k?`X1Vi}Y6D@Cthc0o4<(_?$Do=PI6g#zmcPR_htGLy^#S}f3oT=1 zXkUey(+9SwKAgC+qe;^rseNj?9zLS4b=;cT#vocyniS)gg@gyRHQU z+a%IvEI;LdkjHC?H~x;&WVP(HCE+mmdJs%B?OsD<%lL8X3P`px!RK4I?`yX8>AkjQo=7B4Xz*=ako zOyDj;7Fh1NkucM~@R6S*>TZ`~K%BxCBT)EILk0!zSqN;e{pw_M{uR_Q8xn%48|`$y zPD9$3eE&0t_V-Vh*B|KOW=?e=WDto*Fy$m9Vwx3+I1b?o!u1`Z(#LgM-dkN<^u2$V z4NDM9RMZL8{Ex=we z*V(TB3O^VL(Fl@nHEVFu-8bpT0L5B0SN(y@?^h}~iz}{?xDfrv-nAu(*R}gv58fARLgM3z+inm+E<%1N@u{wENn?*A=5Tv!CHDS_FBXsenADLv{*9jxp;gK`pTn1PxnYCZ zy<4W-R|418^8#Dz&LW3a6(U$i9_!B@(@6KK5ihTuP zM;csgKY-{}*k~wBT@TdlTAAG4dTdhDT0t})PQ$n+=_uMRCKWEl-?Bez^=9b~(U`5r zv{_FQhzF~GzN!)9CoMu^)pmiyo$P)=)xUK*(#!js_9JS#_4+2|^I*Xxw=b@5RiX{%-k9wV>C=3gOmY5b-8+W@TgREnaHVFA)c4vKjND zx*P%Nu|nh}Fn3c`2zO?Dpv*r037f2}5zw89`mZ7IeSu(H65A1~8TW_mb^ZC|Ph5O- z3|fFZRt!#(eF@+1m4ByuKm?yUR6bTT^2K7RQ~??cMFOQD`e^GE`>+ub{Z6>hn~YOq z%=PER1c-=(D{C(mCmH`pjU_A5cu~1KRLzOh)V!mtoeWXtZ9GbirNK3ygOnzFo(Q_H zE=eJf4JQ}2OuR4g&~NCno4A=#&}ac7l!S&OLF((raHgV7?Dt>$gWzf_(_|WaY#Uco3Z3HDel`Z{to?`Ei^yIvr&%_7l-T#Db{tV` zN}73SQ7S)v>Lj?bk75!-Tnt|!rkyN#TwYkHl%!pilLQm%)g3(a6*{+}9o*#2C{;#N zT=i__m!`4B1a!~6r2TN|=@2z+Ja~BIJa034qbF*`$4G6on^J5cTi|iUwEqd-`mQt~ zyj+zn@X-}#Uu#Tay3F&-#64!7u-)3Zk$zaVGrtG>_H2Y5RJnRTU>Xbp5Du2x%T;Cb z&1P{5ZfYvGt$a33=je4!nK>$Ix#lfytCI|d2@jM#s)UU%Vf85^N7ng#a_sBEf{A@y z>!G?YZ}?gt>Bc=!%Q}N6CEh?)<{bVOW^P?cuJBv;S?O32Y(^7p$*c2Ut9QpuZgBp3 zt}G&BHi6CK#k$JDK%PC96$bCA-UF9ULgoqXZM(ghjtr=$$gmE6hf-zwAeLtT#jEat ze(h@`f+?SaA}@+&@%o2+(r$0QcENPu#j3N_8l=18k=VYYA#0q@+KNn?`h@5C26dXG z&x|XWLmBNot559~8V%-%e=|5n3pf~lnT)DN+yAJY=C%I>c_IQ2mlLp_>W$&dy1oU1 zL9ipsc%FtSwr_V^vo;6eM zk{T4u91LZ&x$!@D!b|=2kG0>}L*m9uwVO6RBivK$hEeT}80shNzMai0jHI5&*7wuq zX$y{eyX*#MqDk;Wzg_rUUUm_)K(uH5N5!j1UDJABW@;2|WWBonUEzV!8p zc{voekDdCK+D_A5jWo4#SlhrJqTWd48}(=~Mm{rD(t9^XIP=cXyMI24nTkKCk4QGB zEAoU+MF(Xbn8M}gPYd4&w8SmhlOxYF9R)E;OOOUA6-Lx2irt4UlEuht51~C@orSbw zEp0H_YvX=QhTY))vj4&|0oSbhOD%?bl46t2?ZBOcgmO-l+!jb2jEq<69M+2=c z=>`#Rh_%B>iNQ(4*u$5FknC?Vhh)eARfZ2X+mP%N1!_)T-P6AUjr0r$r{Gt(ZwxOp zL>vXVAMwY;4!a!{ePaOBi^t1`DR%5-XOOU_UV-_Bu4OXbNd1K|EbE~ zcV9u3_X$!RXY#C;{k~K1-q%ogqnuhIUNgS6FeL{a1L+SL5Dv;yVN0$vHw^T}f@={h zdo9^#dE#Nii77PNfdr-XxG;%jE!!qPma4*z4d0jF$uZ&fV^;SJxE-U^RT*btm2l;V zA1983MYsPNZTu!j!XFY6_%ff~%(}nBs6HH&d$-<|qhl`Jp|H0OVh0}=Jl*jsnGK0w zCGySE4Zy3haDTfAb+dUR!rN%2`kop>CflPJDFZ<#0ow@`TsgFlZI*i;iU=SE5d-uj zRCo!;YTXb=l5n%J&y)fK+?}`;+g-%m&jBn`+Jr#iTN9mF$oV}~YYuN9*(7?9YI7=! zR>#{tULr3BRb=<-d`o!t&=wk9MAenJds|lq2z+rD$&=%fp1Rmc)8GV!jDI@*O^uPL zONGJ{<70q1KHRn9R&G(|KHPD#Q8k+-xyV0=AcL|H}+?$jMtCdH2k^iynl8v>4FF&3+}H55|Z=f z^>uI%$dMB7b3S_daQlY*0VjbBe&kx`rju@%lbJJ)ypnZVp{Mw)nbPRw{+P$B$*4M; ziXqF#x@<4t!$8Z)Z|s2Q0{34rcM!Hsb)~z#n@p%M#EE4`;c{Z6Mn*=0PL6HKZh=Xy zc_-$p3?&8KN4@4s2uEoz%k7g?H`|zmWwWDELJ)TT-HL~cIT`0>rxDMVO=Lr8MO+O} zk_5S>!h6T9F@nw#iZqMOWPYDCL##diRvfIy3mD{JRP>Jd>Au-Uo-;QXEo7155;BP7 zS5&r6v7@%SupkmEY>c=9+9tRKtQ?7BRo;PO;9-lU7y@12HnL+--@}6xl&pjW?Lk~? zgR`rIjbeQ=LP)OT=RdklS&9*@BG%>GW7ayxn1AdtV+UCwE-!h4o2jb)1W#oE0_^~- zlK(G&hx+mzok+(GnwQ5zF!gf zt_;fvZots8pc;*RM+n&kj%(X=N8r{m9yvA9lRp1!$I;mn1|j$#L!6}d`_7jCTTxV7 zq?qP1I|k(rBO48(qE6Po!Y|7#rBnZAe%%2EUOQC*+Xf=fFP zDn=T zIsQpO)we{=$3D|QU6=eml&&r^30?=TF%o%b=IBd+q)?VQ``?)ppR@w&ISwP&&*7;M zC6J=~_5ar|G{vL1K*nrzzur?U{*O%tp))24k*r=P5oFM@A>&t1x?fe#ohQ0wBTT#N zUpMcyt_B={Eo5DZHSbozQ5D(fGz0_gH^d$$`j26EA5~C+ZWv(DUk6JDstGQ`DDZ7h7&+Cf zck^wy9%`erp)o2m#`%h1$8(g?mt^cMW9<~_e4M6x~R?rL;)zUxT}=nGl=NNa_KOG3~53et^oydq(SOOwv?}K*?rC%7|7u> z88MvMbCk4|@k@1Sx!L(Y{NR&x`i6FzH_#*&U$LXU{!t~-X@eaUZ_5?r_jrT0g$B!z zM%Fsct263O5wZTRjMd5Xb`hK(@n!Pj@`fC#gQq?e%1sX$8;9dr!Ga@!{8Kv6=rdiE zj^}AFPgJGwuRB4hpJo5uTxZCv+%E-t;=b9>qV^!K0LBslIJ~ZX(x2g>hCUCB8tc^y z^>tp3h6~KHlX0jkZ?^+1SoJ?2n>MDc&p+>9c%josF$(K zVrX^a0sZL%ZyNwfaunf`MKPOWo}e|ECPwb|Tx>ChB)K@q4;ieQ-@`)7p2H{+&MMN zgGa@l6Nb>soj$jJ*z(&8^f?gR zy)kg&YqEBQpEJcD*FV8S&n7$t-_p!kJW6@I;l<)AJqhj4g6$MAUT(Yk_osi{sIW^;TJ_iq0Hk4aPT9;dn9@u$(zOs>j5 zxWRF;nUrCAazt0P5RtOhts@Eg7*th?(A;e?c{$AdJ)EbB@ZU^FXU1pBY*Hpfq05(p zog?Q5ev(d6$a`6n1cV8?;@~B>;_i>0s;RRAQd04`~Q10IEoM#o0=AH(25_q&sN<}NRV^qudzHtBfJ zbF4P7E662A;Zi}e|Dx_7xqrpoTSV+ECDkD>l-v%2+0?q&lG>|q3| z{&pvMK_2H}C)r=I6u8Q_D)K*3{F%OUobLlL#r*Plg0RH$tZF9)^d9$1o^?jbAI!YhzO(r%i5(c3R;y zh_+6%9q!n)0)OssaOhBR0V6$AYv7so3kN2Whm!UHgYhRdhpJ-+J23pEhMUN+%M+#s zUCZme5kN&UayZP_>k{!vozT7ofy#nTS08UNGM=*tw}4BMKGfL0w2TanyZyWP>PE1& zT&D8l8?ip^!r3^O_h9l1Ua2p1F<|b#b4=^7(ig)Sx7r`21!Ui>U_$|c`_=HY-yG-g z^GCyiJe)yN8}(<@8uf!T-QN<+W>ero(3^Ltfp%}KQK9`$vo&w!{9tgFjw{zcze<|g zx!QIQzEd7uogeh-MCIcd2B#8=^mwx=CY|*jvt!_qv$JmuB#$=ET7fyEJeYNsQeE?R zn!@(ekxT`eb07dT@3MVm284&k56-9+J*Fvv%}e0v#pfp}Lo`7iYvWpvOFq&-SmIIN zY{#-AoUJBbzTN09klzG8k5Y?%>Thw6l}>gSb}M?@+r?`d504|`xm6l()@=nnciwDO zK%(;UjJPcak|Kaf=R!E2LnF6~u4;}+ugZ>$W4sDz++LEJK6lxQ_Rj?=H#sgn3$eq` z@-*`9KPf7F0Hh#MAS~r<#=OtK5_Kxur=TN_MZ4?oCLT=F8Tj+E-eqe{X1dav6WsL_ zAMDDV><7QSV0&e_7QXu1{Z(cA8`UCvVLvvQ2T!Jcw5imWZZL$HpNFiemP@N#7fw8> zV6iMre7oN&F;(Fi*Z1RvmTY{w#ow9AxPWfFwLk`x^xDReRD&KOb7A0Y0u?jm3P;xP zM{><}gVg|LCy7%B*QX$uH2}{b@6Hb@mI?^Rqiou_!61p0hr&UrzjlC30F$V~t$R7h zH+#lo{=UGxW7?ZD^ACTfE94-22*G3_^HOwrDl`apl^5NV=G0~q`$$UBeJdvm^ZpSK zQsDKmDJ2QnbU!f6uRi*Vn$h}9FU-NpW zSWlFH^AbJu96!=1(A7xpaP1C_96#td9J=};D2!K3E;NcaFEJg-1Evc1U3|nlZE!vnkS>YUhmF=LwRcX_%u{svk^{ z5n%(KpqN{Ln#`sggm&~!yV#DK!r*9s&e26O?FgR4CUkxClCX9tEn(|eqk3Yc1^LTR99bGRMQ=a&7TJnFsK$76&ZR)5dHi=VuA)UP=F8SP8S}W)c}k^{@15mqf@fRHQv{t}sjaWwMzo&AV@j$T$Fj6I(I^=@fy+;7_EuPArJ!}b0c zoHk`(#>~+aY<`{pq)yAku9Nx2ycS zHDY{p9JU`%jfrMg5i45q!Xh#@cZQA>E zHp@HQ+cICb{%H5kYvUY$%M(;>Ufld6^B%;x$ri;3*q~i=zMc?M%9VeP>GsdBw%{ZE+6j#)EA^UAmL(7mAC0cZ$Q}CpkXYmFr zUhMEw%tqQi!-F$GK2?jlZEl1KB2+v6bu18%kyFlqUO59qUFJgN zWo1RxWT-ot{h$Y#BM~rc76<{0SxerYE6k3RWcvHv_2~KU&8TA1DZyX*o^<3|{!U~5b_p!CT2YbSAHv%iR*YvXy_&ouTod~O>9t*9~c8+;dG@RLwi6t85hrvJ;kXayhj%#Cg{$|9o#|=Ua$0Lh2kqDImva4~F?h~6a zcDOQF3mJR66y=hkz6gmEN;D}9GPmB|t{eB>+`iz^b{0Cuuv-1*y}Oh5vTkLbA!~v$ ze$8dlkPRGkRXOs(f8K+;2Wv;X8TwOia|&In&tf<9(N^5+>#PE#iKvNYN&rpfefojji=eHxNGz! zwQiElMxlrY%wL2L;%oPz(RT>x^C7rOI9$>RPaRb8)Gj7+%{hNNqYT)*>18XA?uy{P zZhMC@2%lYUx<1i^+X(2RKGa@;4y5SCU_Hh6**GKjsg>oR%*g?}S?g&^@r88lVf;Xs z2`e@WDx>jpECR{Vy}#P}Xyi%zAC7v_n_nv&jH5v{Ddo_=z|hoN{@h7Xv-g(icV7yi zfcd=8$9(vYkn$l48A8EZ=zn$g<)D4s$BLbWqNdgE1gQey#Hng)Qj7p*g4Y;(Of!GLkI11qRGz;}`%D>%c#A27>AEi< zS&!PU%J*?YqJ!>HePpBm8vBKuQ&WDr%Au%mCk(!pBmmKPC>1is<7H@rs9`)W^&B+w zTo?1t7m8+=BQ&G1^?Qrj{WyCM)+5K;#Ijp|so3QmQ1+o!^2KbXnNyDG$C|Kd6;s(C z6%I{YFPxtxwv~8W^1~0-`m=aUs?D9GnI5A?iBj+3pdrLw{?@+i&0^qM(D23T3rl`z zXfWP7_v}n2k3g-Ntk5x~^`IGpkUj0&(RsSMRt=b#b53@Kr5Iusd@dC#Tp{d~csSNU zcb7dVB~w}h5VJ|o-DN*D@X23#ZKec{Y4CW90u`sGBm8cV_jH|u1BdUy>h3@s1eIJ4 z!EIiD*vjV+9uXb-U}X2kr-u?M3^u`6{d9ueaR3Cmf_Te$;P_nzk9_-fy#m7uT>~O; zEflnLcuDDN=RXGJ!aOt8Hu1KSQ&h~+D1Z(Vg6qg{*7bEvTb)Gs8LI$9R$$&Ot=5MP z&}Lw#IB$k*oT@pcAziW}>PlTUj)NP7TdBGJt@_zm;Dsv!HQJgg%Nom9QFC9)9nX85 zqWqgbW3BJ^Bb)x(-T6xv)Xh~7DTuPWKlRQ`YmEk1B-%q8y*S5W)|x?A<>*FQ8X^E0 zFI^R+a$K{vR==uOg2a^b)`iV}{Rw0Rm0_O3OD8bj+lZ#w?Z+s5;d4ARK{6cqH;OFK zly~z5zS;jr*IP$b*{$KbYtbPf-GU%pN=gZ$fP}C>x>HJ|q~j$7B&AzeG}0(YcS$24 ze6)accf*-{d+*;F=Zte)WB4C~mv_!J<9Y7;y2HY_f8|x%OehhP@LLW`veu>ikS3C( zedhHv2MzV?^-^z_C_L?-Fq*D~_ppMUYL|&e_nLRF_5Fw?vF9E19=@+#CL8cVYez;K z+pQ4r{kA25Ni#G0?2gsOrinXgJSXw1*FtlcWGv-%clZYVd9m{~V_xu@xq8Mc425Sg zGws9997>QloRxz9i5kpWlG4kx~SO6Kc)}5Lju17jHLBsn>xW`tWPGKX7Glq zZ+-FmWUN|}0x?>1CQKN;JMs3rX=mFK;7j86rPeJ&$PHyx?9bX!vV!v^YwSYKlT|U< zR?PX<7mJ4w&oj+FlB-;UMvrgvtYdJ=xS1qD@04P#rbV7)vwg9e^}N)vLzn`|6OE20 zBG)jMm!P~gMoVB2flKd&Atc4S9y9zCj&O5)TJkRW8zl;B3-Mt>^l{JnX2&_2yK@OT zLZu!eoU;EKvH~P<{6k7^-7-=E-4DDK354-0jv}BJu_r$nKbxiC((S)^#9&`!jJF61 zR)i6{AWw5loFM4=VsJ$Bd$koQF%SS3&tGIv>$t2o8hnf`gz@dZfX$pD-PW$20?vb3qZ|F|i@X6+ndJdWE z;(1>jTlN!GuqlnW>71dwhvM?>-B7(-ZlOZ8lB`gEPUW$b3vlj_r)xQV6=+c6A+XHH zN~u!V8US+X@)DR$9D1xH7)Gq(e*2cSV16}DYC0FS81x1O`Td;%&P*;6A}tsx&aug* z!Awa(6{5mK`ZJFdk~QKeS1bGGcoz!14kVl*{5ES1Otzo)LIN8pHbhN@QF4-?7^RIq ze7A}LT2Sa;#Hamk!gGV;jOFyAzx&4;()UaCT2ip2%))LO*om;z89&%^V%fq9T+HR= z__F=DP0-d+a(iDrVC-O~+HOH#%J=BCjQM=sq$_xPk?LiOGM_m4h{!;b`2Dk-e;O7K zI7s6D7=|aHLxv_&_S0=t&lhSSOIC)9y+6SM+pG!LhIlq{y3XWY;(Zaj*rB-?BYesd z*>aCfjCq1C=A^^a_{mC(a`fD@8#%N)YI>KSRONk3)na?IwYSEW-6Q%))RiWxWLofN z>x*9*wESuwHSW!`9Pd^6Y>toJ2EHG>jFpN(6s9~PDdLQ2u6gwLCl}jlI5a)Be;wm! z1Tx;63=H($UIrBpCRcqT?zFR6OsCgV7(;&H_YQusZJS-xG}JK$1LEFq-!;^zCOvSux9h44{u zcuXl-qbT~$L)Tc*G&7eM`6#X=8AEdU32b9517;Pu^GrbHj+L({M2J9*au_%R(QXf5KklSm2%Y!+0gtL{qQY| z;UQ|;OqS#YtgLW;S_qX8hw=5QQM$4oVZ?`(Uk&5|?n0~Sm78U@(=K&4V3ptSrc0&Z z9G?dqK#Kii@?KAilSY>^Gr>;NOLwbNbI*3e$wb{Bo~yX=9D0km68)iKhM#PEg9SZA z?V4JNL!x$SuJLOZLpiPv4EDzvOFBKckpiqBLONXlby%)#%jviscA7^`}y4CB;8iD;|9^7uZ(W{k(Fg6A@Xg zWe8<9c)p#tIhB9=M$9F{;?(@z;Fm8#^M`lzi>j{JDKC&?rZd5lq@^395glIz;uv-8 z)6HxouFe9WUHDlI!wQV#E2iuARIZh0qW~wRT7@j`lFJL5_QS6(Lwu}ylw4d?fr7H_ z{>qT)(hp1us?(dNuX7dCBCBP@l!3}S$1bA@r|y-v@Vm2ePMeTWOc9;T$2v-(E!~6! zn>>5fkwArB_A>IM;>fI?$JFY_4NcJ(q1%-Rsoh6V`o+!V?(+GqSlQ%hTAj+xvHdzb z(SzuUlZ@NljFAdc+2i|}D@Fxe;1SoH+CF0Qf&;$6{1>$v*Exs$JYtu66(_T{Zeo5X z@*(;vlgdgjWO_AmmcVcHb2j-m00|hBNB^V>nY)5S%HmBez%B^DkJdLFD><{ zs&mV=19$A$ba4+L+ON7w&4rb=ZJm}u*9v!9`dEN_9)RuO@8%NAHayITq#~RPMjQ6@ ze|{aIxbSV~8nd8TAl0ZBTu*IL$ixGsw2uHrFka47rY{RXs!85Jqs&G8xl4HZ+`oRA zdEmxJ28IW^Y=DfS5xXYA-*uVUs1Zb}t;!zdXrB)5Row)&Hs~eN&tlw*(Zx#|#Mxfj zm!Om$+NyE;{ce&-u`!|$$j3|mma`!!HZC~yYCi5iH)IYY-{G$kUG#rh+T6iaJsz)H zKjGRqZ@FbZ=sDB}H8|WxbY)7PCVgX)uLY30zWNN8h~~lTKTMOtp8gbF_h)w>^ksiP zTshtPpB&^vNpa8j+r#4C?e|}VUx#HRt}qRqUq2%(Lj@vzaE9YTrK4dL_c%J6{zF&^ ze_C^(q8G-XRTHLyAe<1zd29S!UUxW5$(}&$!&(${@yd~|^^ezH2HN>z-RXAqe6OzQ zd`y;fsC=`}H9sF#sn?lho-HU@jhV#5FmXOcKa5ZwOj}T(P7&MD*EW0hYXpZPWo*~E zI?ifD(KP9m!k6gj^lAq_!?E$&+onr%ixEBTlV6yQGr64Xzod%7yD~p)&7Km`&eM?c z($dvH%^4#OJ;A#c?6lO!jha-c${Lo3Z!?IV?GNLSD#g9!a#9GCXE7gJi_R@RzX}_7{|J79JWqm!ya@*ysIJL2|0N$YvYz2sd@)eEKY>yt5&Mzy zmQeUzNxQKwkIUQ@v+W|@ko;oAP%S=^!5Lb4iRVit^gc}t6oULe=Tys(E($xXSVqeh z;-Xr^K>w4_FhG}rJ`Ce$O0!YZ-OXxLV#y-pk5b{<59V#b7R-r8WfmWHYG{&G*lAg9 zPpv`Ap>}CdbWC0fH3z7z`^`S5=3*1U5HJA~kAkT1Tl1Qkn=Cv@b>7MsmUB+fNt)Q{ z?yA|w*jFFBv?Fp5)Tf?)nX#IqW+J#CUY&|YqMh4`OA9*w1Ejjg-)PwTcW0ZYs=W0Y zEMKE1=?Y$!HHN+3Ko_zG9lLB@jXI8n4JDZxFAbk(G1FDv+8nnIsBbf3PC&07r8%fSO=2iv#+?xmC{w9^+(MYZ z;y-5KvCcFJl3?{*Ce3k@Ua7%9U39;=jRJ7V-ql!%#Gz-GiW(&bx$7kB*)Kzg9Bry~ zL6wHvac4#t-Kc$@qV#;loq0Y(#996}f(a)BXa}1_ITH#jQ4zp@OD6tt1qfx1eOY`s z|Iq+FG<3vvD!&G-YG>Ncf17<}CyZR=EgBLbHeCK7zxrOVOyr;FH3_;mR$EE^Jw%Ndc>L-8+?0BzV~pdoA=p%0rem82&Y~)}jUv2^^0+iJ%X{aF#`-%R3c5S!fWMP-ZBnan1I>O`(M>+W8%2poba4JJUZH* z*1RhjT2BvWJtQHV{zz8xDw;v(CS*2D7i#Sq_Ux8nKn!87hqLD(`@KM*qU7EsiE zOncr5E3im+mjRcnPLJ9ru9U}wJRgoQkRWIxf0A@EpOxvAu5vq7$BpzqFP^k1pGf$I zT?4t|D^GnLT(SuVNMhU!O){USa_#82x2UY(cap^j-6{+xVV}5*!_9iM?lrq_R?YXr zcNa@KcBX%K_tB6v#&g@=G|K#rm-p<6B`dm`(@t_ZlbR*n-?pNlF9Vj>hPog1#iwXS zG}lugHC(gV6LMPx!k0&`=`*ST1gOq>5NndK#mK3-B~T1iW`ZWIe>Rd)udb12+Y3#I z?q<$>b-++Q`DjTHqZcwccgjUYf~phy#8{n7Y0Xt=^Gwk4*J-ZV9I%t>IG@?XTHWIq z0|-2IOBBgWb=^w0=%bzAx~=!1OP`ff;14P*AYBLl$JvzD%IF_`iY?TsD(q4^M?u^b zL?GmFwkN@MA{vQOUnq>-fRp8>xX<2z7`#I?Wcdw|W{pWF4!TTL=~pK2i=$ujR) zomRmag6m=Gj_qLZo%1J#oJj$g`5~rXW-^6_%}ppQUG2)MYfB!w6{^zcT4ENvyPj1| zdZtVAI<>0H%@NU0yWLO`<0zzqv)i?HLU59F5<0Ilzpl+ieK-%QS}Vb48exW-_9epB zrtE#!MYsJNkouU;PDE3!a}ybu;4Pgn4g92mdbq!EwpvSgYR8K28-sFvV4=37_*h+!;#oB#Yx5blHd=rLoE-CU^6`~w@=ethN^ z`_3Pw4CHSl9B)q8G}DTFh`K8?Q}J8=A(n)1zz_sSo|3aiBNt&7(lAn<3Q-@f{>KI+ z|Gr$ryN1I-QK(L~Vqy&1bypV(Tb zRe2EnZ4Lq*aAH2mr3u=m6koaLR@zJ}=_lT`ImQi7p!6B{qnV_q4p|@=$^q&7kq2z-uHr8-a3Plw5R^{wO#V^SODJ!{Ts_w+Ht%3PKjX$c94l=f1KCE5 z`g{2^MZtID8R3fQs#F+Aj#od}4!2Wx7zVurb+kWo21^@M*6N_()cMc|s|=U3g(Wg0 zia_E5A^w=V`O5LDR%X_?`c>Ol)aChV#_#LTHwck(*B4fx<1Z^ttw2mO%ix^|4x$!N z3kvJ4)$&CA4Rf^Qo0Br=R>4nIBK8O6DH^6^^lQ7B%OzDVgpwH1ls&)Wqc2)bCbeYm z^(!0$QL65+W9SXxdS{>d+FpOI;q=u`<~3KHXX8LAz&WP2oZ6V@(HrBBd(&L0eRieu zcbz+wT*Lzf(P7wNypd@9yOrrP8qX;-4>hA?tdHgwnSXVR?;R<&;;T(^REcejdh#7~ zFLFSllr>ihfD~4Odf3U2A&e!C*eu=rvp4?@qx0`} zTY~BjB-jcGoY73osO%1YzAT)jk)k$_EvQdHIfQ1JsF1R+R*qr+aQ~`wTOl=&rv{ur zxJB)5-f(ck#}02Vy?UWhW$`Z?<0}%vQq3cVFaeNVj*f_-KiIlGEw)ae1rqalPffo; zuIibr1O~Kfps>u#d)nOzCufL}M(`6&V8B#B*PLHm`1C*&dLgQmZAbLR$WO###3goN zH;DQ@0=P9%DzA5MZ_3_Y_pv`goWDki@W{hgt6kyJFM1OVk#0&pm)($z4~*pg?KdUn z8u_U4wshQ&SNBmB(^~TlYcj~3Ii&3ktlS=%A2{_m&zQ)cU$DA-?=$2*vNd=4hUa;` z_`9>mRrC)V77T>C!*O4q(Z8`zY;i4a-4gyNu}BOTk|?{^a~Cgz)J# zCYLq}a+rq>*_r$;uRlMPTntdr@H+Y>oAL+*Nh?8hCk{f*K(zDqGb*dGcjsZ2ZR(E; z*W$LBb`Oh;m^5sK_egAwk!H|kdOqm5;F!1}`Fm}L3gRk%8}rEPvY`Z(^l#jmbLtnN z=ARwX4%kRV_8pQDhgeFk@2Sl&oQOl{=8!HKh{62H(3mBMCiyJX|Fp&M+qX!55VM$X(y`@x>!%u_gj4H7 z*&lXqkv8!oER>1;(uT^3BufDq>M0}pK@fLPBJg;$$Z|0Y)>ILMVx$ZrSIwp-J6GF8ItQnA4bY`2K>M+kb~`dw3D{{AMdFaZ4s zndaKX^p(;bY$Tb=gGV;NSxp|1GxLlonu)q@jcKvW9Q>0HRIfk9g3PJz2d;ea^{Cvv zQWASL^<>cp%lE?VN}1QVZL3BrYf^Fi$J>FTUfb$w!H{S8K@B$t%N+D{K9BVQK{_`i zx0Zq&oEeyQYEPMKSJPAV))U;AdLf&ZfIoh$$AUL?D*A^a`V(?_uQOK2ecJ*P#LM4r zi*TL7?KUT~^P`NI;sjS4SIy^YZi7ao#~UU(Y!mFf^DApEttQSk8wdqQ<;qm&b=b!d;SG#v%oGNfoVSwRgmy|2VvwPc>)whjB_;J@;Ks zVhdiId7q34Ve%+dpJC+llz&hoZi`+Rs86s#yax|Sh=1M!{6fJdwniXO>-B;#vTGlo z20aRb4}be5jTiVUCw(bea9V<8iix-#w$x9C%9^hWTd)uv9G*}!?JU8FPpz0u4Ywoegx&$cPpY+B9}rX?Cs)EIhC0j zIfT=Pvd?UJRPkW3TEgqho>Xhd@G0Hew~+JACBH~5N5(JvH?Zj}vHj2H{htF#=I<#A9j^kDiD3v~m{SCKg^@jh@f z|6=tGPyKzT{*__k@Mp0Y>JXvbZQk_>iQBY$n<1LH)!F}`sCg*t_;_{KQ>NV+lo{JO zJGv;gn!+R8Vp_QPhU$R6*l0+ujh`JQqv^SnDewDs*F5qH?VCE(tRqcz` zXKKIik2c9#j~67bClnwI#lG(j^}>4LD!$u0v-L%t_!AsgYqT)w28E`pb5%Hrb_UiN zOHc%-iuUh=zw5&h>LQCu-Nsl+dBSq~yq-L7yABr#C3EMyZKo@RE^B-pY-Sdn=`;(q zCbyPMnIEYnNlq-2lY#e%Se(e z&{4W`yKhE#;e8B%rT?tN^}PV<*ca(7NH~cy#!F+injcG5R0=O!wWr12&=;Yy%AU1l zeIzaP3Aj*g)NzjBF?PtNm8@@BX}R_JKRRAQ^4NGT<#b3=jdLXW(7#xpASH((d>*CX zA2BZDi+;n;x^xMfA}4W+Ohmbot!O>5La7ekN{^oivUNG1XVHkJFkGJMM^Vlz#|=HB zdwc2t8~x%^rdF)=#8=dGR~6`90fN&|=Vk+6oUe#n)NYO~BK!;(5v;oO&D!EQG^~lf zN>Xdfgg}!yk&Kit4L!F7wt2?Lh56*L^CcVE}PhcUhgvk10wk;yofemr^8rsu)mp zc;eb7@;6C~cUm^*&KraU%gb=z|(48&kq%@+Yy1j^(lQ%*}wKN`ZN z^qV01aH+)c*cSQzpd8e5ZX z9}UhO7?}$Jawl{U1RsbcRakMEQDwh-yIeB1MPvXvOtGmqvB~qaRBIUljErI zLUUA9floc2f zxQmyaZ2=5JrVp}gNz!KbFx+HWCwsEC8@Jq~vSAuFRLgFQRnoY-?$BD{XiAq@;#b?{ zB9$mtEUW&Ww!|Bg)hIfVcLf^a&HG){Rgnb02W@dB&|fu4zKyQf4aeIseh+8GiHs2r zz0ezLVleu=*zU(~lEtoGoD&!f3jb6;OH|x#qZy=gyr)kto=MysnQl<@;JnFtiJdHJ zS4tLfJPt~poE3W|%m_oh<+UG=}gsP30D%77-SF6J?r)e4?raXF&EZCvFzE zPpJ^8^W$E-sOyd%VmD<{D4s{t00WQO!e^|Re#{6`vVQ0q&MXQggDUkvF0|z zI4Rd~h^sNzA@_LT?ohn3DbjzIDRJk!ClmigZu8|gV0hUqZB*(7p4T~EerrNiUenRM z*E`QR6&vB49w%FD>P2cmXnG4NnZ5+K`}kfyK&0mcQpng0&1+6j%}B)G7I_8i+Uk=S z_BzeRjXuq;^K(d9El^Ok`lmc|tW;n39OVjLemGj3q2@M}5cRq{P^wBXnK#Axcrh@n zP)m9;F?uBG`JI^fV0D1U5IU?npw8EK?~GTi*RWNmY6~c>-w}Z#?9N@fs=s=bfofk0 z+lT$dP95$}%_VKHZ#Di6+z+N{aO*Kqhhg@v{yYk5c#%*&lB-~YAxzR%>$s|VAD2!B zpCR-4W1N-nxm@$Ggc0Bi*KT}U<*zJjg{xc!)SmqBN3_iE%|-9N%4?aj8!t{CAvwp5 z3Y6_l9RF5CoxJ}lMC|*fEnH&TeEicC4^<{_AxBBC^vA)bmq@ zGF}fX+!c;{0uRZ}KuSWLim2hd_WL!j0G&6_zMMf{PgK6DGKo_R$H9<-wsSd<#i6X{ zA*SR)Z});`q1jip>Sryh2c&&fS_0mjI+d?TiK|`e)sycZE!(^52|EJcr#CY2V=^+6 zF_+7~4OAKgVBbMGr_j4|+Z0$EzgY~WH^T~mT)f0dwc(|MxMU)iAu}#Hx7wb9p)YC+ zWH-|ANc-Cdx#1kwk5b(xp;_kR#oAMJ+U)0F_B-X@fVw$rXNEgDF7hT#7#As}+v!sF zQKtFGRmf&a5md}eELH(0^H{9dMO{GLW+9t3KtaZ|TP2|ZZp^XwUz zW?iA_WZFf+S`MIBYQDll#DntWbCTZY8>-RnN#FO!dftHe_iERSEB?!TwM=LxlTr*j zML3~?CBscMl!Tk1tl=V*-S~@3=W=2TLk}#8k%(w{FkUyPLGY!xSWPo9G~!?}`v|H$ z&nfHre&_n1_deC%&_m@_Q5`$#1kL<7 z<|Be6cB74TV>}U`e%9y+h_mG|$xdA`-(mMOZ^kT~Rq%3dRuewlPxrQWyLvXgu(AY{ zF=ssbO6p2g#)&lSwSV`0*l7Y1NqxROmlYgjmGiU)Ewv36Qr6{fr42bQS*4~qk485? z89Pfu%cyJ9O6&Y?HoD}w=v?K5`C1#`52F+q{iM1QCrM^HSr3Nje6XfywlI7bqt;v$ zgq&yp+MfZDy!rAFM{d_|fgC{R^(fc(>0B-hNddxK-csH7wbJ+xo))B(zizgT!J`%` zfDQQ(P+Z-~xKl?~cb>6dW!IXRs@~%g7S@Jq=&X>5k+!?$OTciVKC~YT{p6vBo>V!V z6ul+DcycmZ0o^I(tG4z_f$BUFkYd+&X*#T^`N#xc(gv6Yd7EI_&Z*cZ9pNBJz`Rh5{|h;>pGLUbDyRdR^+_tUzYs` zD=>|mpZ#P0?X0kDn-Jg5=@x4Qr|(7V4`3#Tehh-sP=wF=E{0^Q;UA_u63Erx8_!+M zu+Hh-h7u%JgxD&*1&b83uhDhgu|It%?_GiWApcTm>dpM|HOyayrW#b2bhE<|vSe|1<=m1G- zBvRK~>a6nQLu{BQbZ23~s8`1qgA*KdogM+CXpi@M@-tYIV1LE-JBV>C4QLMv4@i`8 zl;Tpt{VEuZSj-fnK-Im#1lM4+%us@KXWrwI3*AwS{11%&#UGhp6Lsn~=Z;f}*0ZR- zzX<$=9uU?X7?)T9(Yn9H-+-#kafltp_^6fI*Z+OPXVd*)3pmCf2fB#cS6mXUFji%w zf$6k1p^!sa;}8=HScnvP=Ti1(vBFS;zZG^(*UY3XcTJpQnq*$*PuD|XE6>-~ zhJ7dT{RZ4OdLqzCr*W^}<(%iPzN0hlL(5@2Sm#e8C7QX%EckXw#DBgzWOvm~vTK%d zGKD4%j|7RPhjsvHWI9vhp!10;==uh8mm{Dte@MKL&|~-d^a$YuBhlf+ zA0g<5)(BhdxdalBv%vYV41kClEvVlu-rE;_e4ZJSwV?9{s@dD0>8A766UgqDopdxm zrpGLP!UGfj7`SWUL&sZNf}r5VW}JJ@$yf1-n8z0m+~}~1tC~2cau5#9ZswTq5O&&B zRDHnS0wZHMBV82~!+lQ4#I$+@#B^m+WjY?=3S@0OZA+MRUdVKky*jV+!>?6zl0mB0 zQ%6<@!!1xSn&!Jf0}^FeL;{yCyZs$YWUN5tMF~0#mj8*DjCIfqhDrs=TFES&KWn@! zShfKb`5epWEMCh$^ysuT0jq#eDB`+NtRe8xs6pud44l_Dj5oLMXXEKiQK_w6xZ8{K zFcx%{2lUsMr{tkgS;ZeU_b&gk#V-F`U}U#Mh0xqET0ZSFBrzr$30#7vHsZluIae6u zJm^^0;3Ged5Em%`S1~rX`_CuCA1mP2@CS1+H62P+P3e}$Pb9RePp+C|yOb%)cHI1Q zkgya+negGJy>CY0Fg|4`k-KtslotW4)uV&Tcw3D3jPT)3vU~USfCKJp0;Col0R$Q} zjY?!(202$L0%*RA>42|-(`C`0l=)ktinPOF}TN;{<=@z$tK4}%$ z69Oe@gxB%0W^aucoK>==kEgjfnD3r%9SrocOCL1-nz@7A5lCE>+yC>O9L8$tQV@Wq zFY}qQVxe$H>p9NSHcJ`Qs-^Iu}7MpI=aNF{UW!QA03Y%iiIWR(s)s zFBquyFcR43GXZ=Ve!t<=7tXD1=xw4}uw~D9ZQQYWD>E)c)5L2TP0(AQx*nO2fZ^Y^ z90bJEC9-kK`%=gGr1jJ|Eieao--T`loFdp7{|QCkS27F%j1f*IyCpKMPRNn#!~KaFj#DiDe3)W8TXpDh!V#_Zwsb%)qRLSD^HIm z9{~jx_b)En&w$4PCk7E?>i%-lH=fj9#_LD^Zj>kTvo4ncoK-o5$%Lz`wIqa~if+;O z>8Gc!?rIzY8FkT}OnR`;0I+ElX_ja}+d7-$CF)M-drNzr((rBUrV-1cIoi|GI&>Jl zk&fl85mpTiGTGC1&1PfK_icE9?+cs(eSV?tS-%&N0pwwniUid!@zd|r^FSeXgC_}3 zc}mk$`4`4IcL>h+F(AeKs|_utcqy#rg$UI{y<;Rx2+N)WCV%VJnJAT}FC?TRug5-Y zAfiZQ%V3lkursPGq0StZaDRaCtrsrRtBe;H32-Ql@+kFfCnZjzwUnwd%HbV^;HhPf$eD`m+lEUOnQdIEAfn0jN}ez z`kKtohW-k>1xss*QtM0kfx}RNkgU}`ab)<7frG%YG!g&f|LMw4G$mh{X%2^6o}>jvx3fS?AzKc^aedAy z-5ExznahNSd;bs%V#(`?btHC0kPO0Rg3dYi41u~vr$BXNtaaT<-L1$CdR?S0>Fm`V}wJEKyl30nq z%kUC?$+%&GWCV{t4sl{L-2W;3j}hl%22Msx$vwXnBbm*h->6Okr+;|Fpc^muy;jUz zkP+-rjwJp2Sh6p$Yxrvf$$$;&L_N3AhhG6vzHfLJ?n52n^(a$(NhTh+aV&uG=L#6K z@q`QQkphuH$Di77t6w*>-48TI{l{o{IJ`ZXendW=HKlh|GAC{C712D&>y% zHNY3e@6)g6ITvcjYKJ3-6@LvQf$7%LpcSIi1)HZ<)`uyghv4q=%xT>eIdY0c(dzN! zDq;ajSwGtErkil`{HNU55wH_)XeB8pp3GdH;Tv=~YrT2keWVhBODXL5*APZh>Z)Z? z&i}`UU>iW+TH;@*(BPUe@dFh7+r-&2*)$ThLJO>5(2iVF_ z)-chBfADI3PrIxGbZNZduR@qm6k8bg=8$j?xHT&poFfu9gLnwzpAMY~AXG(nWo5ZW zN36F}dBq7I@#oup_STL6?29?GgCIgxnsZyii&uxSbx;!qZ~bIx;zTrXBrtG7S>dU! z0ZkoY@oc(XMUKSugQBQ;@LAYerOjeV_1s7f8!N`2%_9Yq`+s@#UR1nX43axUN0z6b zK*(m4E7v)1*`=@LKVqr~a=8z8*F{@+d4qMi65N$rxV_Se6u?d~qZ;Z03;d0$iGk)M zNS;a})sHg{!7zk+vt2Z)5M+(o{au@P#kvhM%X5<)BTOSj&ih zT&IgrRwO`!(u>@4A{&Ja&&r{HD zO}yKr(D)X0zMa8jF!88z*ldD<6No_3YZR{bnC?Q#6k~L{`3?TTzYe@SLW#Y6gi#PNT){sdtEpI+cgJBm;ZG@ z`RFM_(5LT&!+XOo2;9*S9fQb!q%QI4vmV}Gnv=e#po2>y&m-@=cW*K1r-??f+Y{0= zhw72P@+ zpGF(P);oIT^1MsTjl>I|+(}FxL9@8g?HQ94&>`pLdNQLOLE9FOfBLG}G1BwFF3!+P zxI%FR_*!kXTsiWD%Il5#qazx~HRzO`a9|zOSh*|?kmH`2B*PQCG0ESLwBFM0aStAP z!&%LkI30Z@;l#NT}=M0A_ znndkywfM(txUcU?tp!duy{(z9RcL*bz^U9jFz<8y8cw|#fu^Q(-F;-^FA%6oqLaw| zMrT%5W$jxKtiC~LoXI2~(f@6QQplJHf^_TOKA@zNhe`5m;!D=L4CvtD?VadP@n2=` ztZA)?gDbMg<-?f_Qg~ecYZ^{$+|>5cZ%+&lM|DNt@$2a-c7E{b(rN$V8-SPi3?+27 z-)o7PtGqn=QGmg#gyf{lF)YK+U=jZOqhgG&O-vG%9wOgP zwlEDTtq{|Yq@u^?A%-|@2kTMf;fR?qP3#BHwN6ZOJ9&2jTE)MJlLJhl9{qg#hXYTw z{y|&t0Gztd-U$wDTE?nXQ)E6kZcZwIb$C@tu#8>N-y!9``)UrD5*~fE%diC`;QM$a zpUme>0fs^|3LTM2IA3B$nF&R-&3#EgnbV5{K9?P5BJeG*W83up;8NDH(Ld>+3ovvw z(tiKGh}tR~|Fv$E0{Tx0D`a3*$TobsbMiNfq)6^Ox6rfFU(-sL`C98_vZC2-QCopg7KlWn>3jzaEMR1NIp;B8Sofi?{QEG{Vn1!n0&|Ww8A-edk_A?MPma%+z8gL z+%>xU`&iMz{lW|jDMkp6N|K4er<rRu*uBQxgn?;YO(V2Wt;o#DtGewY@0 z5nnQTFjg(fiu(5budKj4mLGv1W+zb>P!J`Gbz_1|VG2L^*b!}0%=A& zcwwV+RG4V(J9~l+Cuy(HiTm^kZO=Dbi*G}l{9C}&4@{5B*nARIbd8}A^T8>bLT${b z1ERf`;+PvZ2=y~ykd);$rb<866hE*Zgi7^2ipj-!L$UUbNBfZjZ)F?2`1W~EPQg` zy12RekC5E%cH=Xmvc}(IvME(|8SexouB{@M*TIQXVW}&E^j?G~l%`?HGS{Qa$X)p2 z5`P2~7SZ)V4i7IdyF_uR0JamKO@;UC(V&1F?WNYPloYOZ@>(>cJ(ohrHhJXmnrU)~ z^e3~Eio0&#-RGey{oVcy>(w*~xo-s}=e{Q(_bt?*Z$>E?ad)(MHS{KL@pvj5lx=g` z9sl%e3)|j!KM4>mR4Zi&z`&1rKU-&|zVlzAnN=H>VMGVf|9mYRo`9@z=WAh0t$ugn zTr_b9|4aGs#EB7BSPMsBgR1(KG$Y!5Cd$%3N$%TTng2K*na(+fMAS3qXJOKuU2}8M1`1{sMAgF}o z;=SCx=j7AKJ3rWliR19MqXgm_{s3VUxMUG=kCxgJV`8Fl=uhS(zRS>n;rwoCmSOt8 z;pD#i!kytEuU>O5a9b_aSA+RABlFKY^{g6xHjOb025g8;5Bb5$%I&mKHBashrGQ7y zJz7Ej7uDi+*9pL&|G*tOBm95bG-?Qgz5}nJ(ne7AZXOj3seKw{D??N!{K5kiHnR(C*mTB_SIIT8r!dtfc+^|k8>uW+x^)XqR zxaRB%Tp&_=M6Oovb&yk29cHVt2e8{Z1G92W4aqTra`iiJ3XQzjflgv)^R z8O8PrJAMbAE!Zrq~=Tu?wD&yjY1dJSXu)bxu#TT3`BcSOLFP)>X;(A4aCUj;GI^xcWe6x`7Imme}}e1{BE9o zL_;p-dU)h$T0SkRUh2j4WngB9PQquAyjcNNlF}aWW&^82rStPzNDKv&lQc>qqXnWj zX_Hz=Zau)PONzhS+LBJIPX9Aqv~XUg-z`Rv6bd4mrIFT8f`m1JFMo``0wEBtV7RX* z91t4v?Fr%rF=enMh)yNsCYd5Y!wIZmTR|xR=R4+KXTMN9{o*IlH8`U65fPjHyETB# z0iW&u-C?ibsa{mN0k&%sd*1tz<-DKorQZa61;j`={3}rB7U0hBl5^uYH$!=MgiQ13 zT;g97?`HlykMNUM?EeQPzTxP#^TKx+d(C<%*i&byr|bX2x8+1SO4SlZ**Z1GTz zQ{`vPk^19Ki@p?-`pyXRLL^*O_;tjW8RlHK_o&|*s;-#iU!lIiEYv$B`+={_X4|2-~rBr2O>L{0cIS-*_d zW5{P~NWCPuzii8oy>h`UT$37vy>dEl&^2+I0BFa=~rlMHJt_0Rm12f)L$Hk37X|L!4$E2cy2@djn$B5YK&dH*4c!O=Yp0kNmb%YRu3Ruo~0ct;YBl*V9QP z+qQ}&9EKoW;OARJf0^rl=!@pu!0@%E4xStC^P|w2bqvVPj+ouo2%(yYYoV62{-5M8n=h9v-mbXbgW? z!V{7Y#33~^K9Hj1>(auFiIGB4Zm2^?5Ng4auU^3*HNMy`FA2W|&G&hgDb;~%r5*(E-gJVawo@vF(v8o>kyY%y?kiE{ z_i~SKzleCt-@gA(1PV~W9KE5BX<$*?0QX>a^zVPJBbn~fO-DdehRt#$8%S)#AyUaN zPkp2gmWL8?knC|~^dLnle$we}UyfX?)DJP?M}VDnLvEPuae50!9+;hd3%dujp#ZV_ zW?wn(af9)rWG&#z%#q7x#(1=f*!vFM`p6liaRw96CIENX#$<(8=fPO@V4_gQSh4mS zKeAONkeZtdI{q<^Po&4~v>+aZM6Nr4T9BXo4~74(>u@2q`*L%{8=n7p?zL-QR!(`X z%>NgkQ7tvx-U*;hCaY``i$Q8q0=IVIc>eZub?`fPa17t%6oHWhe)zt+2S+AE2w3%3 zNP>N>f8*05;(CytG}55$bZ))`@i+o8iMWD~!SVNWa{c-b@z_o$N z6pZJpH&^GmV2^VA`;!n|?@oTP+?Ut~0``7;!~?dMu+vHSTq29|gDQ?4D6~Z4aPM z_$`o>QSY*=T0{&BJ4n#{U02`_v_)41BYV{K<&LQ~>pOqw@E?<^E)zb_;F|+%aklN$rUe zBW9KuMMbz;XDF}t#fK`vV_*eZN^ZKm5v^T{6dfyk-9|NYGG1oLGY{Q19<9wmFYsx^ z=nq$iCeQ?C2+A2m*8p;?>N9{iy>fK?#_k$PC&69sx|>j_URY?38>j>DsuoeVR_)E`mHLhN(CEZ z$?(Za*aiJzIY3IeYaLy9gF>ORhfFw4=8oS$iU1(mjsG-H)X&q%k7ZGk2imK@hdj>` z#XV0tM%AWVD?qMAImstVoR=}DALSq2)pU@J5+)=C(3_g5w~HKUeEAbkXap+=0A4bY zRHpnJ*9q$|G7^r+i+VqpcFsq0wYa=%{c!d>Ft7;pM4OqwSQnG*=TZBweyj^=*S{_4 z5>auRh9g{O-4jXJ^jX2(Kn#segIR^*)w`^LDe|$<3ZegO{2!J4ULB+(WI^!Fe_w0h>+#b6WihP| z_K;=UuH@TR$ogS8@BE@D&;#)erhcI5<+jymMiTxsMv>F@-WX7-sD4$cQ?tr|;Uqp= z%)A#ga+YV>9oZ93$(I8GpQv?oijZr19G9j>>tyBsMb}%#MHzKrzcLKnt>n-higXR# zIFurt(kWfTkRk|3NrRL_NlA!w38<8yNVk;IA?4ZlJn!?K^WpfxFaDgFecwCQUhBWE z%kk^|+TKnK%;V`BE4Sxqyuo@Mh%PU^*Df|GuJ3ZbJ0~*EanJa%ewEz=5}@!=XOG2^S4+N3vPsm4b+D zq;)g9=z}{W=m-kAB56CjkaemslNL)H$QMi%UWmwJb>dmRb;-a3KQKTTbcop7LfUCn zo9_cLQhZbK`ZiB+Db$`4u@E7Q@4qwi!LM4|sLg3=YQixv8Y!8WMv{xhOEjL|(m=O{-D0MC)To-oB@=R< zDyRCHs@HsTzDg=34|E)Q?oj(rZH=|z&;@du>9C;J;aE84wdqwx{mE~E5V8L0?>TOE ziKtkLp=dJkSxP23Pt%{c8puQXs9U&{ngu!hngt2Dmx^V3+;|faCJ9bMyk>88Mep`{ zc~$S|wEG1A@R2CiNGe!a#@QzLR&l8NGk`(<@=dFnN*9OVIgcF}O&BaGANO`Z2o)qP zUs(B#4+`GJ@3Gfq3GlfP6<$vv)Q~0+OUQuB_@9O05xMy+L7Ekj7sRVSf2uxUlzVOS zF|oF{?9s0IQDF7=O+fD5_%aa}N_h_;sPcc3eaV(f)YI2bQw&R<_~7SG9QNL71v_1# zP3pu<<3oP>KH-4=R!IipWDeiX7}lcPpVx(OJClXkyhhd#4PD_ZyXt<@ z8oS`YUS2iqi8F3OI%$L3djX&mgi+zWbiTT$y9=Dq-`J?|i>J-wAkG^=$-`;Xt-Y*W6VNTL1NX5^8D!~hNVxEEHp~gi61Lu3;d5ffwuBSkP zW#A+Wfsydy#o9trs3Q6A%}sT1EUYdxdxLum3BN^c4ujgN3T=tK%exnitK>He771ti z&nqiQlrSi$x`m>ZdQ$nkK|g$uklf9qr@I@iTcomsM_^kHbj>UK#q5Xa+y7vp6lT+b zu(RIkd$sEhq$1K+hVZl#rUkQv8N?7X6u!t7P1Yyl3_Li(3kVAO+FZMGMhfLN1g^f0 zBl9xv?6%V(eF{9^X&{Z{Nc75Uf*%A6R~KSOlNb^2RI0OY4^V_Jwu|k{02c!yvt|!e z$*NJl7YiZ@VfCfVu!1@&GIDg*;%D8eDR1mJuPAn=@wgV)cu1jMO8*u+SLPHHay%@x zI|{-&qbn@DMg&Z%MmzinFPT&Am1Wo!fR^ea@sMIJqe(-^=5(td4)ls)^+Op7RQSh= zriI|_-}4_n;p73F`~40GSNEBJ8ccC>AQYnsm$I#AZ)=i6wQ%pdN+-J(n3@imK-c^? zSQ=5}kP_G58=e%x>UpwR2V@q#VT4!LA9vq{eDrPf8{+N~DBpHAa|D*yS}h&{!perCJy0h1n$I^@9$}KATwJMR*G>n)4Gs> zc%SZRXHNp2{}bzXi>FvSzL!;E3)U>ybX<=fJxX^&J?r|0wExtvTTETsyEJm@~ zz))=m;3{<#xcioGE$`1St!bnsGWrcdlcDc?Tma1waMD4fZ0`o4~+k64K*pDZW0t^)W( z8c|~+y^O-kmoIk$=|>{ilvP#h?K@OSm~`9Dk8vX_P2YR%^wS@kb$I5gQ_iMJz4BhS zPoiPGW|Rr+Jt02RVM9duC5!6w1iPAqq6hr2;$S)A$8TD$@~d2q-cp&@c#SmfQ11A8!sYlvZZ98TV{Pk0`#%{L<1ACD|R_W{j8fKcB(B z_V?t^zzyGN6?9Xg^xQu0maMtZu?qc+539;X8iiVI=zkDz*xpbWEL*yNweZ#P3+=CY zi6txVbGcIe68!6T$Op82=KiZmZX2Everl>Eh=0%b02jmLlu_nwwL-yx1G(DWo8#y7 z<}KqIY?GwnTd;qEZl&riNnNIpa=lQr+~z2)_5V;W;#lsIL5Z5KeacV41wHInv&<=5E+^OH$+Bl2)7E@XAT%KNV)iGN;Dq7gmY8L z`Y&lxP|0dX+dya$gMaurt(^cOsR}C?o7ssP5rrHVu^|rBV$q_CMG9`sqrbYlozl}SfJ#3ZMg zM&ShseNlV)9nwLg4J28O&+gdDIS)BnK`sRAmOQ)z_x?UI1w_`be4C~QQmizmUi_I0 z;y(WspJ4K2nmQj=*9|WG!mfVu&Cgf4xd1mPWQ}KwVZ97S=hOk35$E9n}dv zp)PYRY3-SRh;j*KGAp*=IkJCNDH6J#3BhI>U5n)Yo^l3#Kq{k?(<&ae8{kZs04duc zo;nUgh07NsU;c{zou5}>X=Jrl@4oo`)&2qZFY~OHjh0u=-45vH*|v6F|z| zyYn&$bI)!kin+~?PRKrZ@BmzXWeiWyT8>D+uLFaFvE2`1p6(3HPKg%Y1s6h2cX#)| z%Z5jy*DYVM$LwdwVXZ1>2H0}=&6lK7f zGmBM{Zbd~!)us~ggseNl|6<(Vj8F$c2?AM%hx9CK^d7Sro)bl?BtW@IHdi_z8EArQ zzWm9ux&9@aFSC0j;(}xY+0}F8aSL{Ln4QWo2>4*j`VSl{KF9Lu0^;)&YP7|j2ySV= z-}HbEPdr=+NqA4%^M^S%eB!eK$B?oN7tLD6vxj7paC;hi9`j1^lDP*(2Wvz6z~W^S z5aOncp_=!oA&tLf0!=NHqehILY8&(7Qu07)c_CD`!l-jyGBk2kAqsTtK1|Fus>*7( zIqpGkd?%sFOOS?|Q7ZH}bE26@ki9q(?solK-tlM*sLyXGR9@j2$XKU5S`k>Iq=zK+ z0{=b~@O@Sx+7g42=3;i?+<%EK#n9eNNtAmH%uDKA(*jza{v_r%ZheM#lb$NwJwN`X z+7nG&l#Oh18wa8OW64NO9PuKfI&VCj_c-eEofqKq^uU71m@;h-&IDN)oi*N6&Iw>K zO-9zWiU-9j#vSSWWJ#6Cb19G~9uTfT9EJG)y+scwNjY95{PI5R*LE6f#rp1dG3+Yo zyK_6GLFM`d7YqiMAmQno$A7sC5>l9|45%^jt_}e3c~}(3a@^DkM)IO&nq4}G}Gqr9B=fFy}Xz) zc#O4$eP^AbK+(M0HAGDfw?blR(T-by((TAojqYwRw88OfG>K=u{YWX0pBZ04a-Nj! zTA3a_)0Q8hDU)+2LqBYw#lph6D^dGO>Cc}(V;Q8p7Fl|oD-)E?t|rpzga8*Wr}oS8lX(QTFvyDn1h~a$-#(M%Uci2*mwr zl*H*2oI1s#RWp0PF3*X!GYR|p`^UkCFqyHS%adn^+ei)WGq&SUcUyCF^Wdjf+tnT8 zC`>TjIGGUvcRv&pnvjkv6xJnkk`j~Bq5DynVLQOGElMXUcc>8$hJdNnsBg_xh)MYx zeQ@K~OcP)?Lk__ZNweSfs}!x?bUXZfn|sGD(d>Fk75S90Q4sk^8R{Gpuoj|Aweq@k1WtD5di35nf3F_1Kwc^1h7zf#fcp>7RHLBa zO5%G5#F~jHsi^vn(|vd~Sykh0@o0-(lH+5-%ukVwCP|bKI|7Y^cn~F1;zC}~5-y1& zz%w%DL?prI$RH(l1xKy1nQXcs;4cMcpvl?UbnK>D12BZ&?rh1porXOrmp^^a7b#*p z-iMW=d;#~C(BOg>->^C+2sueY?+j2BKLU+yHpimq_v~gm*0FjzgGvP++`8?-0USs~ z5_y;CP!3#KS+!OPSFeoUq)Dh&KJf3=sW7fV)7v>*VlGSjsB<81GjO7*nn3@$vk_I< zjnmNH9&({)YD)iWzAgJ#^(aZwy>vcy)8ey8e-3TL6{csz{X04W1uD-D#N6jTvys`3 zeVOEs0Cm21hz?9%{!e5W{}z=k(wYEV5ZuUap}Ha4>rOA={Ptq&O;K=1JpGCO{{s`y zYPPvs!*r9lk}Ry2O()`{gyjDdm%u3H z)Rc1D#3R`rNgosT^7|uZL_#)=ULBYMtK=w^W1M;#*$U`ZJPSf2YWF1fvU%JxxP zu^|p@7Az?e`c;q7NV4`=IH(D{z@sz>@#q}l_q<2cu3N)>Kc#`xdxAW6i1z?`TV7}C zQ=XSTv4}WZvr^2vO@NG6hTESr60;gNIt=d8&kG})Wlk%94+g;9_S{&Q1_%v~hgEa! z6%0VSv|P6cn}Rjz{%YBpe%z$HE)KmS3|Gjd5#|^!jE_E)k0Q}4o0u@bm!TRGeNgi~ zJdioZ5H3Sj9x^Y|`a_mr{4O)1KS(0J72JXD%?GDY6F{P{F_6jy1>HZ_jlWpu_Y7BB ziXqn(3f^fGvASDN2f73aboKjIt`4>;!lcO4UafNP8uNhAc$!S;;G74lanY=)144ff zA;=~Sjss<=u8TE^O;w{>e)@}0C0n6Agxjf5C5$R_@%MSfsQB=TpT z3;gHZF{Ofn?Je^kxxw3Kb#APEsHdu~S5$T*W%QOUAvK51;iw%_E&uGd|nIFz7-+xlx@D5Vj0tD)_5 zAt--w4}+2GrP6^df!9Gi8L800ybTS?`T;lw#L>)!sy&vZP<)jrp{)RI5>uHaQyTu2KN5bJ&N?}Mm1778Vq4i4HG@XL$Ct~Y}l4M&sh{}BL z&TiEzDfB22u0zZk9ue__m0>oGRVSGZ94j|Zw>B1zlE%Wu9uJ~eSmD~eGq#vEa+QNi z$!4%OlF0RynVI?i>G$`MG!|;9FWbStX#<`;o!b;?))9byYOcNkKg0PYtYiXe2Ba$Ts+{gvybQ4^;C1-=t1!kfVQ zn)FMSj3%4z!Te3i6!``B{?T^++iv?Pcy5Bx7g2OsFHHgk2+mUJTcte$k=#gjBmI% z%Q#MI&=iVnJNIV)gCF4U?|~SW(#UNE3DkLe(nSj2_diNN40vxzLqnqwRN(5I0^Md( zRC}2FDD|K&%nvvmwAk=zaAE8x8ikN7C<0_*jaVotC=`#kCRhBdK$_@iy+-U0!@29c z`_bxViq=n=iJ!|m3c&euH4dN6)VPzlz4)m*GTk_n?Q!Xd<|}UPo(0BX)Ki*EKMV;f zPZkhUw;$aKz^qlc39a9!dM(8zz6)|R6(bio*uVVBx%JeLlkM^4`<7yPlOzi^lm1{w z`r6Y~8ggoOM8>Oc+SQ?k*s?9|~*0^726~T{=M9&);kRH#x!_Ans-jve( zh4s+G8Q(NdHpp{j>zS;LPIF63-N4_I+?j_~y=RzypYNFDjl6}Y>JF7q3g*L`IoH z63&o;sw>Cg2fj?`n)#(ggC;jKr|)kWl$Df7iSv&yx5j{J(NwK&L-{1vnWfO*7ZF+N zQMfqLl{Nqkx8hX0INg1N-r&j-3ROsmQF&3LAu8Yb1iHigpj4qg8QrTHi*%{D-dJLl zEZ5@WE2-S&wAKQCFwcNbjdrz90B8tsRRQmQUAPZug>fAja0LVe=za|=-3fRcP*VU3 z2Kyhr7e;yUlnUH?F3!DK%_19ti3^59eFg8)+VE+;PE(Im|{zD!NaOM|HGs8r)CZgT;SS1Xisqy$dUy; z2tA%>L6t8VM%-NDwLn0PLZa%1Ws!& z!31zLavE>O>riX{(MfQyc#Zk|MO;bS#R&+Va?EbO|A73U|M7(3PpQ|!b%^iUmx`F$ z*wf#W*_ST| z%{L^zYRMeEs;AlgB?F0xl%(@a*U6XoB4#Hpz2wepU`#5a!!C5cXrSzFROsWZ#$L)$uvlNf8%2o5& zht=j^w&P=ymMpOqpmo?^Vt{EG`H3*8xc{>k51zg0!o*4G=#ZM1J zOPzGU845oL2?@#ESc(4%WXHDV+oUM4f|jK-k!l8;dL@HpXo_qddI_({(98>hIb>3O z8(ByG*1^*1V9KTRFxk9Fo8#Ah?VGxfEV!}PWp15+TYI|j+>rO|-bBZ1xiqfF;e(o> zL^s_UAWTFn{%k=pDFfI=>iYYOGsuPJj7V*@XK2ZX$v+GId*v6}PYJ;h2QGIJL$8p6 z59HNwHz8dTHx}lXLprKVAy_}Uo=&}E4i$YB{F+hUG7#j9iMy(#h>) zPIO!RZQ1ZiEO`0LI%m%QQj_m)6HV?GfoF@%fPJ^n3I_H^3c%@wh418a{P*3HX2T&K zyYQ@AZ0QJ!PAfdSr2@>RGh|IgPa%Oi2YpuOnc;XvY}hZT6A?%>Qpvg=c9LahNT9lz zNnTmQ1jYAq@a2?2&S0f~IA3mI9dk08hz`^d>4?yDOLauO4nUVwtDuI_+O^NhZS~p9 zR6sfbr={g9D9(Lp%rxdo_L?uQpf0D020^fDPYlV($!JDS4ziNSV2~YmHSE^95M(9X zvU7}j7&>LKZ$lC=G_~#gz4T>|YMln7fDTaSS6=wLC&M7;DgE(o=S8e0<^&E)Y^MxgRK?xzPwmn246 zEhy{D;=FVHnm+QdW)LT{=Yo(0H<@El99P_Wcd&Zigbgtj1}^|gCZD1By~@r`-}uRwRtZpLu_rULr6O53iVlVcKU=8QQz~ z`G+u{sKBE}Q*4yNoRC9g^KIxBxYh?ZcX+L7l!{tC?bn=c*0pJWuXKq;D#wRSs=d+( zRk~G-BUU+;T2xdNAn!mkmEf)_4~2`uP+cm7kV&9(@a-|SMA%v2?sc0t_dYq&o&`aQ zeA(o_kmqg+7h)MvKH$jDN!H4nl=*WD6@|5mL!fG8WYh@V0=UhpZ)6_fnmuy^24k~w z_i{tfuG?>#aF&&H9P2}Ut2g&NI#JL<-zgkR6xLvO20U5(;^U9m^quL)B^p#&;#Mq7 zMjlUZY;q-okAzZ~KVvLn-@~W3e1kxmxZYaV;}Cox==7tGeI-It8;Wncxl+a?jL};~ zMXb91tHDukP%&D*pj~e%0$LvYK8t!LSP}xo7e%n3#_u&JO!a=oqx>dP2@34r;nD*&cDC5-fNh z;CX$y{lNThP5bg*w8*q+$>`_9tO`-57`=EuZ{@HSl-OJgbdEs< zV8qe%tMzc!zbc?S(`C;^nu7$%=ep9^SojMD;dBw}8o!Il!y=$^dT=4(rX`6AFTz)w zMhFTdY@dB&HjQ7*v0*_7*>?sJkz@?A!ZP$exc zxkS_#u&uxrpS5vz1iHV3cO5d~P5=Tu1OE7+#lETw28AJP`s0yS5G;H0#^OpSzF72) z<6=hvOjg`qF6cZyUBE;GkA}iuH+F|D*DPzORYv?9yq= z5MP4+AlcXwl$Xma7DBzR&V7DMDGiSPv}km>^MX+(i{uPcaqW5r!}>9_-hy6+0Xm7^ zPeM7pr9^E@NKgnlWWh+Lq`a+U(&)DIyGRa3+MEVF;m9et{CRsJaZv=eRSvaW|4=XBc#x-+#`DlR{8=B~TG(ZnDKH^S>Dx701Upo} zE5`f#azgM9|J8w)cQUN46hC-uoE^(`9PoSj{*n$sp5F(CBngLB@)1MUTHadyRf0$5 z@>4YqD^Kvm3^|gaKD0>>k0=7aPZ4S@pwK!XJMW${Msh@OTRl*OtR@Yf8*8MAlL7B* z&w`6LKl;yXq@efN>nfoVkgK_FaGHD`6cxv8LSn%6vG_cqhncu&&V0KcN0MH}Qz}6b zNy#>o+2|-mo#E+{*Z^tcut@TXc^$dG7UHGqX#c07K`h zcSPks#qs}p8`Z#sjdiiI_t7z6?AQst=roNwi4Fj<6m&3Z^&|B`WNQsW_chUScs;g& z>r0O(DC57Ef1f-kteb{sHnDO-kdrl5iu2i?^PS4r!?0%Y=uFjz5Ug0LXNG_9*M7Vt z7`9HLm-mYQ>4_aC!y=&Q(G{=x9n2M)77s>LJ$11e%(gkQ>w*Bs-spkH{x9>v+(GK$ zJIz8bZRw#m(fFfXesZp)H-a5!$^W-rCKYBv{oi_7IFdG>pPLsIejh)XX`{iWsP{A} zRbByWb7-_n@vC}*-s2wTikCiVJf7$!!&YBGh!Z=D5tuY5G5@!V%7Z{|>XysuOAwm5e9MZ0)`QW(Y%)d>m zNIuVQZ5jifCB*8dk&e%exqyC@Io@5!cSlyh-Fmuv{G~mTP`ca(tjuY>FDUmS-;Gd1 zu)L!B7zp@9@9<40o0~zmzK2A#?LAHbGtRVP8Gfbq9c1aU2ybQ_iNh#yCyFeFTkUF} zQIRDoGTRr-Hy8pT6|F?ybSxgWydDU^NtY2+q)(T0+x&&kyl_qb3Jb~BH>C5oswKro z;I45-juZJ7*>Qd6k9N~<7~%ob z`3RgW$>ss&Q0Z;m9}x}Ou)95nqcvS?+-lV4(Xy|9n~1sZukjD_p3%?CdNN9SE<^Ku z*zOQusP4%8i3(d{LnH|;*_mDbwC2!ij=J@_lZd5vbdRn{prZcCWVRqfv{sTXm$pXFdK^P|{m~ zgv>IMM(E^9-bGyi|H*r0o2Z(lZ@7`xKu2ui%>yO@ibfeAlOdp)025VJm_xjVzEqss z5ZqE&Zkg;(Pa#T04*n*HBgcJEyWHon{yEd-{_{S8V<(;_hT678)JJKnyDD(oU}AN~ z9ikcx%Py>bSm2KaiogN!qzno)$kW%Q7DB<5ahQ#xr3p;K`(6hjl&awIZPs#u`;S!5|e>w=X)gw|sC zpK5lbBZLfDezx0sNW9zDECg~Ty*RXTh%}K7wr>XV7`FGKI6K#>-n`<>R$n}zrr}PG zPYh!{$-7hl@{R=_J=WWhmxhoiMY-0`KoKXUL>Q(Z)Ejv|`v-kU?|e&)Jt@!<6ErkS zfB4#K#pDzOqYx7BuOcai!Nd?{J@)?G+IVqysF4yb2TAhOcu<0;U*Kn_wj{F{OvBg8M;*tPXcD6~IZ98jA(SU61v-`!1r&hTxJBagp43g)%Z7~NH6g8MFtV>Zd}X+>r5 zAz>DvqyoCa`h~D7li#@D>*)!{buhn;k>?gNYxZZh!7rmWV0!I))^)*TJEN0|+3loL z0F&vR!f;l*?oVL+q(ouf?7>TvCgZ;gP;RP3yd6^b7VOV`P`#Lb{R_e0hYhV{Z_C1+ z(kK{KdHaY5un;un+A#WAUXWNMjz+4yAIs;14voQR3J;j)Q0J1vP#BKKTT)We&R6e7 zE4w6^#gZwL&|q?9t5X&Wi_2XeeSL70zi@4%2EkIl%k}{Ap!W~5@!=TgpNA8{&536dCig&)WkU1+8suH)UGBhtd8tJa>{{S69f>1wQpIKho{e z(&9P;OmgC>q0ZsK)hLp3WKyW*KICP1khgS%iP~YLi7wE7ejQ0HXd2lKzthXl35klr zXCq1GZQfQ6kHl8G#fG&NEp!Ggi=dHCeUZ(zL4Z8i=hP~Iy+pmx2cnqbr4V{fxFN~$ z@;bFGL=Hx6g*;3=PuK31`SYcTJBdYk*~Yf>8AitlylGg-x4?DDwS z*w}Y2efNBCK_FOCe6vCkm05%a^68)PCg?)uEz552^)<&LDe$V?ORf(mPk;4n>3?HG z*pF1%QFUpu z9jwQSyS2{8_Fyy*}{1Z!fy!PS7SW&PIX0;}V4^a|+5&@rz z2Ny=Q`(7`U1!_+LQ}r;t^?h-)Lxw8w`7h7U6x0$)B+>-;qGw2M%d4m;>CJRD&ZaM7 zcJZZ!C&=W<)%?8`H*5tn18B&%{XT69>5N&t} zg257251CMX?TDdGkwiabtcd(2SFQKQYb{_Q@{6zl;&mH~I_W4KM8(_!1Wqo-&nRw@ zV;u;sRNSdtf08JS$441~LsaT$p|jE~i(B{sG}oBAI|XJ#Rh&uGC``k{=@nRT6!nAp zCmpWRPG~7mf$eQ+_>+|EFW;fy+!ra%--}-S7=ySq%~=#d%y<)cAt!4?ZJ?cy72TcI!Ix@JnnmR3?2{p{pI27;{mCgpO%~`^iY&7vz?&vDiI^7w_*Y-;iY(7#)&rQ(+5| zsA)7VzzirLC5{G1&7l1I;F;>P&P~8_)Wr48V_rO%4@zz0RFHK@Q`#h>(~jWF{72u9 z-nn^oRpeZjFJ2Ie%7SXTZJBaEe*uf)ou)7is9EU|fe5~fX){*jPf9iM#U&!-zHFrQ zrpt-OQz!-lE69~%HR>r_@RFmocls~4L}D=u?Zl6pV&)XE>Tr#Kkd;w}%+ergYZ(tvV0=L{e z9%4%C7mqm*FMo=0i>=G?K-hn~Hhg9TAY)`4Gi83p50kR=^U5_(Jw^?w4(b9(x*q~V ze8tNZX;N$LGu?VOD}hP-wf%$2$-F;7VpMP%dG9FrkC<%fT%<7M`WV}oXgKHlnR?C7 zRuzMxPXsr((UHHroiOU&OD?rkSd`Ozp~OeE@`qKlAIM|wvYC1+RFN>+E#4u*uxAcY z@DP6~z1C-tau15?j)Kok|AXezc)m4E?GK>8?DR_?8=vH{5zsiDoPPVJ{rNh%VU5e= zEf-d!ABP}vH;CuS9R0b~Hxg>Y{-@?+M+QrQTC0zasAuS=Fbg=DgXpWk@ zW~RIY`0~lP(5Kq#4zeWg4D6{eOct~TqI?V8ZMxx(g0%jF3k-okU!m5W~NG` zp+od$^Y6pExZhW0v910ug=ve7a2a_0k!U{p$tPAUu-~+~-%};91TbVA?$o)!hUaf5 z7Q{WKf0J>51R6|*!q*c;{7$PRL-T7yxi{jMf_;U5#Tqb#B`*8{S;+f_jHXoo;W5#p zqaa@J$m7t?8a+a4y&29uzX$f+kb?Jons<(yyQGSX^c>#)cld8(bnpjkpe`XaN8)}_ z@U?y0q-F4}uwBxcOt>;Hs9#P@wh}01`n8QBU%y^({y!{%I>5yKOs0s+4{H^y59cC# zphL+YA33{*KbOFrh{Q|D07lRL3&zyi3)`L%-zUfhS3RKkTOlbDTYIn5@OJ>X=qI1} z^bu}qZUlcKU>D%m%!}J!a}CdU9+lhIjz1kU`9WZzQBKOmUu5&&LBjL)gl+l>IS}vh z4qo_2!Vmz9vw=i0sj|tqA@>qC-lS#*{pUNfE^1TA52*k9QUTKuyuu~$;8vTF4!>J1lI1eXd*Y=}Qm?fMJNJ5UXIe&4%(0DZ7`4fpVFaYvZihM zQvgfq?3%ty0Qw;U7i<%#Kl^J|Nw8XOe<=9)o)@a|q!ADc8`LPth87;GcJ5UxZAb|& zX}Y2A6$LrD;b3n$)?EARqfSdz=ngk~l&V$#C28VLSLl zz%ueI#HjFqQZxoDE4a=*fBs__<(;BG45}np-+XUI=~VFlhcVtzj70yV7*l;JbU?{G zCHvf!JB@c+TSVF<#W|CQI{!nca00miW8sXSz|XVUDwEh$q{k-doIwXjwFS!mheFPd33D)!G?e}%+Hgq6-;s~0=S{Ym9+j}Bp*A# zjeX$)iCd&7<9UWD?03d#x8kbeVrP2D&x(1c=dQMEF2;)$!osS85*fcg(p}<(QW&%V z!T_uqkq-#V>_LORxdm&QVO~6gy`_9gl%(s#TWW_7H@kNE<2bOGKsA@}|6R>>WzmH_ z_!T}qW}hF5RB@eYEHOk@7XsJG(Wof`-<3*Lm!02{wnA(Wg#tnemvINFo3NS(I+TD) zSH%KL3`vI8*CgXw>!0nxM`^r*%SS!HO-&~oU*ULI>2TL+tPo2(^Inv3lan^6!XI5Q zb$mTxnBnh_IsjDp+iIT@vL01Lwy5jI=#t`nBv7aXKE_-!q1O55j5QOiqQ?D32e(Nu z06an`>{uIe)1A_w0o;YqYD8Cb$?gmmJSL5NRJ;#A%X;U@g&eMHpDjC42mOrlNNGX-AK4lgZ6vUUHROq-q=LPNsaEbuX?BOV> zUMj5s$2DNY;O^U?#%Kc_dbggu1W-a7(9FMYyGMF({6z+AUj#)R>6gl2yyWC|pEd@B z!{N-T1&u6-O;JR;;Tf_}V8=l8Rz1k2>U_2phzylfW(!x^pi-I|4#+K28ADvpHBs$s=}pQhMTG5 z37Xzo({(>?2aip@5Xx-;B%8`Y+s;9of=D?a9lc0(vwqLFeR2wUC16%qPiDV4HD<#X zqM@b+I5cudnre#X^Rf6aQm!wf>M9Yig44E-dUUuJ}-9nWkUMpauT6}-eQlLRkuRiFjdK(M)e|};0gH?7gpAX2jxiu4U zZ`3CFc=uFkwWK36Ed!n?niydqFUtwYiqxH{DrkueFz#G&@{1*8*j}X zL+L&&CdV9-f|U?}TS{4?3I3&hD;eJP)dH{rm02%E<~or1i(|W8iWkLJTwej9EfaXD zmcl?Yebem9qgx2AxJxaln^Ta5n_Uw_!Za;&cWJEV**Q*P((xL>PK zU;tAu>3-#{43<9X?g2Q?J5E;jtHqMcHedN1e!e^N#e_y8$n8QC)EsZfc2T_=MDiY=*ur(4>}xzES@wdppOifEvp z$Nj!-ZA@0sqUjvk2duMNPJjOx0Vmd z`!qdF#>dCc{#pd@V^Q?vBmZQQ{PlBt(}e6gVt6(I9QVFlVU8)>L9c-7XlAi{UtC$B z*{7QO+YuzPFhD#F+cP?Un(%w~^v8kS(ZeBG4bx!w+6c7Uo)=&)~*J88)zsNS(Kk6%En8K%KZd?G1_3OBkB8z z*+{Rj1v=Xd%JlfwDvV2<`ftBAdLOQmNc%-))x$9UOvrEd$L`{)p+!#oyqpAlbp)3d zG}tml7E{Z^{*omf3;w&j^ zEKv`w7-Du=RTD;mgVR|jQJE4l#+Lzi96u>T*nXtu(~+0mCp-G~SA1p)C|ArfU@)px z0D87Q#+>s$dTW~{krdegimzGsb(4jT>4(}q53zBcKtb_fYj3GXD;Re7@bIvRGckX( zcI9KU4k@a$6Mj&!{n2)F`q=rM^QpjdNv1w1zT0v{L_{Wi=iKl4t03g2-ZJ|&Xc5)` zy~d*V-6k?7v6t-d_q7X)ir!xnGe{;SCX%EuHOHz}R_)+U#4v>pm{%w~h(4%Knl5g=ET;Ahp%2TwdCB?%&ZJ_ao|BJ_?oVbhXSz?VT0|wBm3!V8he&?SZIm-5Caf27_J}HH> zvaoLatBTIn+TZ|ltId*=_1$~01_=ntSY(>_m|Lv?!bgo{b`zX|vztk5K6}!7UD`aq>%#|Da6nrrhU|Vb|5Hm3D+RvoM_{V*n+JW60#hDd0PKw!~u9i(3^Vp?vWXM zZxZ-e8SV|yBT#z#v3{k$ShFCU1p6!cWs#9_#7K3ic}Y?avtSuk`?oa_4tJ2@wCC2Z zNUzE$zPfix<1|`;A#&i}d#s@B+8Ku7{cB4^Lr1+dW8*dlC+=_DLIiMR;*i6@z@UGy zeWO2Gz^z}3KvJo{f{=4{b90cqvJr)J z-30P7N1(zlR2J4kKR~S3d$HR#VoS1L8%=np8;W1sB9 z}9*)Ry{!AWz=QXLpfdeChF^Mlm#4iewDAWONghRFz8pKFSKE>FzCdmeO zZ~H?{fiVTo)GS!cfbawpf$%ekP0dWyWww+2&byzWLnEEk=l4|RHV%qbCeS-N6|dKI zC!^MMeYp=AR|dLEE*S473?hRcn+Q#v0`cZQEn75Qfs~e z4uTyxT_zVjijW-uOBoi&CZg}{B}?YC#1R>AM|NDL?? ztpRL{Y#ni)kxEZ@32X#U1khfecs%TjaAGK?SoxUv2E_Xu28bCXW~@=YFOjAb>}q2| zy)RR@-22Q0z@vABF@{+=;`2L)`Cttwr&*Jp`NEb^AIGs}Wvz5Ao?z}WZkI%?aS{JL z%kmxbOzO+_Fgyd9j==kXJK0%X1S;NZf%adr`Xcdc382_Y*yQx9e?i~5L+>*(0Iwg< zN$s}wM3T9}k$=I>W`GfIX%%C&{648HTo;K=M5|;b>1XAIBk0eCDbVh>Gb^7T ze5oGojjzY8=CLP}!zE7TV;w^Pi$}t!5q^p;>}qcip`crfLi$krF-&gHD5ef3@8zSv zU&AO{=tK8y3@nAI!?_EJ1$(Ktz>0c2*-q2^xi*$UOiNt_iGNry5bt?0{+qk{*$^G0 z0sj9()_cao^@VT0F$RNR^v>udy6A1x=!Os>N}?vB_dZ51!RWmN(Fqbs^bn#2AzDK8 z=pi8y-C2JB=XuU^KIf%3*n920_PVe8y1uuvdoqjMDFr6{8`mV&I8UX%ih_u!Xv#H= zL(&0(YDm^@&F_sQA47(tVMb5&ix)Nz@&5*NkgX>E9^U@5Vq)!2q z1Ryp|!OMoqEssi(64hhQ)C~{Afx{v8Valo8>qP$@e0WgO77?+WTwIMqETrS^!{1BP ziK~sPDCF*i+2U(vE?L$C=+$IN05LZug4#WVG=}2W5wLKgDIkuNBY6^zF?jCHBm{UB zDh7?rNde)+E}Gb-I~^uxYuv)Wpqr3N;lCCUK{%P{xFmqH9oev<{9p+~`cs1y zIe|z7UTvBhcn=F5>KU=|7(#dAF9QBJH5z=tlU0z=5 zE(n3}3_3Xy^;_Hkp@hxu&SQpgz?j`Jpqs)u%KWB`iGhtR_1~lE7W; z-7rIW2ze?MF{lBWyLQPAcT{c(#ZjzvBZw!mcN=s~cVCc2(puqRNDh|d{t#|wpoDKH zvc>ZLiVozE7w9>uQxHT^AP55p6d_Wb06OLl zDDFE|g{arXK^!5-+o1R$K!DiN2cZYtG7jPRgY$qCWmduk`X#uKEvQoyJ4#SMl1Cp1 zuc8LJvehvTjP)d-LcB@v=JtJf13CBzNJ6-h>1nQ)3n(jcKM%^6vFNbwD%>(Z_v0G| z=jvPDT>s`(K zvC!gfz|p_j@Xke_jvJqD@4N)2<*rxLk|$QzLVyeefx7>&qx$XZShe6DWnXrW-xDb+ zN!f4qb^<6>^WRCFA5Zag~14Vb?T<7_UjxT;#w%RN@8|Vj5>po?5_LH!eF#&Jbl3MT2HB zTGn@SpfPI68AmGA{bQKk&;rziNSY#S*h8A=1}3~Ck@LTvjsRulbZw6#F^1@#B#JOg zENFtt+eH;QmPZFOCV9!=Tof1f6_#@&M)o?{eI0qVML_jqz;?s0`F!b@1}xphR@h*R z7TkkpmlLMzRwbx7K z+?%e}P-iqFJ^kSJVuM#Tsu41$sPOJ8Y-ydTIs{?zPkCT;bU0&9f`ycoZ|-rT?o1SDTF5_MTA;D`lGqeSearmWJo;8!8qie36{sLv z?4wH{wz3K4*#YX5!A9xmU2RVyh0ud{OE)rd@i6kp8_;UYJU*vRDH9`89xzIZw(OctdrpXZkbb`xyA@*hvpWhRPZ!qz+gy{;u z9OP>pAyE-_prZTw!Ky!rStkD@$R&RrCG1*8LNGoG3JO&qZ*YDFG=|rXV=U&KHZsg0 z3a6P~ZpM~FM1=DtAMLp2|3j~dm|FO^u_bBm59y5#F6}~^@5WD6Ihj2$fEc!`BJlCT3!+ysQOx|noX%+J5L**y?(tDDZ#UWBxbo|8k8~L>yqrkm-icOj9a~**vp@0ht+T@FPYpy6d?8PL8}X>(GVXC za$c@$eA5%{%@pK+Et3$DIDAqj&Fd2Pe|ZT28t*A#64y7p=F;l`+^Slg<=d3?kq4T; z3FFkpK6&L`8FF4EIBpdw44)U9mD4WDkNuX{DGx7Fx@xaw+6ivqDgXAR2Ho=M*?uF& z(GLB)iV;}{+@h&RY9|!(ERmu<*Ysb@P8%qo)b6H zI=h9*^TDLyIJ5H9!t1q6SMLn_4iqOVyW!QVhqsvv@8QNd=X6EA)eB#{}bgqG8|tXdSbjNy-QR!NTv417wGSs4^`}BT}16SlMrH z;W^;8ex0|{{E8wS3OfB$HF9!r*e%<6{NAhLGu3AOb>@zhN*VtVw>!GQLyF9wSCLFM$;BgIos@6D+d z>SAs5=;a@a=$(1yigo%`bX_Ux-%+8MIxW2HeI*ccxl6jaPQ?+_^oQN_9)D)bd4BBY z!)XWA22W0R87!Mu_^)tVJ5x7FZ@#(HR}@T#Z_o}x9~f|6Wq5b^Gkcc5MEKH&j662d z4-u>I$SE?EAF6G2z9T#LR-E}#;DP{Sq|~AX&?3Ad1J`H~<$~t=jlj_?zEy3Vtlbw& zxS~#@j&dq3PwLmGp}QLIL8^&eM`+EWCj)?O>o^U3s?_4d?hVUOw%q$FeWfYo`q4ZD zEvdu`4rV`rfw254NVc!9@9WeX^yO;N?Tex>Imkx-MXwNV!PH7n`P@x*Rj4#?cB0RF ztgTJ-^ZVN@i|-ay(PAZ$)rc~C6NTyny&pi8aB*@2OxG%q##vigQF-vz#V6LAATQur zaqIp-z;t6`!Pr=@9n>?a27;{SXh7Uh?^m09(5o}{AtK{{D2Tb=ZS!`9Mu@@!lb)PY zP*8XZ$iAgpx6gY~d#+r}f6bPeI&gJs?CC9At>s{51fS~R0pRbW5Q;{%6PfRB8h5G? zGKjq{NM(_D_%n*t;o_m8i*Ut}zrX+I$>j2h$lsKFdn$ZH3w}ke=*7U5&2^oV@UZW& zbQ7T=+&Q_ltbYopCs{IL8$UHrV(sxTc_UKJB&MP}G3mI3(rKUX4$(rIL(Y7;xlk`Y zgYf4Z{=@(|M+?(ky8z8zpr&7ET@Mca*t~7Nju3KuzH0nM7B5d*rS@075@^i)zTw1R z=SoI5acdT~qJf6D=l<4ZPUAY>8wcaG(tPXoz5I-9&(SU-a(HSggX`W~*To@(H?}NW zrwcitn?jK=`M=EsqT2HOL7#z3?L^WCIBo^DAUP^5no|hTh+7N!6!JPG=pbn>9-c42 zs7xLnkK4L>P51h=v^G+a&1UQ1q|BInkH>jC=jZBmq6s$hAvz`hDHT=}{y%kv z7*T@3KDrHjRfP`!37R=LhzMAM;<545;NQSn?^*7sYbdE4v-ijCgE}vrb$Y7k-ce}ppM5Va% za{VbJE?lwg-gT2}OuL&P;wy1lKo8FRt5zN`k09GB=qDZq1Rw2FKUdq|;+99?V0EJk ziC*v+h@lr%uG13QE$JmA4S-XABoNiZhJL z83ZMnf}o^QpI%r45DA;+XtAehg+B8UXGUhG#&DK!1>ny}M{9x5bBhb;O0IZbL)U(C z29%C5QhnwOo36JWE*qBzbnK*WZ=UF&0r%nFo_>OtS1aW4Cz;r^YXHDRV8Tr8;W)G5?TjwW--u%19<-SW^ zl`a7DF%0DgBHBN~vdX~bZ7?ia?%?-tw#{PIiPvt6iam5W$*i$LA~?$e=zHySHTFQr z%!;t>5s{ao+ur+`hHs``Tz%(8&9&R!vb7eLZfx^&`ffP=h^qQ)6WV)|XBH`d?HdR6vTxCeE#M4YR z(C1kD`F6p9&%pQ@yQreirqKk~=^+~J`Ck3-&xvEvVfWP;YiEVuLO{%1d_<7eMyf6^ zQWNBDG(jJN3ro**u3QKZGX|{Oz@CDJhQJ{p|x_F>bhR|GOw5 z2U%SEO|f~acaTz`5AtYGTl=#HJll}tP9xP#tI8(r@xY@^075=0-I;m$DmMWXPFXF^ zU&S(K?Pjt-Zgyg)!`jh^p$I-!MtTBl(k|S>0K@7~V%Aqmlr}MvC-nuK(!db!N4g@< z(Hj1zpSEtolwV&*Ah|seEm)SRx(8Vb54(MF0KpWa+e4@UbUp92O>!Xu|J;u+tvUkF z1!C=-w1G+9j>I;4|FunJo0I{A9VNs>A-7VRJZ-SQt+a1L*iv!O12{&$LU(k?wi1!e zgq{u%3BYpVMrF3|neB*98^9|G(Ujbx7};PEN3E{rCQwg+B4{ArkE+S2o}4_4Wb4P8 zyR82)Pkho;(I3L>bw!ogdA)3B(NnF7M1`$G-*J!7tbCJ|k}Itx8BJTH)~NwL*S!@}dS- zsN`1&w6orY!pxTzqjm!pA)nQX5u&(+$Nt;>-haQDa6iE+CbCKua*4SQ2La7L>8a;k zybr^<*G06EFmy0NtzRCK-Tj#X#1HZKL+c(imxGM=Z6Kg?%gg-50#OX5N z_Bt|Xn$VaxXsZSO*7PALT^C?ffTD-~pI8vhSdfelBkhy_WN|a^Cr zOHX#8(V4rP+}xhutE#HJMb;1mcDD_<35`L-iI3KzMO79LZ6i{A6{gEcILZ_FcUK+6 zraAol{4!IKMtBqWgG2*2C-+K9OTh6%*{L{n1~km8jb0GVDP4D~w6Fo;f1CmME?c_9 zLONK)0L($$0P;J9Q#3+ilV@@xiCIAd>`lW6l-U_1hH4u0co;Z?^>A~Iwx^6g;zW{J z(T||hfGKl;VH>ji_uoGrX*P@<(Vp?fNKyaShhzB3;efyelFtnGa-j{6B5sAG|DH+h z;@vVz<-Ym%SBVzmYc|mQ=))XTJuw4vuuS?Y(lU0gw5~*@P{$3b?Yc^Q5-AY*Jxj(K zMqt@`P67Gl5ckgDkz?kUt=Q(F=>US1l1l%Z+YP=Hmh# z*}HM6T=z{zECRzN0h@GaM26D{VTYi+I@>|(8lj{(WfHF5W*0EM4|N3DU#7$DkwYlI@mU!4q_A>*=*8H}Ke{c(%vXE< zmy9fL0F&%7@aQa1gg41ggztjM;(yogBohk5JL<}dJvbNZ0L$~O;xx3UNKHY zWoi&tYi?i$ml2OaVOa+HrQOPHhfj1WN{AF?Ina0*oDh%>4dZ~#`qo)@DMlbMb|Bq3 zgxOHQHtbR`ssGuZDRSw6zZwRqemnw#S84bd}k`4+};|G~LqEBa( z=*0>gTzuZV$d=c3UsJPeDzjn6S)-Qsdk9QA0o&@;Ay)9g8J;J__ zJ~Q-%&T?YJfD=^aTf+1DMa3uiCeUhQ*IxeMX3+2)Hq~#0@^kyXof7+j$s+&Eh^hYn z5*u67Cvtmpz>-0@wZnn#V zgD6l`7ft#0tCavG#m8+&P1u{WFJnrLf4HB9dlRI>rL70Ht4NcWfGDC1;HSpTPJn({ zMG@oz$qA#CpPdbhMCy?ieOTwIf%5Z;#9AhO7zqSks;^sU)nAy0nQQMei58h>(P7bK ztfQl>?Ck8bx0v${-_t{kp~~#rGcOGU&0jw>T|B^*8^Sx3MDAN zhUH_PiLbG*bcL4V6)S#63?T+6KS~M+Jkx6EpxeGANyj%=d#_5)+l#5Tta;V9A!JQx zE46+Nc;f9K=QVv>LRDt7GFh1+CF=39OD<*fj$%35QZ99Cdc>Qow&AIXka7`jrpz*M24I??`2s2~>l#=_& zuTxVvVOf>e#;ES-Qeeo>*5R>AkKtY@8-!@Sawew`A9_2%x zk#WBrLB8%CsWgJP|22h0-LTM6(5(HL(XdcjPCB2K+@I`nUC9>UJqAwgdvO-X%(OB?YZL}1P6vXYNxr$q$zL!-Q@a=zZ%kZi{`AibDSaUi6NQru%q!W70JfIC^&Dgy`DY9XGfm`?eKaH1;Lx6>BVc} zT9&B{0J5`dx)r63Tn?uN7Tdl}NGh}x&+JhEXp9Y6UAq*{1%MDLuGMDc<6MSLy7T@+ z2V2drDPSI}=Y`-=D=Og6p}`Q(mY3c4AbkW!y|e^J^E3Hggw(e($~rI1d8ZcddH-5c z&IRb{qe3U*I7mRa-{iIaBEeK_L-MlB-Oe`BZpZ(<54s`uU#b-e6z=E=2$r?eRMQdi)PH1L`ZgD!A$eOumMu5hfOGcNO6J5f~ ztZ{3M8FG^LAeq2-)=;Ixe<+=AcZ6ztDmH?k0(qU_prKZNd*>#rGN{-d`K$x7`YWmF ze(jK6m{#)YBk26Ig^xY*8;Od(6PG|HTjp z_zfShJU5?P^{cM_Npk+OF)>&}s zRUF;fSp%uihjv*BK6DBC;5w`{+wl@AieopS3Q95 zw5v>=P374U7b&jaHq7|`FL371s84|TH$(vNo2k6FyxwFF{%$!m5)Se?%3xiMpMkHv zEf;~WV5N*u^0S!L2EoeIf`Wo7vo8+UBP%?;d^n&@T_HP?Kj*rVaYs2c^sPAaUBLVg z%VX!&%Hieq#}(;pFgj)&Rd{8IdO~$^xYg~zKaV9zt=j!XzqbdZ4H`Gv^bG))X+LXg z$cFEm)Iy^d+J;~>{ovwa5m1r0fNTq9EQw!x{s?0F7D1$Y+yR0p>FQ2m}`0iss#Pry_DF^FgC-uT$jE_g(!~V&ygNsH|e;3mXQ$uJ37d- z7X;MeWb+T+SbTLJR_zRW^6yx3c=9rTNc=Q10;DSaYt0c+;KE!wH zBf0v$wu}33~eaaVU>s2hh=zZaDFH*v~+cVOObd`|~4S>J}nq-@WnA z?v{A#6~gib1MPij=M@NXj2LD7bWfMWJ)G)QQHH zylVfJkNtXk*O13GS*uR&>D+CV+#2T-QI-~`V!JoQIG^U?D7l-yFSvDGXc&$g3V1-4 zyMsUeL)@Po6ICMHRYGP7TbBb_Ee&Xzp^^s{oJDS_cT@R*E1-x+m-a0HbOt*-o^4Go12f@P7!3+$sPyjP32PTu}54B%4Lf8*aDd&!apgL58|&~OsiRe=c^-eP|^>&#w7cSt4D zD@1eYAovFo^n9*OL*a;@5D!8SRWOmS#!|q4Cz1|H8C;y9|5iwyws)72oLoyI7Qsw9 zgUDLsH^4p_$~y$E(vwU=ml*Qfu1O&Is}k=wKIhgk5|dDw&1)}W9Rv8aB}xDz?z1;3 z1BiQ2s9ie2+l~*545N`^>|)`xch1qvB~igoGwufb{{8!a6oguo-*==Iv7_$RQdFi- zY995V{S+N)9-Wp~M6dT_n5A}rZH7%rNvRchZmgxRu^>}f4MK06$5UrqzAz*-lX(;X z{&1hra7wX6nCj-V%L6_>zBs~r59g4b?`7S8-H=Ou{OvX0%Z+!HJyC}D{pkv1ed6c8 zf8S~zeL0@7cI&m*hJL5~2X%VI{?^zJj=nDF4=gE*d`cfVy3UNVcWSKaWc5xTQgB_M zk4D8Bs-NzB1JxTFTX%Sy9MsRRBAtd0Py4x`V6*l9+`qG_2Y5t+_sURZn&Xft6%SB1 z6Yd@SbS&qQD)Y!{32edUcg$D??Rrg{*^Y?zw+Cx!B(q!yOAzm+f@x)_p01#Ot}s({}t=vR(yNDAHl&&hMeh9Cno7$-3VZ|!SH)6VEnVxH96 z+43Us)Orp!`-aAqv^n8q*d?wvI+b3^H9ucAWIKars@_Gez@0|OoENC}CN>kmtaIUM z^(HuH@};%4edyEaT0{DU2C|r~tW!W;<4i`ge8&rijbzKLW`vs9YzV+&;n&l|z?iYx zl;>Z)9EsD)6ObpTr`uUdGY=X)qiBL6ZvgAAmY>yG$CN?bXUqk`A9TX}_YanS)$vq( z6xK(TE)qQXQZ^Mp-t_nKZ2xKc)7}B{Ccj0i zCkwSLRJY7^fz7a<(eqIviMq1N+em*F!VrV1lRz-wF!l1{ba?PCkA2S= z0z?c39gBj?TfV6k_*=hnSn7~=`omj=>I~PYjpc0|T-w;FK<9M@>qp;YDaq?(y7Ss; z_RfcNKJI;o$MQi2Gm+#bE5tMC&JN!EsSQuOHkw^{C6+TTR_=tsITTcinun?u&}wk>;y2{KSC*}T zMNuPMSWdf9TeHtgT>~2##~ki%=of(5dE#iXkdNLMIr3k$(s-WK6L2QVoB|q|te=-q z>OKjhlV zeP`M0*^E+dhE<=*)RNMn)AZ9faa0iwFiu8DhTsFq?^*%JOV6JjKZ?ZCN<4gjgGc&ULhBY!-^8qihrLWzEzXA- z+d;Qa6HAjK_X%VIZOlAdnmHv#BkwoFaHDR#4<1}e4akuphWwq|T{I5YMSI7SG^s=p z50)q98OyxiYwhalvgq_Si>DQ?=ZtITDbaHdZ4@^sIe3~PVtvGgPFls zb5M)u*6-B8IN>0mnHIxdsP?R1C>n72uyLT^@^%y0cTKEJ}D$hm(YwQ zLFbcywrUO-uUJ@_p!yBCyvj^pTJc{*z*2$@BI8B`nDosACSv!ctkewq63yH%7p&rL`rV!tfCiryQz{UW?VF+Sgj=mo)JhaRH`v5zNZr z$i{Op&84;T(cQ4bdy+V_Ox&r&TkHoz>q0xsCGXbAnhb|qG3|;6#RGBNrg{J58A-Cs zmoz@f^2<9ePyUkMWg)@LvSXcLR8sNjB;tg=%;Sd#eKCwAO$rP?WSZ}9GIOm+ta-m% zA>VUA^Bitx)bRg6PH3#fYSn@MXl}Ny5^%UqmpN#XP%cK1_$4IWlc7PZpc-jA70)6D;$bi3p{@& z^lgeY?a{#RG&Ma>-kE+&p7u)t2=!d@vjTJx7H0R4uUc|_oDG#t9Vy8PQ*iKfKCWdW zwq505W0d>oygA>-Xr~{BLvl#qMe4;sd_KT%c!OB%Cb_iKo9n&rIU&_CM}|YE)j>bS z)IJBKPAw2ZpZ29wQ@~hwIEfjv7l&*4pM*6x;L(#EQyBnghLSwAp=L*$ z_g)FzD-S#jJEriY-S#YL0u)RIMJ=HNFynznkMHjgRpS8~gumrW#98Ddk6@^ek&pR{ zLL>=}7kG#wy^9jD*u@QtM45OJoQ13HLcisbb<)BvcPp>b0eh945BZoasKfo*#OyoR zUHikKa+&#Bx6O#1CX6m0oi@6LN!}fmXi-ZK0>_`=n)jgm4sYqY9HT}tMEg$9B}RYC z@$YD&J^T7$ zMW&`yTMOGx8GNf-UA&A-aM^k3jurPArSN{Aswu|Njd+AM*9N|<; zhCE50>Mz+r^Xg-JSuG3n9B zbQskm>ovPa>b@&PaL@2#$k~|bRSJ;l`0HB}|Al^@Z(~;@-&*@hNYe3vYCqx?F&$)( z35Ru`J&BkVT5S^P=efG>(JV@)SB#)lE9HHhNG+0blV$~-{qbR|3Y3msFvW}=9-JBH z1h}xOrBnlvUt6NWQl!GXcQ+eLPvxvQRKlM{$HIGWXpyFnmdwkXdunFzn|Nns_qv33 z1q?e#=p#G-tr{q8lQz_ZmEUsGD0);t(D-mBC!cx#N2Q?@`Z749A&N!LwuEekF0NJ0 zxCg;TW59;~r}4kd1iSFA+-9e0q8noUqbFbooeZJs!8L@C4RELtN#Kx_^vNgOPCiF5 z9Jeww|9&4~^W#%D_O3v2iU-9b&OkCl?&@?UbZ7_}ma#{?JD8`&^xoHIuPiEIL-|jQ zT;6};&a`7UQa)ASGj_5)gwa9b;;(P1HCmQ%DwPqV&16T7an>Fpo=j%0%3{Ur+Wru9!xVf=WC4~G#=;?!0gw5WD>$1rT`^pE zXCjbAk^xlSB2X0$I2(pNAS<8a^*h2T>^-dUlacGk2B*lG6axkarjUajA&F|tV_7cP zdGHoAegL;6YK;`he)?w{T}sZ)X$Qs;n@S0=GvCKg+mAK8dB_qV&GexwWm3@cjq&%| zST#@vwsYnLjn2z#UlKlHkP6M${wPv|seS0Isz8(UdvK2PeKE?{yXy19FbPBN?{DejR+{y`fE~))p92bTsC3_Urr}rl+ZT< zm&9%;eUPrwR`&5nP;}>j8He>QNUL_AJoX4lJl-9H83&g<4Pk^thq1wfpwEzfEUG)* zU@=w*arT{~CZk>&#|;op;@|1lYtzq=TyJ?CRwcl;G8r8P{Qr|j zy+1=qFm{pPE`z6q>zju3b{?i;GHM^IS}bzwJ@$qz?4~}ACw}m@rk3Fw+K#Heo*mpya&@V z2Rc^hoFh)8qVcl_P^L*>{+M7Hxypd(mpkCh&>E~YLe(I`-5QRAzuHSE@@LkdP`*lX zqh3_wXn@kD+X`eCGw!PBi*CKYND6wH5O>s8Xo;pPNxVT8j~%D3Ly*OV;oKY(2*UkU^lp4E=7n& z@$UCW;3q8_uvmCgY9-#_D4VHTjQz!INmcO=LKvR!f96l@>Ne^6;N9crLiKlzbifw@ zTK#mU{s726b$Su!*l&7cg-UntCdOLa6n7Zn%@Mbs{${2I^PuV-$8NxK#meS_sC*YYEG^Qqgo!`|G^fpBGwKSlBMqv|?!STuV_dtjtTp3jX&+ zFmURkffmN{GIB}lDTW*vyDqjAEcMDaD-pDzRS`nupd!k8Y;JvhJrywjAQ|Y;Aco(v z#4_>%e?hiRa@|9(I2P3+&}a6GvMzpLLP!j;5^V+B>j|aK~ncz;d?ejESCusT}mXA!G{}^Arlgmo@r2lm^dj3-+ z|6>4ES={Z+E!Da`p9^03UM3t4lxY9x=xy+j`fDKaOmRlQRfvf}d^EU%36ce#ZIHkC zv z`re57!8)TG{*aU7o0bGGkXT`5NRY%j_gx`$FtNDK$Ih@bR? zixRiwYd(m}jjny5QIsz8IbUm|OJmeAmVE`>nRFmmKBoGV%Z3J5d-M8gKa7lufvbfg z!PSOl{L0|_s@PKln5!BhDboIS%8+OLPpo=w)pV^5Q!Z&noqu3nDv+aHn^&jNr?ryd z2FE})!buXL8D5IhD`-#WG&x`FQDD6x%G2-UCiGaZn2ygXiB21m9Uz9??nr(H~u?L}Dz8ATj8#~i=HBBbl-$(K#?w&O| zfY(q|^*zk^ll!Yz*VQl%904(@Ifcg{eZVtLxL~j2Yt&L|I^Dj^5^y#eFhV_t_P0u9 zLuqc5tlH3}BA1is##?UPt49%X_obMXt%Kt4Xn#={OmH$JWW|ef``*cgGld(8W{6Zu zJkcUr6E0p$f=z9*-l;LXqawwDB24r@w<3e6E4zKVi#SkwW`(}+NhDSNCK~(%46aYj z2a8zAHd6n+%?@BgCu&TExlFOb>{^o%9D%pd-N8P%QOHNU5miJ+#+piNl_taYPESF6 z+PEojXIhZUhMM2_^S?h}SQy%Fzq8|a(CTy>8#Wxgy&2)G_1}TWF;5OI_kmY2I}ZvT zfqQ3F7@}TR4o(}IV7%JvchVC%*|&EGJhtb$+|c>jk8NIUCXKWh_Jw`-%;I2|+nuU5 zxqHIW7gf4y9^Z-tCJ&`x%Ki(_L3hc%bx zf3uzyiCB$rFEAbyuHU&3*(UH2)7`v&d!O&~Nm!FXrT|ejX0t@zJ$%&|Bn7pJN5&&iY|O4O4<-L!NS_{| zyt(Xf+YR++H`*NUMChoa6@gj32iQZwC67>GXW+ANq-5;4p4}EOwx9peE_E zP}ZgMER&(=j?Gd&Y9i=}sDM6dI7`&RRQg}c=L7?|Pv#4cVBs(tJ8z^jtW`N^ugJ`p`I zV->W*qc(jfHH12t5aSBG73FwbjRYCD*=a|uu^@N(L1+DU$ zsK&cl%rEtXvh>BrjTAa*o=FMSb)?{xR{*N8K%79ptVxZ8`SjoZOR>%(p3B;ObxCJ1 zlZ786h5IATKym!MVm}(q!?MRB=DviK$j?@^Z3X^MtHG&R58EF<60#<)&_6Fv53?I> z%K&TQiXEkSIuKEUm0p}nX1#jLZLad@(Tph6hzPxb1)l)c~Sg*kR1zN zJdO?F{Fx2B0A!G^9no8c5))nB+w+f)*-5s%Kd7mt9vnK3{qUN(CfMZ9xeX?WIuyPA&mTk5{|-VUqVo{r&iOpUqL*8k{!F4SK&rZK zuoyrsoFaz8pAAkgTOFMDd36vFnKZq(g+vd$V~FC-lX%L#e>3=wJ>-?+*gaYn*aLrk zMRIh@c_b&wBG2+vzU6RJV&cpt?xJ%fk*=v}6BjM^oB|22Z_&;J2?bG~C#Q&@kPegKzKJYHR?5GDMEtbIK zMWMSZhq2bZslNNp7B-h*IxN#a(6`ygsge~z7iN8(ts`Oy={mbq<}Xf~0_UA(3$Czf zMsI#*v-Eo?jY-dzxVbUW@N=;v=xseYNK_O*rPMX8UxZbksEJxjx}u(Aaii#?BtS_R zOAp~oMdGMiUA2pWfPTh-vQ^;GE+1%bnr@Lyw9dD9-fFaIQu!zvG>UfS(hzV$$OEN+ zhW8$b*RPFbp!r1+=Ob%=7vFpeA{!tDA|BxsIUrCpG{PvAahxyjVU)xqS5ih~6LXi! zE{c$BgISXs@0Hud6}$kD3H~jPmr$HBRL;E^+)4T2O9_C}Yh`8w1Yr7eIzw zZav46sX({0s=6wfaAH_3LJ*Z$P3~ZzcRIrdw&9iEk|JL^SLu`cMzk?ebRKh~uoYKl z=quX12NPab|InAFT2xRXdTWcGR&>PO0Tl*Ad z-_q&EILgPyM-!BwsEpv#t%{dx$iBs?TF2|{LhFQe!E`7&drUT}GDUa^v`O$|KCl^v1pS)|j@7!$t#uiE0`SIZ zZ2*9g$o|{SH@g++cA3w;)<+F*S#BpHwm!7pHm$cf_{vknXUM+Y8-9ab#dz~ro)t@FJt zv5i-_$$~OoByIns_t>l*a0&1r8x+kI3sm<;K`dc>L_%`pIB7{PjD&ApRk9|Qs^LMG z*Nnc5&$h9%tN6VQ224D4FyFeJj|qoRmtQs16+Fr&-U$jXvbf(<=R z*xn|9mKbFVTUmzM)SYBjuGQoi5`{+cFl1x4fN85&r0V#o#%NFI)pOGw#=fcJL?j-Y zR=V&Tpy!Z{+Bl1_0T@t5K&FhWV|~8o(I>MxSU2&12Lu5= zHXs!l7u9Y#^Pe7B$p#m~*la{RhFRO?Nx6|u|J10P8%S0W>efeOzQ^KwcL+?{P8_Q^ z&z08=ea*aR(JhtG9IK>OX_WS*3hG?Adzgwn{4tcHV~KXtFH#g#y7-gkz4=+0!+1>w zWJ=I???%WI7fAQ>U?Hg@%$RDA-6bLMote`)R>RYUOX&cIjIU_7$HHSFm^}+EVaKMZaE>ekhuqSAsoCR)8WMOYAFo zKYd;rPeItawbqc$99W6|>?^eW(?llTknvwC9)0O-Of6(_-+t0rwvn$l2|zniLL{GL z7v)4;i~3^;k^sMW5j44JsGT&s6yEUL?bPg*Z{q0kClj?ctg|gccHq!&)M*hgMW05iR)<&S zbFN!}A3u{qNllJn=VOY9^WOS`uKEp4LB!InRru{Moc8HXxqsq)=a>~qHptRe51?fx zDiJ4;whFb-S>=$rv!~{a;g^r$Lb{^L_dkHH_(V%QrC6ZzKlJ1eZ@$=Lfq zAatUO4fSbEtrY9Qf%b41eE`}4M?shK0f`okT7Bh=c)s+wd2Vr2K&hiS_+Lv?5{-NZ zjTaU#;ea#w9$g-IZFI{9pzQ7PWt$j?cV zyA~8xB?yP;XY|HOV-oITU*Ck}4ecF!eq^fvO?71>l?|QN`{Op>gssHu78FrSB*L0r~{c#!bQ~e%?u2v#2t9u@u^1v%w(G<9q*u=Q-iYCoH=Sw{rar z{m3OWbzz!;u|=pz(tElHCcy?Z($&_O*>=<2|A()yj*6;X+kYKFLK>8Yp}RvmhekS; zP(lO&>4xpl-H1qvv;qPm4GIWDO9?0;Fo<+Z2>5%(_dVx3=lj zI;xjISyyACUs&yqY*bfV7oR43MUB+tMQ=3&0PP+wf)MG(7L*+JbL-0ht?Ytc7X8^C z;0de|G5$_;WGOrXf-QAmY$|Gi*cZ@e)W(?)I=IdE03|kU{1$^>l+z-4vpmbgtAnzt zV(`OB%it%=tT_EbVQyfZ3WXy_03-!rXe?r(3fBfzDI%Pi)Xs_e=zDFLk?17AK4^~a zS!jWD37O#bXw4>V)>?o2aUqsoTM_&&I=dN({n64#w)km`=<#?w>?^bi&PHqx+}*%F zedgW}_csL11tWc~zRMPIo%U2&g7yB3i>@+8BNm47$YuH?qp}zUWr(2x)4Se6V1;o~ z`fF4*V+*w<<~o40?_)%Q+oiLoH7Lt9CO|~BU@1|7Z9Pnzo$?keLz^L;a2z}e4ESIv z&ATMmcvAO#J8P{dLtE9hO$i+)x@J%6N>JMZj!upU+!_FCw>GjYH~9vDyNT0 z?+W<5eaqx&NK((;Ux2G#fJKk%sP9WKW(=sQ2seOzOi2B-ku_{cANMKIcKr6dffrvz z&5Nyt>W^>1{pIXrH^5}o;qf&VJ@;$&t=@K6 z-17zXiG)`$br?BTA)-LZ3XemY;AWv#Aza(#4%DE3OPYZEAtFyO*l)4X{)I0im~Ql+ zlC|Im<-i~10%%y_HK`tKMg4?3eT!kE-DVuAaT59|E{?8jS|)13qY41%`$b!F;0W+0 zW)c^0;?3$5c&&|}?~BpuU)RWc6H}~Di$~dtbIe$lO?UAPPHAr4cjF;#17?9SLP%|D z&$94prJ1Qh$b&BLJ{I>LUV2M*%gM^FA%~&qeM<|=8Mm(KXb)oasuhrQac<`3WRsSm z&kvn9G1rrZB$^Z%OB7S+qYt)Qv3)eNE{TAXT%|Q5q8;`ayJ92t?~4QaH?EiYnkU?# zSgO0F>XEjqn6tk!0&xHoY{D_=OrL}^HgOr85g*i zKYbb!0S*VQA|;lMc0r+1VJC~OVd?FRh`5^-5&llj6sd#H#a?Mr)y&?GR+Y_&rxtdq zv05)T6TwQo3ezm0qdkaSVNO(#4L(yC&RgKktlsuruP%lzR&}34t|=Yk4=u$nJug4} z{(aD3=yz*yAxdJ(A7E5}t&P<`4?4{ksK06CY}Z3Ef4Ky8B)?!J$e37qbBF*$i13;}br^suE@qb^lSOZm{x zrl&3^z?aAl3{eWd-hS5G`(V_v?h8qVrji_sQ}(+AAGga7eU`xt9JYyNdA;qHV|93l`-!h0|JeWOou$`iNwaSq+E zVb56J<~x4XH54p`X-u})>aA9jUPUnAOg);c%>eMWNe8OId4`#;%az;C3Fy-TxO?sU za#URAcII1NE&GaD*{){z91lZtrnAINy##9_KV%~_ig}P<-nc7!wcs6`fG+TA1%VM%N923#MVe^;bvWO? z>oa4{spuJ6Nb1&Ftc<>wEF7-T_X)HD_AAyrEWBu)b{9YoEYjVPixoQ9icVokwPfWM z9}$+4_*(Plfq}uS;W_{&Ayc{G3tY2b%p0O|N>5mIB|Xr|0p1%LF=TA$)4J7K+f^dV zbfXF*4aj44u=3gsql{`m%cI~jK$OYossuXE%q)i6DZC+~lYxT=i~)>%{*n5(G8=0! z00W3i+-!UB-Ct4ZyDRhobViNPR{hl54#21G8#EX3-kW@*sEkKIU^Mk1l}$^tP3ou{ z;JY#7ACz&t&wrdXnE(PJn_-Xk()aJ|nf#FTn5Ry@6d!P>o1n3bUtancSL=EqB^j{! za98?DVedv)B%xua+C|k0#Gl!#%Mj%+i%ZqAbYs9Y9Rr%abNTvKm-l-mH4IoV^$^U6 z;cz=)t>bXO#AK}^v*tKj-nx7>DYgR&h;d0GE>~?D&mJ(@x9b_CuJu|oXbN71e6*WS zrzC64y-fKRv{>}k#vS6^cyWkEG868q^;Ld#Sh5>j^foBF$3%U;eQ#}|JZ-oSaq-OI zC3CKwCvNF1}0mb~IWC9dZO8zoO zlGbRa%F0r4-2TR)^&~y_p#+9jx%IU^6aNM`r~BPE&^;s>iyaq$F*7ASS5ctKNCu=@ zo*HH$DZmz#>H#|a{$wWQ^oiG&Ehe0H1Wim%^?rLcYg5(v6kHYvKt?XUe%+@>dEfx} zpH@q+BFfFwb7Z^?e7(Aw89pCg{5sCek@V;XMCA+$;52r(uUJeOE<^$zXb4sAj2&ud zafMN?@kL6C<*$&0hWko|(X7e^(YW9lzO~R{eL`-Ej>%3ODr|A?)1g^A=%+FICe!M( z^4##M+kv{Idw_B4qV{1qkopQ^+Ncr4{LDG_!AnG*MOY}pdu>)FS+myvaLae0KS$Ct zH6dL9+${49@!rmVZ$2dP3Q!J^=veot?Nn=I_BG@;dHjx@v}oWM5?aL4kn-YRd9soYTO? zV&&p^VG1>FY8_l`(5x$WiKvqd$vded6Nu(smJc{A`;Hx64_~G$bjH-#6g*bEvl~o1 zVRQ`g?iOHf_K=ZW89qvi%dDqN2Rl;j##(2R)O)#lc*S8I1991SslzWOpu!2K7Qvs>!$bM_d6@sIZcMq!)dBe z)}!O{gpogwwe%+nIV7OO!*?{+$r}Y;QY%sV@~9ce(_#4rPiA;EP+O%+YCYiFZ$T%q zje(5njNXg=0Rc@v;1co5r13hXK)L~&mIjDQ5|dJyDEY(6O^JaQU*D;RVdFovi+eZ% z06#Cv=N5pQeaq26lC1sHa+&$d2cT`iwLEw*l+nmkec*PStSfhl7%SDPz8>Yf5PS=I zrXU-1sP|GMnaRf}=DB+tAT2vylCM6VeYyDDVa4Xhq~}U{ zs;-1Jd3-mZrXMZZBky??|M3PApIvX*JLG&%?LrnSakl!_yq_|833T$o9ffayMv|(a zugliDq4++aW9BQakux+O@8g=ooGcWU^yQfv`(ohkw-V7z*{Hod@~n=yNS(}JN2V$Pp*@Fv=26-R5+ z$Kn-=(>WZEF?@6xx$MDS=rry9#?@Y$$l0SF`{ zUEaI5c@+r)?B0--Rc14MY^*NwDv2 zo3-aU=OWq^B)qZNsjp5meT2if?wc*cW2jl8$BH!{5ly@Zi}KQr&!FS%xqU&leRU0D zVI@|ManQ}zsQgmz^%@|j>7^ri6)cQRCUjT*J6S`csfnP&8=q~i`ZKROPrp&5Kj~~? z(1DVLnWNBlm^zI13N8g%d;iKpyc;Jm!?k|ZQlg0}BhVq(V7Y_AJWeU2-S0*J-<=|c zJ#+?jUs?>>wQo7_AF97qBlce!fikV7RO0AiH*T!9?Oq#b#X+8dT-XeSJq|X2IC8By zerdzBU>-_vRV?lbbRRaBU3i3i3E3W)vtbn+$nNZW4CZA|Zh zVvIxtOIYIFm+5C}0H>3}a_TaPQrNl;WZ2cOFH;3{pVTp-8TqGup>2y~Ps#>~2O^Z6 z66^`&5l*aPuZW-KgxLS)VyCNHy_sDH`bqD@cm)-D9_g_^TWKQ#{soJ>8Oh(TbBBS*S76O{)j~|fh_C#=bsH|MU28GP_jbr2I@2qC&nWtepKL}c%Nq#; z4}CH3=f94otm(C8vws%-Id?y7Bph~>r;1>9ns%(CeT>EJI0army8+JnW_f(AW0zr5;Hv-ZU+IW|8eAhXe-d#5P|K-% z zY1QR^JEacAsIrAOh1aRAPK0W5)YFl9f8Ufl-IurS#}q#H?21;8fPDk{U__2mJ4n1T zkd<3Z|j|yfrue*O4!mMUtCAbAl#if{OhX_isN*rJimZKSyrombV(^AybLQo|Z zXGKC`wb>KH?3aFfOJ4*|RS%KDjVWbi%#OX(u=1KY(a+yr znZb1slFxCS(mB&X_dIndioFB!?#paYg^HHNU^loBJ{Y`>3bq#tV9pB73~u4guKQyk z(7}OnH(oWaEF;|hCm7q>&D7fOW6H_-;J}wMf;Xwmf^9qIT1+21jupl5tD3#JcU3Or z+z&%=@&JL$oX6BN$f^j!pWn`7G+rhFNCH2?xyctD_01jFFFo^eSfzr%&p|;7$5$dU znndAF*hVQxf;;4=cA~+^4IZO(?u=JITI)c!CL3jNZTf^!!VPXg`x|(9udPB;RhAyB z+qvH!>FX4z9?^U_U}3`%M9nln`O5e?Htt*nO)bT~g_3fb&riaXZB(Pw^}|gi(^88a zQdMI<$%n5qk`-fVLK;RS8u(|+%-TfAP`lA6Nty-|8W{gg9Ti0R8$k`%14%F4jlTvb9KKfF|BRVr!D()E&C{t zZg@tZU|Ru}FHd>1>To|=@tC|Wv(}XO45D+a?R{UWCP8B|CC@~CuE|sXjOo8>DcuZJ zo@uo+UbcZqUlDf8NPnF@n~%jxr}*ntl=QWfwRM=Da%RlF{{1#+O^WV!MFuQP#A$B! zObZu{6xhCGH`bxKu)~=7N$!jh3l7aaX zPzk=6{{k|TyszRGsz`=si8%Dn1^Tk(I!LjZ2DMKIy>R?ZEi93#^YY?NN_R)hAUQot z=*LLHgdijQFp=D1BA#q$V-l&=x&c>y!O*R$vEPPv_HISuWB|Y&Y1sVJe zA@zk#D>_l=5)?>fS!(&vd~mpf7%-@^yf+Kze=@Q%VTAHjNm@UZ7Fw^##ix9vy*LN} zRKa4fbDp)9-p&S+%A0|$L{%W}=5xDYziWWkg0)Z368zg*mN{kGBzN15#lINjn@S_R z8g}w1ijnUgPg#Mf==OwiGaqmbne-?0Kx9Q`)je=33S0Wj3xO5=OrA4z4jAJ(jOJ4xM&5EfD)8kMWT ztk;{wmyeF|0rQ0T8vzvkd|BX6N#Le0Lo?A)G6=bQ_++_(&iqHPS{KGgrfl;R)*a_MBFF3sPf`*PL##?Ov6yh(woK%EmyMphUw zn1o?R-AsDZ&mJ$x3TY~{{PCAf*i_Jd7*ujPxNF|zz43II3y}qvbH0ON9LWm|VyhpV zv>y0_%%`T*B2_}#wY|fXAs#3@WY}kCzOsntpQ*KMzJCNE1G5R;NaQt7a$4%4=VIeE z3d0gU2cL(UHM-0a8KwDWTEo;SNZ&kj7FEH^Vf7(ljg2Ow4KS`SNdYaYBZXj~4+rJI z3XDy2U@>CkYV@N+IKfAPBFFp#Xsf0bI{T z-Y+q6_qFOP%2u-YIP59fMF_lqmMrWO3)-CWlp+8kN`E=E!LYrvPQPj=h}JxrX>@$a z{~$T+WOZ*!z^pyI9PkMDn1d_9K=XP2$4*w$VGHh!`;=+9c9cjPckI!E69H0@Gux?v zfv0>15FEO4aol=(V7D+IOf!o2NeEySmb47Y?>(0+phEq$xD5F*yhbpur<@B26MtfeYmPfHuHAd(S8tmCiKCR zD}`^^o&^oKL&9$F6>`S`i`{OQB-=BOct&}l#n2C-5= zNiLtzsGTFK-w_^DwYSm&+4d5#k?@*{yL@5wKR+mIy++p6$;Ar;JhU|~!s zBl}Lr+X>g`O=L&|cjGIT$<3y^Ni0@gm60%sKTT8|$xKr}7$5`yAjl&}Y{$@>*s zVIS;qsW&~HMg*gxxM#}%R7lKvi3$E*coaLq;_zo*r`_PU@L?Z&RGz*zw*zKBG`E*T zV!h&fB8h9|Z|)#QvUI2QJTy))q>t^#&f@U8FkDh<2RY_1LA#$6`LuN=g26;V)JCW4 z{K*i-_cstZhc4><zxfQQzQDJ#CZ8}yCjN*OvVwMickRtkoA5v zMc6Oq^&Q89*|gHCn3P109IfMcS-^5+*>2NB!Dk@aD8G*b{7tI-wGnA$rM(-6crTcf zQh(qX+&$x1*TSMJ4By~FC&(}qH4VxKxmxOjd8U3qYu_cDMX&FY^a~-T?8jGweX%tE6ZPh4z=+^4mJVjSM>Dm>+j0OE(Ck&jA{A zqMxxI+Ye;+75HA``j91fKUDK5p!KtIZ=B~6a0(F9Z?}s{g@j$sM2GS#?v_^Znl&o0 zxK?{Zw*fZI3PyXAALrgAMV5C>MKh|R0F%}@5x37yQO$1a!MEGgLN?(RZ9ZZV5bx&C zRD2l2Hnk{DMZih}3*FdwDe)x2W}^yEpGGRh427G`o6JOWsttrgV=_u|T2xYS%JDni zi4(-WnsU%}vykS&j6nZ- zIF=2o>n2rHyh8BFVpQ|VO#S!_w$Dw8O}OC-l=l(bq4bmfaUb!1ejwpfC+OyE#&QBQ zg2kV{OWiT1^NLj7aC+vb*hgv5l^3S~Ar|{U2a9+cr9wF>Xcz3cI+S5ps_RG*W5Hf% zg;rcNI|@je^4^PJ?HifWEq+3ghLt%?Fa1iR2LM`L4%xXoO&ue~OA`%`t;>z;o{~Y5 zMd9j~6~@&iw)QwVbetZYN}oG4BpI<=(u$&=NMqzX_+?}A9TNF?<%oc+YiZ(lwDSr?0*#?R5| z_kREhJuCw(B-H8k;RaJIr@;H(ud5l+`kFz+dS7EWZ#I#?V6o)d3)#vu5Qn+Cu1z4A zxv`vHE_XKA_cmf-V?A=|pbNS4hW1#t@N3o51R^!&uU9!g*X!*bSLPPHFWHX+Jd%~i zf3*Z-*5f{I5iTqPt^e){4cG?a!Sp2}yxOUov}|<*Vvr_lXS6#=tPyI~@zNbzsyaO= zfBUpmqo{_2bM{CuKtU(|EhS84OM{}PFL4G@jKhO`*7ExieMUlfg7$wz->ieM~L&~%(#`7_|1|#|M`gjY_ zQ!F@O%5oc5r~`+J;*g88I}a&sDd}WRX;*bT2)N`DmD!;*k){VjG-B_VehW8I1|;mP z<~{#)>=moFEMs0dc9*nscXLL$GWeZ$)XGKi_v5&^7EIkt1mtvNIKC|v;6I2f;9SQl zt~0jmQ5+Ce*U)zOe@)V%aC&P9Q5^eI1SHNdC3D1^aGEMKU5U5n?9mG6l7Xp&U$3M2iygAW$)q}MNmln@EwcA2=k$D zzBpP?sZQ=&yZ5&iLRdC#z0L%Uf?5DxJR7jgX2Ad_T7-ju4vDeLdDLXCHHVvgfL@$F zD%Hv~t_`Go>QUUHIog?!OBi-Y3iG^NBC6k(_jI*gzxHW+{tr@#Pyh-l<|nS1n4iCl zMX$sTM=W)Bzn^kZ1KFT6@N+H=LVrt%aA$~MKAWig`Up(%l5*}iBllOJ_6gXmK9g>h zqH-bN88c%T327`DDi4M%cxZ3scsaGT#Y<87*u^TtUGCh9;(B2(fNWF&=QF$H*oL9`(6Jt$ysL z)%5A)rN5uP|LQ>J;i;A;T^#~=3Bkp^GJ+rSdc!&WoU~K2sV%-s z0uDThj1k_M2olI|ol4#LtKL^xu?Gj1feYivQTyFP_qC$6e|HUW~vQ!Et-cB%jx}- z51Iw@7^l_D?``8)9<+(W%kuUhhN3IP*~w8f@s5Q{LdCQ`~G)8=6?EaCny zGQmoJDpT+^zMs_D5!MJwcTdkWF#CTi!RhU)^lnpb2Glj&iC?F$1I~q)o!Q#TNzAeb z7&w*9zN5Fr<%CgCRhpFN-=sP^9851LGsEImTznT7q4aHKfHLvxJY+uO>-LULqE&Bu zPB&*qG*Iv#oLv#ONF8R}VHND}@J(9BGQmsu!6p?3*o@h46>%8m-Dy9lnr-tn4WT(0 z=dpdFC+a$FrZH48nW7~GR{-kL^e5|AhAG$hnxNn9e)dPH8@H+k03gErnqwRkMoYOv zcYAJAt(rlC3$dQO>b=L>t6Yq&$H-Kb*zic_i4MolZ?iV#-&jMIZxj^G+im<`@2fAJ zv%l`LLxQEjZP4Zct;!np&cI_D-?&=L>z$(&>MuBO)-^1MCeJkQ;{yBuQDr6ngQWD- z;~cFSB{S1wh#%PPBaP|P2uRi4vHIucEa%sugQpl8%l&Mah441Ueb~2D(azmhbIA)jy zbHTj+?|j)GhiVL}^y?z;2pk*$y!nvk^kCKZbrKE=EDihf@I8;6ZCQFxK;gAp9nSSQ zLEYy{?s$?T4e0<2LLLz6cXn`keYF^fqYMC!k`OBN894M~NHX2PG_BU(Aq_*~2DefD zdk#{e*zl)(hLA~b;E!D+0n9%YvtykA;7Jz+e`Y+WqBKB{v z0_iFkXNx&v_*Gn%su*XWBhlK{mn7Vy_jPEX5U?}7=BlHE0!aVx@IW0jhMPb*U{7g%#! zyUw?Ak+bXIvg^%G6+p-^R!|^gT)-sVm%?Sd1vxDm-^}CKzP{U1W%6V zK1nYqQD!HCB{*dN5P7iLF5~B??7ARXjqM>|Eep@&-*GS0G0nYO6FShh=>LTZ!*TV7 z1L+5Mo@Cm@;Sm2!jtXSptd@&A_#ny-ODZKhxe`!)4MYk`UY|cIHUxJ#ZSGB`2fMDfdx#bw1&a3%s ze<^3(KVqnUWak7p0#cD`Nu2xsrY{vRVl-9?o0a@tYYhvKE+vXD?li%fK+<*#Ez72c zmUQ1vFE>m#^4S_QhU5dco!eGpZ*X&uhPm1IKht1wN|t1H&hcGY&wytw+70X9&1HUl?#V+<~Xb`SG6u$RXIvYG~EB0I{FqincIIJ6svXb(pYVKe0^C0 zy4P2xT(~7*CF3og3fCjfAuQa`EZClHDe9W-4SH^m7xQ2eGvr6z5ZX zZNJ;F@>osNmjxK^q6jwS^=QPLiVUQ2jWYU;)1%K%#*>=eIFdMnn-95NUOpYjYYQwg z(%?`cUq2?vkqbV1IN>zZET+ktyh=))CF}4GgvW=8dQ7OZF+DyEQpW4SXluqy-a*dh z<_)g7pp)sI(GP+Hcr|lXAaCQ5I*bV~hZw0urrsS%#}tw&0LdjZ!arW0F?(`f23f=h zwV%^4F$X9kHY4idsP(b%$CQaEimvxr^!}tbu2M5f2;v@O?9bpUJys}+36u&sZ=~Qg z(ql_p2sy6ohi^<)Qik_^5#vsS4);9-sU!^;D*RkWo=J$l`)X(NksRBMHyZ49KY%MT81G4}?bF|^=FtY9y{CM}b-QYKJXYm&;=fjMbJ{Rtni<@iG^3rJ8AHKuTNWiQo&h+og z&43|}mX^29#;WU|1$eS^qW3{h@N#s%ESXcT1(74q@Uy00FQzAvK~}1BRweLnfx_QE zo+zohK2~fQHYo$M90Ku>B#UH0K7@UQetjKc zI=I1|{?r$T(5q9x^`cJ2LYUcz;ZhT$LTbaE>R zxb77y2R98`d|bSXzk)%Z93C?}T8NZTiQ)U!G9R$ZdD`lKpb7i{>~`|nxROtC0s)Et4#Q$nS7k-~7!z7c0k zUMTgT+7^k_hbdZ)l@}N*^ed=E9&o0x!6e+r5Fh!N@fN1{Y{S@sNQ_agQ^S@=49SX1 z2I|{1woIM9!7`9e`g?%qyusC@K^`BxM_%>zj2P)&VvEZl$dwDeGn_BKgMKwTJ|dB? zD^3Bw4Gk(fb0?=s0E4I*5V4X{8XoNRr|H7@S1H4XG5gY~gSr4)o!wJj4N;u+^ROrn zRLbPsZwIE^iJ+*P*Kcv%PXY(C@MJd6xiiwqJ5LJ^-8};-MBeV|Mvs7Tj_~E`o**W9|-haK8drXIZT1IW>hx*1u z^8w?;MPsrn?J~2W%?AXdd{4j68_vk~TlbdlVXo&_HQC6oG z=4h^(yqv$4oL3-BOGb;;$)?AUxXIRb_l~=#2SKV$;1X()K~hqs>iiL|5y+I8s3)S% z)}rZJqW9=rc*d8`Z+ItxK!RwELHs*fC&hvM>=?bT_x5QS0sD2-P;&obLDUu4m^iZ( zi6FD_ghOX6ZBuUZey;4xQ5tk8!dBS6zsP45HJ02C;Nv2#kQ)f%vB)k9GyfaOoInB_ zc~HKF^+NfxDk3uarY)NQo)Y4PC9?_YF#0bh`=y2Ryk83Vs3j0G@g+bOMYPN&rE)nA z2qE6HRUy>!k_OW6cJwlpb1u!anKeTHiU0@TH$8f?{+u{sQ%LJOujZFGUL(0Y4K%JK zi(T6gT_4=mQ?zhthubLfMTTgn`%^> z`?pGkO7CzA6vi9rpQiVpKdiDCXaMOD%1i7Z!|JO3R|+MNI{x!1|N4V0iUJ>3bgb=+ z_}5wfdlQ3_5g6AQ1Yhe@xb<7U=_Re(}8zp-T{GgozlWIFU$X4MQd%Zy1&7DaKOr@B>ECXy6?1TCn0N zG->j%?@N{ob6%*iWRMS`+i(w`x?&6HQ+1nP+O!yC{78b#n!VPFJc|35R#!QITqq52 zjr!|1rqF;@z9)WjV;Td;{l_JW=EGnXlhvk>^*qq+yisQq@rgfgS{j*NqWFFa@)5|Q zanbickGW{|5_kutO8dH=0u~SJ!O#WOjNsI|}t?QhWBarr|5*7s{zXSy8^_@qP6*qtlbfq;%{iDw}nR$|{ z03^`|@)zuYfM~4BrR5$7UskY@@SV5lOXElqbs00lBO4OQg-Xu>O%(9x=;?`p6Q#>+ z7wF;bhFc{)&QJEf{aUfYA*3tnb~%e*^H}P}+lmu{XAY!@L`_s$nu@!ADKk(-5z2ce z@x!hPg`rmlvUO`DJb%Um85~RW6UlHFU}uhHa0Up273XhM6$sg$Cii`M#%s~2*9=Cv zy$TGzo6ziwq8qTIsfvywz*+^}7P~nxQB3DGH_)IfB3xQ#)|<5koi|&ysW@A9vNSG* zEy&4e9L$%u?i&MvANF5Ay9fuN-TZ+zJ$4ibPBM%DC;C6{bYT5>9mB&*QA#Ls`q-Yx z%KJFTQrhGqPoiGaXEP7tZP?BUEY9yqpw$|B^1ODxI-~iAs$7H5+#Pe^d{4$%sH$~o zR%lSeFc-+Fe`fzhV655r5gQd?6Ma4}BA)-Pz#tpI+Ty$Q^o($whxE?Hv6CW$87O|( zq=*@1)4$X^4U`OJqtP*GAg$y^!V?7O+X<$^S}07Tnsz#*ysH( zox_5AsaQLopcIwb`y0e$ej6SBa=f2#ell_b+Synu>)(B@c1@!zV5MQ;#_s~ps#(KO zB>D*eY;qC}pDjvxsanzP3<@#s`WFA?lK$S35rWuVlKOy`tF+CHYG{j;8YW_wPfcupBjlah5u;KFkPb(huil<_muv$*!lCL_etxuJjzy3NNnWo$UY`R^NHVftY5vD6vPKn>z`TnFb4~bSVx%EAAI9{s$%K zy%83X2ZMT0XLvkep>n&fI!vXz#AV}MD%+!*!3i?VxOD!yi=<^z-s>pf#vO`h1O#AC z6Zl;8RRJJZNYJhJ(U-<@RWBtnPF64fi-QUylF+MJOLq{pZi_4<_+fba;mK6u&Y$$a zB<*F2C~FWC9>ugmLGrlbc_Vu`Sq@Npc*y{_9B6ck>8XV@uQpsy9bXMtg^)_&Qm{S* z4VWgvpC3gSW1tcN;w?9(h-3d!Drl5PXUb5jGjylOnO5E@_WQs0IysU;r65``#Ubn} zeF=`}%^hH-m?F0ExvUaUcqT5d{t`1l5Ro=XAFWH0k~%1MKu$;8Y5*0CGXG3XrVRsfNY-Hk>#sPpk774NOl=6y8x%BbA7NNwXT0N8P;-u%X<3 zKvDg!lCl~hS`B27&+Fa>oqAwoiJG`6y&$dXnvI4i0Uk@_P0;+M0$-W66YAsmERhK_ zSl?_7;6I73BSOfU?%uy2dQP;Y;YwC31xA-SO?KZOi-`3Szz~edZ{Dy5p`-uAJz%=T zL?o7^G1?)eG_vN}@)Lz2K>cEiI+yX?wqSeFNc?~5&n0rqeYc%+qxj(>yAB~^$4P5Q z{I7OD{{}kg8c2yjhTx898u8(9f$zkddxOt}JWe(Bwm1rT5oyt8-{f#oVsx-#BLc=t zZf6|6aqFc{!W{=jHQgL7NnTsClPRRv)@?fOOgi0(TLWQdejOgX7ow z zeuY-10N0g8DZqg-hGExqU)cfuoHqkP3KcwxF>xY$zY_&99v-MM_5tu6{dK3q*pp}< zkR6JEE(chsVHRPS-_iEu-9GvhKD*c#o&|~d68tWA>h*iuzxukqa>T)M19{m_i9!%) z4!w0T5w#h-4(ZYVwS`6F!xqUneUAxJ*%u$MbEbQ>ayHN z4Wl?JfKa*r6Wr<$shxBH=SXIt`<=|%91>X;;f(3tjzhf?pX;Dj1TCNF-J{hhio~Vv zz8a;#Y$7p}^7y0o>8%d6{V0m$1tfJx?}wyo&^vXe)2xToWFOxMmNV#C?mPzI{s#jE zQzlDXhEyeI6jL3VfA5z6ZA6Ki4m!lrwx=dZrWBYg{&@85_AqrL>3vpVl?j5?_O@R? z**`XvC-nW-C4)yWB`>Wj6u()uZTRmZ3DZew;9h}1H}d2Dv>pNS_rkotpSQQmT|Z`( z*e9V0oHmM^Hdmd?^fzN_IG)npE-aya`Tc@)Ig#=pM;_{aa}TNIe#8G)jG6qGQ4q)Z5=TadD=E5Ct(fg8y6>+bG> z^YCm4bR-lMz#ChdUIH!X!-=w#gc13*wKa)=Lrx-kDXflL=X#$TcoTc+d%@H%&rbm4 z9*Me6P?bBD`&^7O$jQm!5)rBD4Z$TOB);vh3`qFx!iU*&_`Gl_DJjE6uPZ(~ zW>*-Of(Pj|+b9%COiOgj^cG#5WMnEon?C3jSvfeC8zMRu!J`+7smaWw1RC17AiNja zTeJ1#IyyR>*6lJDIhJqi*2tiAZe#rVP^bHA`^C{oNzhUor!6DF>0cVAX}jaE^Yi@V zbidVkZ+*FO|HXavrmWDnZ$lf?yjnaQ@i zBUT)V_r*om+t2oS$m?U$*yRMVR^A!*>PQ4^mIZe>DHtQFl7iU%gFZ__YeE09(5{gB zyU0RO(V{wZM5de7GKc6%%yIma`F$#P2^KqQd_O1%Uy z!!B^0Ylb8Of4)x8Z340xMoo$IC;9n|U_2WWm&wmup!rvYNo8z&Ji*&XU$@&8_I?2_ zxH8B(aswFh=3K*654Rr}%VD!bCcWSK+92fp#i(drryc>6@;*=B3SBHofJR)|#)cV? zo9wI?TK&oATD%)pmvj1`^2Lpp>e~KBUa&)1V9NXfOh?}vy#|bE@e%Om7a?tDKCYX4 zW`c@NKvxvQoS(6*r_&VlY7WXJ<+zs?XHul8cC-$w@2 z&1L@pyvxGE!q-Nx4NG-U>LVce6UWcb@16_#4o^eiZ^oUw&0Xl|+O8cwi`Gv?rCO`} z?B={s@Jrs6>lAD;G22`P{#|y6-=Kzy9)gc%Kf(YdoLV zbv>`E0&1?8!NHzd+2(_e(}WFIIa4Z->H7h)GW?cpF7)Dj$PQepp^~Y(Mc<3rj~z4B zyYx0@01Ltk2qP-qd4Y;ime@vFSy_b&B9VSdgDJX2`L7v?>3Fum)N<}q-7w)$LtBy% zjNltFlEA=uW@TmddW_G5B`>%MT-F4#3AG|KSLrFwINw(=cRXOGJ13S9RE3$j!wg7< zdwP00bsHo#@fv)l;_80x@1OIa>g!j0TtcFOZJWumV}X|#T~oZ|x73>*`+WCP-0%Hm zH^00rRqgG+!YUeSVA+*fY1CHNT8A)X&Gr@ySI~)np1zW$tf1>QUx<=iXHgcF(Y_58>{M}<;c3l&v4mRptJiyYCm zzO>V^ju};pt|t{`V>r{d8UMO5wV&E6H4ikML1AJ=(t_=pgeW*TO#W1MxL+@T}e%CY8Ng`JXkNseDQB;%_e5yY#V|VUqo~|BZXT$9>p{DYx zIy$Uj{FbUok_d0Me*WdzZbGj87PI)VUv4!AS_KlsaI-)=`?--*ez#pZ4sp!8_O?C6 zcHOGc;8LOS=49pj7M9nzA&eic46Ph#LERjp=WO;sYPQh6eOgmaPEN9a~6`4L)RT)5L zpY-BV@^2qLTq??BPibx`?_gh~hPyFGO5cnop3NWVJ|-A(!w!2Zy7R7^8TF`p%q_5+ z+mMdnFvdjz8R8Ef&|Mc6ikwyi+y=DJ+Xl*nXbsR?mwEb>eCU_UTX-Ug5H{=Uaf)|b zceWgY!hm--n8=J5P-}8{>^>EbIX#$KF?h48dc@o8K+qYQ?_W<{SBl0Q4EB9gvKe?@ zg`bj+*&_3r8K)!uaEf}06^A8C3N3}RR!h}=+T)ZLpcLJ5sfAa+eBa2(n8xY04`!F9 zefkU=uoxeU?XQrnRdAsTAvEF!->aq2+&g#g#ZJkBTDx3X+z1TzQQ)=p+~!8AiJe<8;o(OW}}E#j{IvBCR`w{63KoM$l>@M70O$-FkL8Sj03B-9o5GDzX2D=S$6*6*O_Crf+6k6aiT)CfU8K1G zwe)Jcd|*emb${Pjr9+U5V0v5|QrB9$=h2C#aV_2xf8OsgRD}X$mDwUX-9h0&=Cmhs zi9en!G>( z%L{lEM}k1afe!O*v!Iaq5o(&cRgJ0;1iwcU{e|)v3dz>~WUa|aMLd}~;y3jY!ayt< zcu$`>!wWfYK{Anb48;#d>-3=*LCb?og_KzZlXF@$pIgLeP<2C?TbqxYW2DjNv8}^l zaaU{?71kDOY5QGP@2sh>d}@*NJ;o7ap3u3h-}Xg?sH#ER;co`Em7@UNc}_nJyG^nIZ_6m&z{VpK{=PD$Giq$g$k|$B!WkW8cB-ci=JmZ1S=V zS!J-V6PB0WU}~o13|@yEk*d+BS(@YAsRC)93WVHlkVz-U*nCUUtvLJ4_*w9Kkd6$t zj4G~59GNCos}1O6Nj-`&&ER6A^F3NH_j9hJU-i6h^{jg^vw-(qM(?8y+M!0qy$zXs zLJ7A}+~|9TRWC?(io?|bf?fLWvuy^i;ogdWQ=spkTshj=vNe{9gy%ea@k)ktTzoU2 zePJE8h2Is$EsfgH5|bAkUKAaYyj{bg^e0YFYYM5SamGqz_5{+3V)Ahrt`WatFMk+V z?Hmk<5Bz(B?S4KzC+5ny(#fNikj)Y4#wnjwg(gC219ra=)4YgPT-_p%FL?)63~{!( zLtD^7C^|HH)|yMFn>D0{R@-k$*uK~>mcsX7-D)QxApU_xT?5}`8mhX1PL;Vq4+d46 zGre9VCXD+8k_IA`bsU!GW%4bA${MUvXB>3RnCoU2$2I45gE2Iy_-D35CS?%soRcZw zt3RWa&3$#qR@EEh#_H8l;dGIH2w~1(Ja;3MfAO&bE`^wCp~PhDHNhZ4@nEZU@Ql|? z*e2l-cZb|e|EK=7m&rk~L1K&)Ndq$ZJ+bHcv2E?w1K-`_R5n*>w4Zc6vZPfoZW*Vj zVq9QLJow;_CJuZiloy$adot@^8i$|+(s+k2&CgN$P%Cdn{Mh$z)rtJm=nEVDY82=xwLPRj>cL zs{gsOOH-sJ$$pacg`@|h+YJ~VXN3$&535 z_S6)5tgShIv}k5He*AbR$##EBYwHYaBME3QK;aKYynU;ZdjImq1`KF#uRZGP0ATgn zH3Kqe;yfcFA~tMS7Z$D}HZ~XC?^BkQmF)%na8;SdeuBW@v|&^_gbCZ&+Eo;9ZM=N* zX1yL8IYuTcE6Y!3?mr!20ro}XZi>z>&dndqZaQ5MwCZ`1lEQT(QrL$}Jx#RXtMKo~ zmE=BMVK)N~JP-?2^ZFG^4E~j@*X2WACF$hxy+VS57P`y(ZGa!TN4XuxJJpuvLHUzq zpo5}h2)H`|S5TcSG=y5)oEfE5OVjcp8CJ!BRnoB?WS9gN)FO#n;WJ7f1Y|5fuG^dT z?`09R{G5=aM3uzw1Pn6XgmP;a*L$ARi*~dE!jAqZ^RwV^PB}$Ik53_juhKaTXTBrZ zX#J-rj4B>&3eCte){VOemF|UD5t3K`N`PZuWU7`f{FJLkFnb+hb1x_m(fVlW292Kqd{jw#>EgR> z8gK`t6}{yDHhx&k?gHxzAhms-k#X_j9fQe`*KeMVafh9`eECRFNz{vhXJu8(_ZM$J zocd1tbF#exR3|e!2^Od%#)Wfi>$*bfjWhB~L$*JD16L6G!){Dhqw#JJ7Hpn|f7;Lb z_O0$`(a3c+O*4m=k*|DvrgwsKc)1^I*!NkiqJJ$_AKpM5e%}fuTJ+>L zE7bHkt%}9ni#x-6s@c^{SwW$^#aMB0bmsPtP!$yww4gFW5{OsJ6JOBHn)F}R*ME=d z6W&yY!Uj30896>TXPqhgs5Z)6{3K0tB0E9@%v^oA1RczVWZJ%0Pq!L%8|4a?bxg5Q z%!dSEs|OY74s%_8fQ?6+hoq^^h7cnucWSr2eSAo-@)cfl>BFK$D?&^|dEbws(2Hj7 zgawc@uLJnOyH_zk%oF74gOKN}QS5vL>4R$Fp9TuW!_SXOo=EOw%|0&HH>*E>NjmoL z2FiKF=~NaaEsLnVE;N~!yIN2EhDMzvyWlSQj%;*v+=J#am&fVbsDN;$^ivZf6!^Hs zfXxRYXk}aoM&|ZrO6Oyige$?+2^bt=g`vegnC;0kH24PPzeR8UkpoQJQd_@%rh59z z({}vGSe{WaY}~OEzu)nG%gC)(v67*Qqr8jLI`wAX_e4F{txDB5?i|j+Semzb(h^)| z8{?}xpCRaF+cVPG2N%BPwVyEw)VBuE?0Y|demy8biH!>nu7#D$$6J-`^AU%+Lut>U z?24df3ld4NA>V(HzZt@!DWK_jK`;G$wTEQ5KSn^-ZP?SxEBpyxvmuQDjA?3{uZ%M3 zg|P;Tq`$rFbd-UC;cL3J#Kty^#^Nt!pqVKI`I=*G#3qX_4&G<%8zU|K>Y&^C6=Za@ zvMFWk>~EhLf)}MioyQw*)|I7O?a>kIeLfPTtBfm;58Y>!U;^K<$gqLLpr8Z9QCer~ zXrKBIQTMtod+evp(fgsPt46ve{o&51qp%d8eflIQ1oDfvGaFL!<~>#yO?k-hPv`PR@8aEq=%uqefUD*boT3 z6K_`Q*pr#^0@<4kubh0N6vj_|+=1_+$?`G9U?H};JxJ&xd756yww&8(Cf~N$zPhk% z#QbpcsUgx*t#X#{h}W$khX2ANVUJ?$cd@7=*g>wf9F3!bBU`74IU@(!qTM{O^h3~L>_J}O0U?#I&l|rgQzZeb*mGgf(A<_dPbwp$ zsDdVd`5fLf*UBqKH8gmEL*>Tfq0}yE_uPr4|6hAu|ioN;|bO$bwkf-iNz|od0!OYVn(q`DvBftTsam-x^vLjto}+8l|zf zxGB@>BEQ^OUnkKcC`_Y|&$l1xvfW!H;q$#w`+S`^Rn+qBrc9;QxM#i$~ znrdnVkSkv#F#B}Wr3l`Bep{_)J7dBvvo*u2vT-;+4f~LO`hc0rJWB8IKEATS@9gMH zP$CEpx^Pu6noe3~k-z z$R8Wvu%d#(5WkTg`VTpPXqt;L4xd4yU_6VYDaIUL$P3j_SrL*I7uji?%NhxYe-eDm zYP*iso>OG{YxTq3P?^;_DsnQi4~$XS98KnijcAtOq^}uxp)11r%FAp{>=c46qinY~ zuLNDm5m?2Jnzd9}d02#&p-))UY$vUjn@5`H?rbTC&;CxHoFqdIV-U3s$*pf8-{WnW z6d!g3vIaE zH)b00lv|Q-CR*IonQd6+JCuVOm&uB#s4)MqTS+k-;;?A#?dj>c7s;mmVaeOaudM&X z)En!x1HYp|h|~Z$lYQ9dyK$73U7%-$Ss0vu``Ysk%`Y!X0oI=JXvqYtTe#-76|CD< zFctx6Js#|dTc%C|M?^cXy>*gJ_0utT5)zx*&C#&0`8d%!?@;}?-D&DCTA`N;CE_;O zkr-vUnSXbX%qhg!rOp7zUL@zy^rztx3dUrZ26+c3EWiNO0)>8XRomMOyr;Y+Ehk6i z?BYV$E`QXGTZNgP-P#%=*WP#&$Th9R>u8NYdh`?u+S1lmKrTn0eUz7%Pr8--%4H7d z7lDw~odt0+a&pUIP5-wM5ih`0TC47%g0fFxgKyr>t#8Qt+m=CER(t@$-Z4*c-CQF$ z9fGCzD9iAq;N)mR_U6XUi!F8=LtSx}^ihpKfGj&hoq7Efm6fYOVP=8t&n%|;K)?27 z!^k9=d~j+-Vvkt*Zw+H<7Cok&Oxoy}Ut-QSaE0~?>UGCt#gs})K>jv(Op)fzW7rJ~o*lw@z&T~y{x zvhN?!QOFPdDmEz&22##B$^ngko;DpXBuR`IV%;N0#{uY{)R?Q=zGTO75xMnDsOv#6 zrGz%AkI;X$)a=FH#hjc)jofR_w=7zot;H%;2vO90l)s5qofT*mm$as z(P(HMsuqvbai=ni*nfCUw>|f1=lS{V+h4aERyy-7Q{p8Htw1B@?e8xGq@pEG;n-6i zy=Wn(+Ey6N`3eNzch65{RZ1)bjtLrfGLOSd>|Owj52OSldYYU=4vfVW-h*eR&(!U^ z-=-dGfo)7gO1ABL!IlTtxeHK}stpOCh;;t7kQh+d1apFVSpg+%>VI6yk$#dSrX-*w z!XX+xZBo3?0Yppi_bwJ%jDd zbs@W7!LQd;qBRm^-C{5*(YgMU;RtX4$yMF7_^V*nFZ+q77x#+oC}tZAe>+`QvY#de zf=U)3D)br{xchMWEB$`D6CW2VAd2j?7{yPJU+ETyI;8ha|eYH z_|}Qau6ae0?Ug0_8MZamG3@()X`CliqLUx1A5*KX;+XtTlo5dyb{~Y1P`;wwnh~ zbqk%G7am*dQtN%1y<1%hl}|LpJu0J;prmt0a&sm?7PSUe%1Y1BTeh!WpB+oo#m;bP zlsl|18gn6ov84-zYq^e$nb70QM@Ozl`SSdE!9;f+KPs0`3=%s{va)VhJz`+Fd8>7F{EkvU(-3{0A=~VLreHY&T08)Y9b$ zaK&j-rJ?nJUhBot$Njd>cIQsMz;F-l_X8oFQhjINh%bI6r8sH9IUB66x~T#vFrt# zq;-mTKfT0+1T9%Hw0xn&Xn^@g0EWz##(Ffa$vqkp%jLYF{Voe$$q1-Y7m_6~_9)(G zeOTN%3MlE#rvWc{{*^}j^NYE4cuA5Nk~+q-C*=Q-aR97D>2{5xwpxtIHL}abu6UJi z?B{1EU1f*QmD>Ja>g^qe@`SJcaYVqy8@V4|^Zv#<)>FMn>iB8|GERsZ7nhkn!ek*a z>^n8V5xLRHjQ{&pY|>D}-CveDdfJ^2pLVMmmDGPB_wMyVmRGH;KTZDnv7`o`9)<$q zaDBedAN$u^l5i3~z3)B1bEe-vCfoHIdgWh2?lld}4_34nrDar9R#pmk8Jieuc@^!n zmo++Ch%~w+b<(E~sDN}doLppS6lg5-rLFxd!(eE&lR9Cl^G7*;*zMjfN4g%;C{skE zIscG5OzA9xwa{Rv@x3^ct$Vbn?iaQt@gPan)+Xd|uJXL?nB71|Vt)d>?FANw66~zV_UxhLmzTPFi%q3xX9Bb=rY*TXV(hD~Eq|ghr{_1RP$@3a zP&Vvx=!1*#bsEJ6R~Q%=2WDpvNk~YX)l))b9368l)i$Qw^#WhMeB$b|aVlEjNeyn7 zZN8OD&13AvGN>vC8u-DnNt2pe@LMko?q@exiRc2HkzRaf+SVxUUg+nK)Ycxmcvv`r zbHW{f)-Y4r9aEwI)LlA6devy5bcjfQccH z`C)WEu1Qo(OzCAzY%G1Ba8I!yxDdYBd~jAP+v+K-0xjy3QO{rH-W_xwUzKl``-87L zNrq7)yA8r1Vf$(CDnI67tCc&7`Y>!|I29Jk_Y*vm@oomxcc7%b2sbGj4Si;*=TlBZ z8L%??Jf4<)=3h3hXMmp4qPv*(PX5C(P(Wbw7Hw8bVc2{d`1m?vn2pL8 zILR4wdUcx9l_l2JewEfVQ!ktw;@t^Z&m{~~${=Fjiv_m}7LExUVF0A2&%Q!)r{H1s zXS2S|NAu~^)~hgQ;uRjwB3dMhr1A$YMZX|?0F3H1KYv>>dy$|PJK?W=H7*BmJt`0X z_&&wAYmnc;Ee0w)GMJ?cIFh1vGFN}zkTGfb#Y0@^#j-G=UG%ddOrQcHu}ZUVqmsdf zPs4*o;K{&a&L}={AvOorpHtsthw`NAe?1~1MPX$gl-ey)6O%1@2jGm)-uGM{6@(`K zBC%Rm)*d19k-92l_nj9IXm3fH5y=V1seG<_!DblmcJ4+*#;a>-fd&=S#G_S}m8Yhx z`tq|yDd805U-;ivB9~iCk0WvMqa7|e8FHGic7}#NsYrK@mhi3Y`?q%`kZJhMF4EK0i(!9cNac_|9ZzWkvO$lRop`LXiKI5~y^BV$fSeLolfZ zLAn?DJ+z3$~_}W;MYvfi| z&2HLY1_Fbv7QWv-8{Z1Yf_OAk#G1!->!-Ri7?r7&3mQ)b1KEyyP2{=b3;6NriAiMN zd+WL4b}4tx_JJ}FOSa%7$f=0?%1~SvazUG#okjZK31QR%VL^@UOUdf&x9+hoRY%s1 zy+uRJOaDG2MfJ!kcl|Fss1mOBko!*gVsOR^Atd5hO2>y?G;tR-ig`sOK7-=RoN!}s zV705kpO#sJvU*W8&2~sD#@n09d10LiGWR?($Y?37=~os=Hn8>eK18qPTkZ59 zl`)7@fwn{S2r8*h%qd(lRja5rOzdLsP}uwSRP7!4Xk@HD+yg3&@7s7RdE7LiuTTJu zuKMg}h>PMq0VW-#oy~?o3u`)W%zSF6_#gD$631@uYML9!6Yn_{rLkDG{4CzV0$t|s zY{AudbijhMk*}on-VN=W&q|rysHmti%tb{-E&S-y73JkgB#p*u#O?;TOUOdr!i>(g zSV=2;hdSW|!>UtE4*;g~u0yYhkDTD?QU{CE{&mEio(~3 - + - + - + - - + + + + + + - + - - + + + + + - + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + - + - + @@ -72,7 +141,7 @@ - + @@ -83,18 +152,6 @@ - - - - - - - - - - - - @@ -104,12 +161,12 @@ - + - - + + - + @@ -122,12 +179,6 @@ - - - - - - @@ -149,18 +200,13 @@ - + + - - - - - - @@ -186,7 +232,7 @@ - + @@ -287,12 +333,6 @@ - - - - - - @@ -340,14 +380,14 @@ - - + + - - + + @@ -357,16 +397,13 @@ - + - + - - - - + @@ -384,13 +421,13 @@ - + - + @@ -411,22 +448,22 @@ - + - + - + - + - + - - + + @@ -560,60 +597,14 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - + - + - + @@ -624,7 +615,7 @@ - + @@ -633,27 +624,68 @@ - + - + - + - + - + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/index.html b/index.html index b8737a0b..dd99611e 100644 --- a/index.html +++ b/index.html @@ -351,5 +351,5 @@

Citing ARETE

diff --git a/search/search_index.json b/search/search_index.json index 7bc5d54f..168133b5 100644 --- a/search/search_index.json +++ b/search/search_index.json @@ -1 +1 @@ -{"config":{"indexing":"full","lang":["en"],"min_search_length":3,"prebuild_index":false,"separator":"[\\s\\-]+"},"docs":[{"location":"","text":"What is ARETE? ARETE (Antimicrobial Resistance: Emergence, Transmission, and Ecology) is a bioinformatics best-practice analysis pipeline for profiling the genomic repertoire and evolutionary dynamics of microorganisms with a particular focus on pathogens. We use ARETE to identify important genes (e.g., those that confer antimicrobial resistance or contribute to virulence) and mobile genetic elements such as plasmids and genomic islands, and infer important routes by which these are transmitted using evidence from recombination, cosegregation, coevolution, and phylogenetic trees comparisons. ARETE produces a range of useful outputs (see outputs ), including those generated by each tool integrated into the pipeline, as well as summaries across the entire dataset such as phylogenetic profiles. Outputs from ARETE can also be easily fed into packages such as Coeus and MicroReact for further analyses. Although ARETE was primarily developed with pathogens in mind, inference of pan-genomes, mobilomes, and phylogenomic histories can be performed for any set of microbial genomes, with the proviso that reference databases are much more complete for some taxonomic groups than others. In general, the tools in ARETE work best at the species and genus level of relatedness. A key design feature of ARETE is the versatility to find the right blend of software packages and parameter settings that best handle datasets of different sizes, introducing heuristics and swapping out tools as necessary. ARETE has been benchmarked on datasets from fewer than ten to over 10,000 genomes from a diversity of species and genera including Enterococcus faecium , Escherichia coli , Listeria , and Salmonella . Another key feature is enabling the user choice to run specific subsets of the pipeline; a user may already have assembled genomes, or they may not care about, say, recombination detection. There are also cases where it might be necessary to manually review the outputs from a particular step before moving on to the next one; ARETE makes this manual QC easy to do. Table of Contents About the pipeline Quick start A couple of examples Credits Contributing to ARETE Citing ARETE About the pipeline The pipeline is built using Nextflow , a workflow tool to run tasks across multiple compute infrastructures in a very portable manner. It uses Docker / Singularity containers making installation trivial and results highly reproducible. Like other workflow languages it provides useful features like -resume to only rerun tasks that haven't already been completed (e.g., allowing editing of inputs/tasks and recovery from crashes without a full re-run). The nf-core project provided overall project template, pre-written software modules when available, and general best-practice recommendations. ARETE is organized as a series of subworkflows, each of which executes a different conceptual step of the pipeline. The subworkflow orgnaization provides suitable entry and exit points for users who want to run only a portion of the full pipeline. Genome subsetting The user can optionally subdivide their set of genomes into related lineages identified by PopPUNK ( See documentation ). PopPUNK quickly assignes genomes to 'lineages' based on core and accessory genome identity. If this option is selected, all genomes will still be annotated, but cross-genome comparisons (e.g., pan-genome inference and phylogenomics) will use only a single representative genome from each lineage. The user can run PopPUNK with a spread of different thresholds and decide how to proceed based on the number of lineages produced and their own specific knowledge of the genetic population structure of the taxon being analyzed. Short-read processing and assembly Raw Read QC ( FastQC ) Read Trimming ( fastp ) Trimmed Read QC ( FastQC ) Taxonomic Profiling ( kraken2 ) Unicycler ( unicycler ) QUAST QC ( quast ) CheckM QC ( checkm ) Annotation Genome annotation with Bakta ( bakta ) or Prokka ( prokka ) Feature prediction: AMR genes with the Resistance Gene Identifier ( RGI ) Plasmids with MOB-Suite ( mob_suite ) Genomic Islands with IslandPath ( IslandPath ) Phages with PhiSpy ( PhiSpy ) ( optionally ) Integrons with IntegronFinder Specialized databases: CAZY, VFDB, BacMet and ICEberg2 using DIAMOND homology search ( diamond ) Phylogenomics ( optionally ) Genome subsetting with PopPUNK ( See documentation ) Pan-genome inference using PPanGGOLiN ( PPanGGOLiN ) or Panaroo ( panaroo ) Reference and gene tree inference using FastTree ( fasttree ) or IQTree ( iqtree ) ( optionally ) SNP-sites ( SNPsites ) Recombination detection Recombination detection is performed within lineages identified by PopPUNK ( poppunk ). Note that this application of PopPUNK is different from the subsetting described above. Genome alignment using SKA2 ( ska2 ) Recombination detection using Verticall ( verticall ) and/or Gubbins ( gubbins ) Coevolution Identification of coordinated gain and loss of features using EvolCCM ( EvolCCM ) Lateral gene transfer Phylogenetic inference of LGT using rSPR ( rSPR ) Gene order Comparison of genomic neighbourhoods using the Gene Order Workflow ( Gene Order Workflow ) See our roadmap for a full list of future development targets. Quick Start Install nextflow Install Docker , Singularity , or, as a last resort, Conda . Also ensure you have a working curl installed (should be present on almost all systems). 2.1. Note: this workflow should also support Podman , Shifter or Charliecloud execution for full pipeline reproducibility. We have minimized reliance on conda and suggest using it only as a last resort (see docs ). Configure mail on your system to send an email on workflow success/failure (without this you may get a small error at the end Failed to invoke workflow.onComplete event handler but this doesn't mean the workflow didn't finish successfully). Download the pipeline and test with a stub-run . The stub-run will ensure that the pipeline is able to download and use containers as well as execute in the proper logic. nextflow run beiko-lab/ARETE -profile test, -stub-run 3.1. Please check nf-core/configs to see if a custom config file to run nf-core pipelines already exists for your Institute. If so, you can simply use -profile in your command. This will enable either docker or singularity and set the appropriate execution settings for your local compute environment. 3.2. If you are using singularity then the pipeline will auto-detect this and attempt to download the Singularity images directly as opposed to performing a conversion from Docker images. If you are persistently observing issues downloading Singularity images directly due to timeout or network issues then please use the --singularity_pull_docker_container parameter to pull and convert the Docker image instead. In case of input datasets larger than 100 samples, check our resource profiles documentation , for optimal usage. Start running your own analysis (ideally using -profile docker or -profile singularity for stability)! nextflow run beiko-lab/ARETE \\ -profile \\ --input_sample_table samplesheet.csv \\ --poppunk_model bgmm samplesheet.csv must be formatted sample,fastq_1,fastq_2 Note : If you get this error at the end Failed to invoke `workflow.onComplete` event handler it isn't a problem, it just means you don't have an sendmail configured and it can't send an email report saying it finished correctly i.e., its not that the workflow failed. See usage docs for all of the available options when running the pipeline. See the parameter docs for a list of all params currently implemented in the pipeline and which ones are required. Testing To test the worklow on a minimal dataset you can use the test configuration (with either docker or singularity - replace docker below as appropriate): nextflow run beiko-lab/ARETE -profile test,docker To accelerate it you can download/cache the database files to a folder (e.g., test/db_cache ) and provide a database cache parameter. nextflow run beiko-lab/ARETE \\ -profile test,docker \\ --db_cache $PWD/test/db_cache \\ --bakta_db $PWD/baktadb/db-light We also provide a larger test dataset, under -profile test_full , for use in ARETE's annotation entry. This dataset is comprised of 8 bacterial genomes. As a note, this can take upwards of 20 minutes to complete on an average personal computer . Replace docker below as appropriate. nextflow run beiko-lab/ARETE -entry annotation -profile test_full,docker Examples The fine details of how to run ARETE are described in the command reference and documentation, but here are a couple of illustrative examples of how runs can be adjusted to accommodate genome sets of different sizes: Assembly, annotation, and pan-genome inference from a modestly sized dataset (50 or so genomes) from paired-end reads nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --annotation_tools 'mobsuite,rgi,vfdb,bacmet,islandpath,phispy,report' \\ --poppunk_model bgmm \\ -profile docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --annotation_tools - Select the annotation tools and modules to be executed (See the parameter documentation for defaults) --poppunk_model - Model to be used by PopPUNK -profile docker - Run tools in docker containers. Annotation to evolutionary dynamics on 300-ish genomes nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ --run_gubbins \\ -entry annotation \\ -profile medium,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --run_gubbins - Run Gubbins as part of the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile medium,docker - Run tools in docker containers. For -profile medium , check our resource requirements documentation . Annotation to evolutionary dynamics on 10,000 genomes nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ -entry annotation \\ -profile large,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. --enable_subsetting - Enable subsetting workflow based on genome similarity (See subsetting documentation ) -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile large,docker - Run tools in docker containers. For -profile large , check our resource requirements documentation . Credits The ARETE software was originally written and developed by Finlay Maguire and Alex Manuele , and is currently developed by Jo\u00e3o Cavalcante . Rob Beiko is the PI of the ARETE project. The project Co-PI is Fiona Brinkman. Other project leads include Andrew MacArthur, Cedric Chauve, Chris Whidden, Gary van Domselaar, John Nash, Rahat Zaheer, and Tim McAllister. Many students, postdocs, developers, and staff scientists have made invaluable contributions to the design and application of ARETE and its components, including Haley Sanderson, Kristen Gray, Julia Lewandowski, Chaoyue Liu, Kartik Kakadiya, Bryan Alcock, Amos Raphenya, Amjad Khan, Ryan Fink, Aniket Mane, Chandana Navanekere Rudrappa, Kyrylo Bessonov, James Robertson, Jee In Kim, and Nolan Woods. ARETE development has been supported from many sources, including Genome Canada, ResearchNS, Genome Atlantic, Genome British Columbia, The Canadian Institutes for Health Research, The Natural Sciences and Engineering Research Council of Canada, and Dalhousie University's Faculty of Computer Science. We have received tremendous support from federal agencies, most notably the Public Health Agency of Canada and Agriculture / Agri-Food Canada. Contributing to ARETE Thank you for your interest in contributing to ARETE. We are currently in the process of formalizing contribution guidelines. In the meantime, please feel free to open an issue describing your suggested changes. Citing ARETE Please cite the tools used in your ARETE run: A comprehensive list can be found in the CITATIONS.md file. An early version of ARETE was used for assembly and feature prediction in the following paper : Sanderson H, Gray KL, Manuele A, Maguire F, Khan A, Liu C, Navanekere Rudrappa C, Nash JHE, Robertson J, Bessonov K, Oloni M, Alcock BP, Raphenya AR, McAllister TA, Peacock SJ, Raven KE, Gouliouris T, McArthur AG, Brinkman FSL, Fink RC, Zaheer R, Beiko RG. Exploring the mobilome and resistome of Enterococcus faecium in a One Health context across two continents. Microb Genom. 2022 Sep;8(9):mgen000880. doi: 10.1099/mgen.0.000880. PMID: 36129737; PMCID: PMC9676038. This pipeline uses code and infrastructure developed and maintained by the nf-core initative, and reused here under the MIT license . The nf-core framework for community-curated bioinformatics pipelines. Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen. Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.","title":"Home"},{"location":"#what-is-arete","text":"ARETE (Antimicrobial Resistance: Emergence, Transmission, and Ecology) is a bioinformatics best-practice analysis pipeline for profiling the genomic repertoire and evolutionary dynamics of microorganisms with a particular focus on pathogens. We use ARETE to identify important genes (e.g., those that confer antimicrobial resistance or contribute to virulence) and mobile genetic elements such as plasmids and genomic islands, and infer important routes by which these are transmitted using evidence from recombination, cosegregation, coevolution, and phylogenetic trees comparisons. ARETE produces a range of useful outputs (see outputs ), including those generated by each tool integrated into the pipeline, as well as summaries across the entire dataset such as phylogenetic profiles. Outputs from ARETE can also be easily fed into packages such as Coeus and MicroReact for further analyses. Although ARETE was primarily developed with pathogens in mind, inference of pan-genomes, mobilomes, and phylogenomic histories can be performed for any set of microbial genomes, with the proviso that reference databases are much more complete for some taxonomic groups than others. In general, the tools in ARETE work best at the species and genus level of relatedness. A key design feature of ARETE is the versatility to find the right blend of software packages and parameter settings that best handle datasets of different sizes, introducing heuristics and swapping out tools as necessary. ARETE has been benchmarked on datasets from fewer than ten to over 10,000 genomes from a diversity of species and genera including Enterococcus faecium , Escherichia coli , Listeria , and Salmonella . Another key feature is enabling the user choice to run specific subsets of the pipeline; a user may already have assembled genomes, or they may not care about, say, recombination detection. There are also cases where it might be necessary to manually review the outputs from a particular step before moving on to the next one; ARETE makes this manual QC easy to do.","title":"What is ARETE?"},{"location":"#table-of-contents","text":"About the pipeline Quick start A couple of examples Credits Contributing to ARETE Citing ARETE","title":"Table of Contents"},{"location":"#about-the-pipeline","text":"The pipeline is built using Nextflow , a workflow tool to run tasks across multiple compute infrastructures in a very portable manner. It uses Docker / Singularity containers making installation trivial and results highly reproducible. Like other workflow languages it provides useful features like -resume to only rerun tasks that haven't already been completed (e.g., allowing editing of inputs/tasks and recovery from crashes without a full re-run). The nf-core project provided overall project template, pre-written software modules when available, and general best-practice recommendations. ARETE is organized as a series of subworkflows, each of which executes a different conceptual step of the pipeline. The subworkflow orgnaization provides suitable entry and exit points for users who want to run only a portion of the full pipeline.","title":"About the pipeline "},{"location":"#genome-subsetting","text":"The user can optionally subdivide their set of genomes into related lineages identified by PopPUNK ( See documentation ). PopPUNK quickly assignes genomes to 'lineages' based on core and accessory genome identity. If this option is selected, all genomes will still be annotated, but cross-genome comparisons (e.g., pan-genome inference and phylogenomics) will use only a single representative genome from each lineage. The user can run PopPUNK with a spread of different thresholds and decide how to proceed based on the number of lineages produced and their own specific knowledge of the genetic population structure of the taxon being analyzed.","title":"Genome subsetting"},{"location":"#short-read-processing-and-assembly","text":"Raw Read QC ( FastQC ) Read Trimming ( fastp ) Trimmed Read QC ( FastQC ) Taxonomic Profiling ( kraken2 ) Unicycler ( unicycler ) QUAST QC ( quast ) CheckM QC ( checkm )","title":"Short-read processing and assembly"},{"location":"#annotation","text":"Genome annotation with Bakta ( bakta ) or Prokka ( prokka ) Feature prediction: AMR genes with the Resistance Gene Identifier ( RGI ) Plasmids with MOB-Suite ( mob_suite ) Genomic Islands with IslandPath ( IslandPath ) Phages with PhiSpy ( PhiSpy ) ( optionally ) Integrons with IntegronFinder Specialized databases: CAZY, VFDB, BacMet and ICEberg2 using DIAMOND homology search ( diamond )","title":"Annotation"},{"location":"#phylogenomics","text":"( optionally ) Genome subsetting with PopPUNK ( See documentation ) Pan-genome inference using PPanGGOLiN ( PPanGGOLiN ) or Panaroo ( panaroo ) Reference and gene tree inference using FastTree ( fasttree ) or IQTree ( iqtree ) ( optionally ) SNP-sites ( SNPsites )","title":"Phylogenomics"},{"location":"#recombination-detection","text":"Recombination detection is performed within lineages identified by PopPUNK ( poppunk ). Note that this application of PopPUNK is different from the subsetting described above. Genome alignment using SKA2 ( ska2 ) Recombination detection using Verticall ( verticall ) and/or Gubbins ( gubbins )","title":"Recombination detection"},{"location":"#coevolution","text":"Identification of coordinated gain and loss of features using EvolCCM ( EvolCCM )","title":"Coevolution"},{"location":"#lateral-gene-transfer","text":"Phylogenetic inference of LGT using rSPR ( rSPR )","title":"Lateral gene transfer"},{"location":"#gene-order","text":"Comparison of genomic neighbourhoods using the Gene Order Workflow ( Gene Order Workflow ) See our roadmap for a full list of future development targets.","title":"Gene order"},{"location":"#quick-start","text":"Install nextflow Install Docker , Singularity , or, as a last resort, Conda . Also ensure you have a working curl installed (should be present on almost all systems). 2.1. Note: this workflow should also support Podman , Shifter or Charliecloud execution for full pipeline reproducibility. We have minimized reliance on conda and suggest using it only as a last resort (see docs ). Configure mail on your system to send an email on workflow success/failure (without this you may get a small error at the end Failed to invoke workflow.onComplete event handler but this doesn't mean the workflow didn't finish successfully). Download the pipeline and test with a stub-run . The stub-run will ensure that the pipeline is able to download and use containers as well as execute in the proper logic. nextflow run beiko-lab/ARETE -profile test, -stub-run 3.1. Please check nf-core/configs to see if a custom config file to run nf-core pipelines already exists for your Institute. If so, you can simply use -profile in your command. This will enable either docker or singularity and set the appropriate execution settings for your local compute environment. 3.2. If you are using singularity then the pipeline will auto-detect this and attempt to download the Singularity images directly as opposed to performing a conversion from Docker images. If you are persistently observing issues downloading Singularity images directly due to timeout or network issues then please use the --singularity_pull_docker_container parameter to pull and convert the Docker image instead. In case of input datasets larger than 100 samples, check our resource profiles documentation , for optimal usage. Start running your own analysis (ideally using -profile docker or -profile singularity for stability)! nextflow run beiko-lab/ARETE \\ -profile \\ --input_sample_table samplesheet.csv \\ --poppunk_model bgmm samplesheet.csv must be formatted sample,fastq_1,fastq_2 Note : If you get this error at the end Failed to invoke `workflow.onComplete` event handler it isn't a problem, it just means you don't have an sendmail configured and it can't send an email report saying it finished correctly i.e., its not that the workflow failed. See usage docs for all of the available options when running the pipeline. See the parameter docs for a list of all params currently implemented in the pipeline and which ones are required.","title":"Quick Start "},{"location":"#testing","text":"To test the worklow on a minimal dataset you can use the test configuration (with either docker or singularity - replace docker below as appropriate): nextflow run beiko-lab/ARETE -profile test,docker To accelerate it you can download/cache the database files to a folder (e.g., test/db_cache ) and provide a database cache parameter. nextflow run beiko-lab/ARETE \\ -profile test,docker \\ --db_cache $PWD/test/db_cache \\ --bakta_db $PWD/baktadb/db-light We also provide a larger test dataset, under -profile test_full , for use in ARETE's annotation entry. This dataset is comprised of 8 bacterial genomes. As a note, this can take upwards of 20 minutes to complete on an average personal computer . Replace docker below as appropriate. nextflow run beiko-lab/ARETE -entry annotation -profile test_full,docker","title":"Testing"},{"location":"#examples","text":"The fine details of how to run ARETE are described in the command reference and documentation, but here are a couple of illustrative examples of how runs can be adjusted to accommodate genome sets of different sizes:","title":"Examples "},{"location":"#assembly-annotation-and-pan-genome-inference-from-a-modestly-sized-dataset-50-or-so-genomes-from-paired-end-reads","text":"nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --annotation_tools 'mobsuite,rgi,vfdb,bacmet,islandpath,phispy,report' \\ --poppunk_model bgmm \\ -profile docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --annotation_tools - Select the annotation tools and modules to be executed (See the parameter documentation for defaults) --poppunk_model - Model to be used by PopPUNK -profile docker - Run tools in docker containers.","title":"Assembly, annotation, and pan-genome inference from a modestly sized dataset (50 or so genomes) from paired-end reads"},{"location":"#annotation-to-evolutionary-dynamics-on-300-ish-genomes","text":"nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ --run_gubbins \\ -entry annotation \\ -profile medium,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --run_gubbins - Run Gubbins as part of the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile medium,docker - Run tools in docker containers. For -profile medium , check our resource requirements documentation .","title":"Annotation to evolutionary dynamics on 300-ish genomes"},{"location":"#annotation-to-evolutionary-dynamics-on-10000-genomes","text":"nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ -entry annotation \\ -profile large,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. --enable_subsetting - Enable subsetting workflow based on genome similarity (See subsetting documentation ) -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile large,docker - Run tools in docker containers. For -profile large , check our resource requirements documentation .","title":"Annotation to evolutionary dynamics on 10,000 genomes"},{"location":"#credits","text":"The ARETE software was originally written and developed by Finlay Maguire and Alex Manuele , and is currently developed by Jo\u00e3o Cavalcante . Rob Beiko is the PI of the ARETE project. The project Co-PI is Fiona Brinkman. Other project leads include Andrew MacArthur, Cedric Chauve, Chris Whidden, Gary van Domselaar, John Nash, Rahat Zaheer, and Tim McAllister. Many students, postdocs, developers, and staff scientists have made invaluable contributions to the design and application of ARETE and its components, including Haley Sanderson, Kristen Gray, Julia Lewandowski, Chaoyue Liu, Kartik Kakadiya, Bryan Alcock, Amos Raphenya, Amjad Khan, Ryan Fink, Aniket Mane, Chandana Navanekere Rudrappa, Kyrylo Bessonov, James Robertson, Jee In Kim, and Nolan Woods. ARETE development has been supported from many sources, including Genome Canada, ResearchNS, Genome Atlantic, Genome British Columbia, The Canadian Institutes for Health Research, The Natural Sciences and Engineering Research Council of Canada, and Dalhousie University's Faculty of Computer Science. We have received tremendous support from federal agencies, most notably the Public Health Agency of Canada and Agriculture / Agri-Food Canada.","title":"Credits "},{"location":"#contributing-to-arete","text":"Thank you for your interest in contributing to ARETE. We are currently in the process of formalizing contribution guidelines. In the meantime, please feel free to open an issue describing your suggested changes.","title":"Contributing to ARETE "},{"location":"#citing-arete","text":"Please cite the tools used in your ARETE run: A comprehensive list can be found in the CITATIONS.md file. An early version of ARETE was used for assembly and feature prediction in the following paper : Sanderson H, Gray KL, Manuele A, Maguire F, Khan A, Liu C, Navanekere Rudrappa C, Nash JHE, Robertson J, Bessonov K, Oloni M, Alcock BP, Raphenya AR, McAllister TA, Peacock SJ, Raven KE, Gouliouris T, McArthur AG, Brinkman FSL, Fink RC, Zaheer R, Beiko RG. Exploring the mobilome and resistome of Enterococcus faecium in a One Health context across two continents. Microb Genom. 2022 Sep;8(9):mgen000880. doi: 10.1099/mgen.0.000880. PMID: 36129737; PMCID: PMC9676038. This pipeline uses code and infrastructure developed and maintained by the nf-core initative, and reused here under the MIT license . The nf-core framework for community-curated bioinformatics pipelines. Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen. Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.","title":"Citing ARETE "},{"location":"CITATIONS/","text":"beiko-lab/ARETE: Citations nf-core Ewels PA, Peltzer A, Fillinger S, Patel H, Alneberg J, Wilm A, Garcia MU, Di Tommaso P, Nahnsen S. The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol. 2020 Mar;38(3):276-278. doi: 10.1038/s41587-020-0439-x. PubMed PMID: 32055031. Nextflow Di Tommaso P, Chatzou M, Floden EW, Barja PP, Palumbo E, Notredame C. Nextflow enables reproducible computational workflows. Nat Biotechnol. 2017 Apr 11;35(4):316-319. doi: 10.1038/nbt.3820. PubMed PMID: 28398311. Pipeline tools CheckM Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. 2015. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes . Genome Research, 25: 1043\u20131055. DIAMOND Buchfink B, Xie C, Huson DH Fast and sensitive protein alignment using DIAMOND. Nat. Methods. 12, 59\u201360 (2015) FastQC FastP Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018 Sep 1;34(17):i884-i890. doi: 10.1093/bioinformatics/bty560. PubMed PMID: 30423086; PubMed Central PMCID: PMC6129281. FastTree Morgan N. Price, Paramvir S. Dehal, Adam P. Arkin, FastTree: Computing Large Minimum Evolution Trees with Profiles instead of a Distance Matrix, Molecular Biology and Evolution, Volume 26, Issue 7, July 2009, Pages 1641\u20131650, https://doi.org/10.1093/molbev/msp077 IQ-TREE2 Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, Lanfear R. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Mol Biol Evol. 2020 May 1;37(5):1530-1534. doi: 10.1093/molbev/msaa015. Erratum in: Mol Biol Evol. 2020 Aug 1;37(8):2461. PMID: 32011700; PMCID: PMC7182206. Kraken2 Wood, D et al., 2019. Improved metagenomic analysis with Kraken 2. Genome Biology volume 20, Article number: 257. doi: 10.1186/s13059-019-1891-0. MOB-SUITE Robertson, James, and John H E Nash. \u201cMOB-suite: software tools for clustering, reconstruction and typing of plasmids from draft assemblies.\u201d Microbial genomics vol. 4,8 (2018): e000206. doi:10.1099/mgen.0.000206 Robertson, James et al. \u201cUniversal whole-sequence-based plasmid typing and its utility to prediction of host range and epidemiological surveillance.\u201d Microbial genomics vol. 6,10 (2020): mgen000435. doi:10.1099/mgen.0.000435 MultiQC Ewels P, Magnusson M, Lundin S, K\u00e4ller M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. 2016 Oct 1;32(19):3047-8. doi: 10.1093/bioinformatics/btw354. Epub 2016 Jun 16. PubMed PMID: 27312411; PubMed Central PMCID: PMC5039924. Bakta Schwengers O., Jelonek L., Dieckmann M. A., Beyvers S., Blom J., Goesmann A. (2021). Bakta: rapid and standardized annotation of bacterial genomes via alignment-free sequence identification. Microbial Genomics, 7(11). https://doi.org/10.1099/mgen.0.000685 Prokka Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014 Jul 15;30(14):2068-9. doi: 10.1093/bioinformatics/btu153. Epub 2014 Mar 18. PMID: 24642063. QUAST Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013 Apr 15;29(8):1072-5. doi: 10.1093/bioinformatics/btt086. Epub 2013 Feb 19. PMID: 23422339; PMCID: PMC3624806. RGI Alcock et al. 2020. CARD 2020: antibiotic resistome surveillance with the comprehensive antibiotic resistance database. Nucleic Acids Research, Volume 48, Issue D1, Pages D517-525 [PMID 31665441] IntegronFinder N\u00e9ron, Bertrand, Eloi Littner, Matthieu Haudiquet, Amandine Perrin, Jean Cury, and Eduardo P.C. Rocha. 2022. IntegronFinder 2.0: Identification and Analysis of Integrons across Bacteria, with a Focus on Antibiotic Resistance in Klebsiella Microorganisms 10, no. 4: 700. https://doi.org/10.3390/microorganisms10040700 Panaroo Tonkin-Hill, G., MacAlasdair, N., Ruis, C. et al. Producing polished prokaryotic pangenomes with the Panaroo pipeline. Genome Biol 21, 180 (2020). https://doi.org/10.1186/s13059-020-02090-4 PPanGGoLiN Gautreau G et al. (2020) PPanGGOLiN: Depicting microbial diversity via a partitioned pangenome graph. PLOS Computational Biology 16(3): e1007732. https://doi.org/10.1371/journal.pcbi.1007732 PopPUNK Lees JA, Harris SR, Tonkin-Hill G, Gladstone RA, Lo SW, Weiser JN, Corander J, Bentley SD, Croucher NJ. Fast and flexible bacterial genomic epidemiology with PopPUNK. Genome Res. 2019 Feb;29(2):304-316. doi: 10.1101/gr.241455.118. Epub 2019 Jan 24. PMID: 30679308; PMCID: PMC6360808. SKA2 Harris SR. 2018. SKA: Split Kmer Analysis Toolkit for Bacterial Genomic Epidemiology. bioRxiv 453142 doi: https://doi.org/10.1101/453142 Gubbins Croucher N. J., Page A. J., Connor T. R., Delaney A. J., Keane J. A., Bentley S. D., Parkhill J., Harris S.R. \"Rapid phylogenetic analysis of large samples of recombinant bacterial whole genome sequences using Gubbins\". doi:10.1093/nar/gku1196, Nucleic Acids Research, 2014. Verticall SNP-sites Page AJ, Taylor B, Delaney AJ, Soares J, Seemann T, Keane JA, Harris SR. SNP-sites: rapid efficient extraction of SNPs from multi-FASTA alignments. Microb Genom. 2016 Apr 29;2(4):e000056. doi: 10.1099/mgen.0.000056. PMID: 28348851; PMCID: PMC5320690. Unicycler Wick RR, Judd LM, Gorrie CL, Holt KE. Unicycler: Resolving bacterial genome assemblies from short and long sequencing reads. PLoS Comput Biol. 2017 Jun 8;13(6):e1005595. doi: 10.1371/journal.pcbi.1005595. PMID: 28594827; PMCID: PMC5481147. IslandPath Claire Bertelli, Fiona S L Brinkman, Improved genomic island predictions with IslandPath-DIMOB, Bioinformatics, Volume 34, Issue 13, 01 July 2018, Pages 2161\u20132167, https://doi.org/10.1093/bioinformatics/bty095 PhiSpy Sajia Akhter, Ramy K. Aziz, Robert A. Edwards; PhiSpy: a novel algorithm for finding prophages in bacterial genomes that combines similarity- and composition-based strategies. Nucl Acids Res 2012; 40 (16): e126. doi: 10.1093/nar/gks406 EvolCCM Chaoyue Liu and others, The Community Coevolution Model with Application to the Study of Evolutionary Relationships between Genes Based on Phylogenetic Profiles, Systematic Biology, Volume 72, Issue 3, May 2023, Pages 559\u2013574, https://doi.org/10.1093/sysbio/syac052 rSPR Christopher Whidden, Norbert Zeh, Robert G. Beiko, Supertrees Based on the Subtree Prune-and-Regraft Distance, Systematic Biology, Volume 63, Issue 4, July 2014, Pages 566\u2013581, https://doi.org/10.1093/sysbio/syu023 Software packaging/containerisation tools Anaconda Anaconda Software Distribution. Computer software. Vers. 2-2.4.0. Anaconda, Nov. 2016. Web. Bioconda Gr\u00fcning B, Dale R, Sj\u00f6din A, Chapman BA, Rowe J, Tomkins-Tinch CH, Valieris R, K\u00f6ster J; Bioconda Team. Bioconda: sustainable and comprehensive software distribution for the life sciences. Nat Methods. 2018 Jul;15(7):475-476. doi: 10.1038/s41592-018-0046-7. PubMed PMID: 29967506. BioContainers da Veiga Leprevost F, Gr\u00fcning B, Aflitos SA, R\u00f6st HL, Uszkoreit J, Barsnes H, Vaudel M, Moreno P, Gatto L, Weber J, Bai M, Jimenez RC, Sachsenberg T, Pfeuffer J, Alvarez RV, Griss J, Nesvizhskii AI, Perez-Riverol Y. BioContainers: an open-source and community-driven framework for software standardization. Bioinformatics. 2017 Aug 15;33(16):2580-2582. doi: 10.1093/bioinformatics/btx192. PubMed PMID: 28379341; PubMed Central PMCID: PMC5870671. Docker Dirk Merkel. 2014. Docker: lightweight Linux containers for consistent development and deployment. Linux J. 2014, 239, Article 2 (March 2014). Singularity Kurtzer GM, Sochat V, Bauer MW. Singularity: Scientific containers for mobility of compute. PLoS One. 2017 May 11;12(5):e0177459. doi: 10.1371/journal.pone.0177459. eCollection 2017. PubMed PMID: 28494014; PubMed Central PMCID: PMC5426675.","title":"Citations"},{"location":"CITATIONS/#beiko-labarete-citations","text":"","title":"beiko-lab/ARETE: Citations"},{"location":"CITATIONS/#nf-core","text":"Ewels PA, Peltzer A, Fillinger S, Patel H, Alneberg J, Wilm A, Garcia MU, Di Tommaso P, Nahnsen S. The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol. 2020 Mar;38(3):276-278. doi: 10.1038/s41587-020-0439-x. PubMed PMID: 32055031.","title":"nf-core"},{"location":"CITATIONS/#nextflow","text":"Di Tommaso P, Chatzou M, Floden EW, Barja PP, Palumbo E, Notredame C. Nextflow enables reproducible computational workflows. Nat Biotechnol. 2017 Apr 11;35(4):316-319. doi: 10.1038/nbt.3820. PubMed PMID: 28398311.","title":"Nextflow"},{"location":"CITATIONS/#pipeline-tools","text":"CheckM Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. 2015. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes . Genome Research, 25: 1043\u20131055. DIAMOND Buchfink B, Xie C, Huson DH Fast and sensitive protein alignment using DIAMOND. Nat. Methods. 12, 59\u201360 (2015) FastQC FastP Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018 Sep 1;34(17):i884-i890. doi: 10.1093/bioinformatics/bty560. PubMed PMID: 30423086; PubMed Central PMCID: PMC6129281. FastTree Morgan N. Price, Paramvir S. Dehal, Adam P. Arkin, FastTree: Computing Large Minimum Evolution Trees with Profiles instead of a Distance Matrix, Molecular Biology and Evolution, Volume 26, Issue 7, July 2009, Pages 1641\u20131650, https://doi.org/10.1093/molbev/msp077 IQ-TREE2 Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, Lanfear R. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Mol Biol Evol. 2020 May 1;37(5):1530-1534. doi: 10.1093/molbev/msaa015. Erratum in: Mol Biol Evol. 2020 Aug 1;37(8):2461. PMID: 32011700; PMCID: PMC7182206. Kraken2 Wood, D et al., 2019. Improved metagenomic analysis with Kraken 2. Genome Biology volume 20, Article number: 257. doi: 10.1186/s13059-019-1891-0. MOB-SUITE Robertson, James, and John H E Nash. \u201cMOB-suite: software tools for clustering, reconstruction and typing of plasmids from draft assemblies.\u201d Microbial genomics vol. 4,8 (2018): e000206. doi:10.1099/mgen.0.000206 Robertson, James et al. \u201cUniversal whole-sequence-based plasmid typing and its utility to prediction of host range and epidemiological surveillance.\u201d Microbial genomics vol. 6,10 (2020): mgen000435. doi:10.1099/mgen.0.000435 MultiQC Ewels P, Magnusson M, Lundin S, K\u00e4ller M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. 2016 Oct 1;32(19):3047-8. doi: 10.1093/bioinformatics/btw354. Epub 2016 Jun 16. PubMed PMID: 27312411; PubMed Central PMCID: PMC5039924. Bakta Schwengers O., Jelonek L., Dieckmann M. A., Beyvers S., Blom J., Goesmann A. (2021). Bakta: rapid and standardized annotation of bacterial genomes via alignment-free sequence identification. Microbial Genomics, 7(11). https://doi.org/10.1099/mgen.0.000685 Prokka Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014 Jul 15;30(14):2068-9. doi: 10.1093/bioinformatics/btu153. Epub 2014 Mar 18. PMID: 24642063. QUAST Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013 Apr 15;29(8):1072-5. doi: 10.1093/bioinformatics/btt086. Epub 2013 Feb 19. PMID: 23422339; PMCID: PMC3624806. RGI Alcock et al. 2020. CARD 2020: antibiotic resistome surveillance with the comprehensive antibiotic resistance database. Nucleic Acids Research, Volume 48, Issue D1, Pages D517-525 [PMID 31665441] IntegronFinder N\u00e9ron, Bertrand, Eloi Littner, Matthieu Haudiquet, Amandine Perrin, Jean Cury, and Eduardo P.C. Rocha. 2022. IntegronFinder 2.0: Identification and Analysis of Integrons across Bacteria, with a Focus on Antibiotic Resistance in Klebsiella Microorganisms 10, no. 4: 700. https://doi.org/10.3390/microorganisms10040700 Panaroo Tonkin-Hill, G., MacAlasdair, N., Ruis, C. et al. Producing polished prokaryotic pangenomes with the Panaroo pipeline. Genome Biol 21, 180 (2020). https://doi.org/10.1186/s13059-020-02090-4 PPanGGoLiN Gautreau G et al. (2020) PPanGGOLiN: Depicting microbial diversity via a partitioned pangenome graph. PLOS Computational Biology 16(3): e1007732. https://doi.org/10.1371/journal.pcbi.1007732 PopPUNK Lees JA, Harris SR, Tonkin-Hill G, Gladstone RA, Lo SW, Weiser JN, Corander J, Bentley SD, Croucher NJ. Fast and flexible bacterial genomic epidemiology with PopPUNK. Genome Res. 2019 Feb;29(2):304-316. doi: 10.1101/gr.241455.118. Epub 2019 Jan 24. PMID: 30679308; PMCID: PMC6360808. SKA2 Harris SR. 2018. SKA: Split Kmer Analysis Toolkit for Bacterial Genomic Epidemiology. bioRxiv 453142 doi: https://doi.org/10.1101/453142 Gubbins Croucher N. J., Page A. J., Connor T. R., Delaney A. J., Keane J. A., Bentley S. D., Parkhill J., Harris S.R. \"Rapid phylogenetic analysis of large samples of recombinant bacterial whole genome sequences using Gubbins\". doi:10.1093/nar/gku1196, Nucleic Acids Research, 2014. Verticall SNP-sites Page AJ, Taylor B, Delaney AJ, Soares J, Seemann T, Keane JA, Harris SR. SNP-sites: rapid efficient extraction of SNPs from multi-FASTA alignments. Microb Genom. 2016 Apr 29;2(4):e000056. doi: 10.1099/mgen.0.000056. PMID: 28348851; PMCID: PMC5320690. Unicycler Wick RR, Judd LM, Gorrie CL, Holt KE. Unicycler: Resolving bacterial genome assemblies from short and long sequencing reads. PLoS Comput Biol. 2017 Jun 8;13(6):e1005595. doi: 10.1371/journal.pcbi.1005595. PMID: 28594827; PMCID: PMC5481147. IslandPath Claire Bertelli, Fiona S L Brinkman, Improved genomic island predictions with IslandPath-DIMOB, Bioinformatics, Volume 34, Issue 13, 01 July 2018, Pages 2161\u20132167, https://doi.org/10.1093/bioinformatics/bty095 PhiSpy Sajia Akhter, Ramy K. Aziz, Robert A. Edwards; PhiSpy: a novel algorithm for finding prophages in bacterial genomes that combines similarity- and composition-based strategies. Nucl Acids Res 2012; 40 (16): e126. doi: 10.1093/nar/gks406 EvolCCM Chaoyue Liu and others, The Community Coevolution Model with Application to the Study of Evolutionary Relationships between Genes Based on Phylogenetic Profiles, Systematic Biology, Volume 72, Issue 3, May 2023, Pages 559\u2013574, https://doi.org/10.1093/sysbio/syac052 rSPR Christopher Whidden, Norbert Zeh, Robert G. Beiko, Supertrees Based on the Subtree Prune-and-Regraft Distance, Systematic Biology, Volume 63, Issue 4, July 2014, Pages 566\u2013581, https://doi.org/10.1093/sysbio/syu023","title":"Pipeline tools"},{"location":"CITATIONS/#software-packagingcontainerisation-tools","text":"Anaconda Anaconda Software Distribution. Computer software. Vers. 2-2.4.0. Anaconda, Nov. 2016. Web. Bioconda Gr\u00fcning B, Dale R, Sj\u00f6din A, Chapman BA, Rowe J, Tomkins-Tinch CH, Valieris R, K\u00f6ster J; Bioconda Team. Bioconda: sustainable and comprehensive software distribution for the life sciences. Nat Methods. 2018 Jul;15(7):475-476. doi: 10.1038/s41592-018-0046-7. PubMed PMID: 29967506. BioContainers da Veiga Leprevost F, Gr\u00fcning B, Aflitos SA, R\u00f6st HL, Uszkoreit J, Barsnes H, Vaudel M, Moreno P, Gatto L, Weber J, Bai M, Jimenez RC, Sachsenberg T, Pfeuffer J, Alvarez RV, Griss J, Nesvizhskii AI, Perez-Riverol Y. BioContainers: an open-source and community-driven framework for software standardization. Bioinformatics. 2017 Aug 15;33(16):2580-2582. doi: 10.1093/bioinformatics/btx192. PubMed PMID: 28379341; PubMed Central PMCID: PMC5870671. Docker Dirk Merkel. 2014. Docker: lightweight Linux containers for consistent development and deployment. Linux J. 2014, 239, Article 2 (March 2014). Singularity Kurtzer GM, Sochat V, Bauer MW. Singularity: Scientific containers for mobility of compute. PLoS One. 2017 May 11;12(5):e0177459. doi: 10.1371/journal.pone.0177459. eCollection 2017. PubMed PMID: 28494014; PubMed Central PMCID: PMC5426675.","title":"Software packaging/containerisation tools"},{"location":"ROADMAP/","text":"A list in no particular order of outstanding development features, both in-progress and planned: Integration of additional tools and scripts: Partner applications for analysis and visualization of phylogenetic distributions of genes and MGEs and gene-order clustering (For example, Coeus ).","title":"Roadmap"},{"location":"faq/","text":"Frequently Asked Questions How do I run ARETE in a Slurm HPC environment? Set a config file under ~/.nextflow/config to use the slurm executor: process { executor = 'slurm' pollInterval = '60 sec' submitRateLimit = '60/1min' queueSize = 100 // If an account is necessary: clusterOptions = '--account=' } See the Nextflow documentation for a description of these options. Now, when running ARETE, you'll need to set additional options if your compute nodes don't have network access - as is common for most Slurm clusters. The example below uses the default test data, i.e. the test profile, for demonstration purposes only. nextflow run beiko-lab/ARETE \\ --db_cache path/to/db_cache \\ --bakta_db path/to/baktadb \\ -profile test,singularity Apart from -profile singularity , which just makes ARETE use Singularity/Apptainer containers for running the tools, there are two additional parameters: --db_cache should be the location for the pre-downloaded databases used in the DIAMOND alignments (i.e. Bacmet, VFDB, ICEberg2 and CAZy FASTA files) and in the Kraken2 taxonomic read classification. Although these tools run by default, you can change the selection of annotation tools by changing --annotation_tools and skip Kraken2 by adding --skip_kraken . See the parameter documentation for a full list of parameters and their defaults. --bakta_db should be the location of the pre-downloaded Bakta database Alternatively, you can use Prokka for annotating your assemblies, since it doesn't require a downloaded database ( --use_prokka ). Do note that there could be memory-related issues when running Nextflow in SLURM environments. Can I use the ARETE outputs in MicroReact? Yes you can! In fact, ARETE provides many outputs that can be used in the MicroReact web app. Some of these files are: The PopPUNK lineages tree under poppunk_results/poppunk_visualizations/poppunk_visualizations.microreact . The reference tree built with FastTree under phylogenomics/reference_tree/core_gene_alignment.tre . The annotation feature profile annotation/feature_profile.tsv.gz . This file contains the annotation features in a presence/absence matrix format. Since MicroReact doesn't allow compressed files, just make sure to decompress it before-hand: gunzip feature_profile.tsv.gz Make sure to check our output documentation for a full list of outputs and the parameter documentation for a description of parameters to enable and disable these outputs.","title":"FAQ"},{"location":"faq/#frequently-asked-questions","text":"","title":"Frequently Asked Questions"},{"location":"faq/#how-do-i-run-arete-in-a-slurm-hpc-environment","text":"Set a config file under ~/.nextflow/config to use the slurm executor: process { executor = 'slurm' pollInterval = '60 sec' submitRateLimit = '60/1min' queueSize = 100 // If an account is necessary: clusterOptions = '--account=' } See the Nextflow documentation for a description of these options. Now, when running ARETE, you'll need to set additional options if your compute nodes don't have network access - as is common for most Slurm clusters. The example below uses the default test data, i.e. the test profile, for demonstration purposes only. nextflow run beiko-lab/ARETE \\ --db_cache path/to/db_cache \\ --bakta_db path/to/baktadb \\ -profile test,singularity Apart from -profile singularity , which just makes ARETE use Singularity/Apptainer containers for running the tools, there are two additional parameters: --db_cache should be the location for the pre-downloaded databases used in the DIAMOND alignments (i.e. Bacmet, VFDB, ICEberg2 and CAZy FASTA files) and in the Kraken2 taxonomic read classification. Although these tools run by default, you can change the selection of annotation tools by changing --annotation_tools and skip Kraken2 by adding --skip_kraken . See the parameter documentation for a full list of parameters and their defaults. --bakta_db should be the location of the pre-downloaded Bakta database Alternatively, you can use Prokka for annotating your assemblies, since it doesn't require a downloaded database ( --use_prokka ). Do note that there could be memory-related issues when running Nextflow in SLURM environments.","title":"How do I run ARETE in a Slurm HPC environment?"},{"location":"faq/#can-i-use-the-arete-outputs-in-microreact","text":"Yes you can! In fact, ARETE provides many outputs that can be used in the MicroReact web app. Some of these files are: The PopPUNK lineages tree under poppunk_results/poppunk_visualizations/poppunk_visualizations.microreact . The reference tree built with FastTree under phylogenomics/reference_tree/core_gene_alignment.tre . The annotation feature profile annotation/feature_profile.tsv.gz . This file contains the annotation features in a presence/absence matrix format. Since MicroReact doesn't allow compressed files, just make sure to decompress it before-hand: gunzip feature_profile.tsv.gz Make sure to check our output documentation for a full list of outputs and the parameter documentation for a description of parameters to enable and disable these outputs.","title":"Can I use the ARETE outputs in MicroReact?"},{"location":"output/","text":"beiko-lab/ARETE: Output Introduction The directories listed below will be created in the results directory after the pipeline has finished. All paths are relative to the top-level results directory. Pipeline overview The pipeline is built using Nextflow and processes data using the following steps (steps in italics don't run by default): Short-read processing and assembly FastQC - Raw and trimmed read QC FastP - Read trimming Kraken2 - Taxonomic assignment Unicycler - Short read assembly Quast - Assembly quality score Annotation Bakta or Prokka - Gene detection and annotation MobRecon - Reconstruction and typing of plasmids RGI - Detection and annotation of AMR determinants IslandPath - Predicts genomic islands in bacterial and archaeal genomes. PhiSpy - Prediction of prophages from bacterial genomes IntegronFinder - Finds integrons in DNA sequences Diamond - Detection and annotation of genes using external databases. CAZy: Carbohydrate metabolism VFDB: Virulence factors BacMet: Metal resistance determinants ICEberg: Integrative and conjugative elements PopPUNK Subworkflow PopPUNK - Genome clustering Dynamics EvolCCM - Community Coevolution rSPR - rooted subtree-prune-and-regraft distances Recombination Verticall - Conduct pairwise assembly comparisons between genomes in a same PopPUNK cluster SKA2 - Generate a whole-genome FASTA alignment for each genome within a cluster. Gubbins - Detection of recombination events within genomes of the same cluster. Gene Order Phylogenomics and Pangenomics Panaroo or PPanGGoLiN - Pangenome alignment FastTree or IQTree - Maximum likelihood core genome phylogenetic tree SNPsites - Extracts SNPs from a multi-FASTA alignment Pipeline information Report metrics generated during the workflow execution MultiQC - Aggregate report describing results and QC from the whole pipeline Assembly FastQC read_processing/*_fastqc/ *_fastqc.html : FastQC report containing quality metrics for your untrimmed raw fastq files. *_fastqc.zip : Zip archive containing the FastQC report, tab-delimited data file and plot images. NB: The FastQC plots in this directory are generated relative to the raw, input reads. They may contain adapter sequence and regions of low quality. To see how your reads look after adapter and quality trimming please refer to the FastQC reports in the trimgalore/fastqc/ directory. FastQC gives general quality metrics about your sequenced reads. It provides information about the quality score distribution across your reads, per base sequence content (%A/T/G/C), adapter contamination and overrepresented sequences. For further reading and documentation see the FastQC help pages . NB: The FastQC plots displayed in the MultiQC report shows untrimmed reads. They may contain adapter sequence and potentially regions with low quality. fastp read_processing/fastp/ ${meta.id} : Trimmed files and trimming reports for each input sample. fastp is a all-in-one fastq preprocessor for read/adapter trimming and quality control. It is used in this pipeline for trimming adapter sequences and discard low-quality reads. Kraken2 read_processing/kraken2/ *.kraken2.report.txt : Text file containing genome-wise information of Kraken2 findings. See here for details. *.classified(_(1|2))?.fastq.gz : Fasta file containing classified reads. If paired-end, one file per end. *.unclassified(_(1|2))?.fastq.gz : Fasta file containing unclassified reads. If paired-end, one file per end. Kraken2 is a read classification software which will assign taxonomy to each read comprising a sample. These results may be analyzed as an indicator of contamination. Unicycler assembly/unicycler/ *.assembly.gfa *.scaffolds.fa *.unicycler.log Short/hybrid read assembler. For now only handles short reads in ARETE. Quast assembly/quast/ report.tsv : A tab-seperated report compiling all QC metrics recorded over all genomes quast/ report.(html|tex|pdf|tsv|txt) : The Quast report in different file formats transposed_report.(tsv|txt) : Transpose of the Quast report quast.log : Log file of all Quast runs icarus_viewers/ contig_size_viewer.html basic_stats/ : Directory containing various summary plots generated by Quast. Annotation Bakta annotation/bakta/ ${sample_id}/ : Bakta results will be in one directory per genome. ${sample_id}.tsv : annotations as simple human readble TSV ${sample_id}.gff3 : annotations & sequences in GFF3 format ${sample_id}.gbff : annotations & sequences in (multi) GenBank format ${sample_id}.embl : annotations & sequences in (multi) EMBL format ${sample_id}.fna : replicon/contig DNA sequences as FASTA ${sample_id}.ffn : feature nucleotide sequences as FASTA ${sample_id}.faa : CDS/sORF amino acid sequences as FASTA ${sample_id}.hypotheticals.tsv : further information on hypothetical protein CDS as simple human readble tab separated values ${sample_id}.hypotheticals.faa : hypothetical protein CDS amino acid sequences as FASTA ${sample_id}.txt : summary as TXT ${sample_id}.png : circular genome annotation plot as PNG ${sample_id}.svg : circular genome annotation plot as SVG Bakta is a tool for the rapid & standardized annotation of bacterial genomes and plasmids from both isolates and MAGs Prokka annotation/prokka/ ${sample_id}/ : Prokka results will be in one directory per genome. ${sample_id}.err : Unacceptable annotations ${sample_id}.faa : Protein FASTA file of translated CDS sequences ${sample_id}.ffn : Nucleotide FASTA file of all the prediction transcripts (CDS, rRNA, tRNA, tmRNA, misc_RNA) ${sample_id}.fna : Nucleotide FASTA file of input contig sequences ${sample_id}.fsa : Nucleotide FASTA file of the input contig sequences, used by \"tbl2asn\" to create the .sqn file. It is mostly the same as the .fna file, but with extra Sequin tags in the sequence description lines. ${sample_id}.gff : This is the master annotation in GFF3 format, containing both sequences and annotations. ${sample_id}.gbk : This is a standard Genbank file derived from the master .gff. ${sample_id}.log : Contains all the output that Prokka produced during its run. This is a record of what settings used, even if the --quiet option was enabled. ${sample_id}.sqn : An ASN1 format \"Sequin\" file for submission to Genbank. It needs to be edited to set the correct taxonomy, authors, related publication etc. ${sample_id}.tbl : Feature Table file, used by \"tbl2asn\" to create the .sqn file. ${sample_id}.tsv : Tab-separated file of all features: locus_tag,ftype,len_bp,gene,EC_number,COG,product ${sample_id}.txt : Statistics relating to the annotated features found. Prokka is a software tool to annotate bacterial, archaeal and viral genomes quickly and produce standards-compliant output files. RGI annotation/rgi/ ${sample_id}_rgi.txt : A TSV report containing all AMR predictions for a given genome. For more info see here RGI predicts AMR determinants using the CARD ontology and various trained models. MobRecon annotation/mob_recon ${sample_id}_mob_recon/ : MobRecon results will be in one directory per genome. contig_report.txt - This file describes the assignment of the contig to chromosome or a particular plasmid grouping. mge.report.txt - Blast HSP of detected MGE's/repetitive elements with contextual information. chromosome.fasta - Fasta file of all contigs found to belong to the chromosome. plasmid_*.fasta - Each plasmid group is written to an individual fasta file which contains the assigned contigs. mobtyper_results - Aggregate MOB-typer report files for all identified plasmid. MobRecon reconstructs individual plasmid sequences from draft genome assemblies using the clustered plasmid reference databases DIAMOND annotation/(vfdb|bacmet|cazy|iceberg2)/ ${sample_id}/${sample_id}_(VFDB|BACMET|CAZYDB|ICEberg2).txt : Blast6 formatted TSVs indicating BlastX results of the genes from each genome against VFDB, BacMet, and CAZy databases. (VFDB|BACMET|CAZYDB|ICEberg2).txt : Table with all hits to this database, with a column describing which genome the match originates from. Sorted and filtered by the match's coverage. Diamond is a sequence aligner for protein and translated DNA searches, designed for high performance analysis of big sequence data. We use DIAMOND to predict the presence of virulence factors, heavy metal resistance determinants, carbohydrate-active enzymes, and integrative and conjugative elements using VFDB , BacMet , CAZy , and ICEberg2 respectively. IslandPath annotation/islandpath/ ${sample_id}/ : IslandPath results will be in one directory per genome. ${sample_id}.tsv : IslandPath results Dimob.log : IslandPath execution log IslandPath is a standalone software to predict genomic islands in bacterial and archaeal genomes based on the presence of dinucleotide biases and mobility genes. IntegronFinder Disabled by default. Enable by adding --run_integronfinder to your command. annotation/integron_finder/ Results_Integron_Finder_${sample_id}/ : IntegronFinder results will be in one directory per genome. Integron Finder is a bioinformatics tool to find integrons in bacterial genomes. PhiSpy annotation/phispy/ ${sample_id}/ : PhiSpy results will be in one directory per genome. See the PhiSpy documentation for an extensive description of the output. PhiSpy is a tool for identification of prophages in Bacterial (and probably Archaeal) genomes. Given an annotated genome it will use several approaches to identify the most likely prophage regions. PopPUNK poppunk_results/ poppunk_db/ - Results from PopPUNK's create-db command poppunk_${poppunk_model}/ - Results from PopPUNK's fit-model command poppunk_visualizations/ - Results from the poppunk_visualise command PopPUNK is a tool for clustering genomes. Phylogenomics and Pangenomics Panaroo pangenomics/panaroo/results/ See the panaroo documentation for an extensive description of output provided. Panaroo is a Bacterial Pangenome Analysis Pipeline. PPanGGoLiN pangenomics/ppanggolin/ See the PPanGGoLiN documentation for an extensive description of output provided. PPanGGoLiN is a tool to build a partitioned pangenome graph from microbial genomes FastTree phylogenomics/fasttree/ *.tre : Newick formatted maximum likelihood tree of core-genome alignment. FastTree infers approximately-maximum-likelihood phylogenetic trees from alignments of nucleotide or protein sequences IQTree phylogenomics/iqtree/ *.treefile : Newick formatted maximum likelihood tree of core-genome alignment. IQTree is a fast and effective stochastic algorithm to infer phylogenetic trees by maximum likelihood. SNPsites phylogenomics/snpsites/ filtered_alignment.fas : Variant fasta file. constant.sites.txt : Text file containing counts of constant sites. SNPsites is a tool to rapidly extract SNPs from a multi-FASTA alignment. Dynamics EvolCCM dynamics/EvolCCM/ EvolCCM_*tsv EvolCCM_*pvals EvolCCM_*X2 EvolCCM_*tre EvolCCM is the R implementation for CCM (Community Coevolution Model) rSPR The outputs are approximate and exact Subtree Prune and Regraft (rSPR) distances between pairs of rooted phylogenetic trees. Each CSV file contains these distances and the tree sizes. The PNG files are heatmaps of these distances and their respective tree sizes. dynamics/rSPR/ approx - Approximate rSPR distances exact - Exact rSPR distances rSPR is a software package for calculating rooted subtree-prune-and-regraft distances and rooted agreement forests. Recombination Verticall dynamics/recombination/verticall/ verticall_cluster*.tsv - Verticall results for the genomes within this PopPUNK cluster. Verticall is a tool to help produce bacterial genome phylogenies which are not influenced by horizontally acquired sequences such as recombination. SKA2 dynamics/recombination/ska2/ cluster_*.aln - SKA2 results for the genomes within this PopPUNK cluster. SKA2 (Split Kmer Analysis) is a toolkit for prokaryotic (and any other small, haploid) DNA sequence analysis using split kmers. Gubbins dynamics/recombination/gubbins/ cluster_*/ - Gubbins results for the genomes within this PopPUNK cluster. Gubbins is an algorithm that iteratively identifies loci containing elevated densities of base substitutions while concurrently constructing a phylogeny based on the putative point mutations outside of these regions. Gene Order gene-order/ extraction/ - AMR genes of interest and their neighborhoods extracted from the assemblies. diamond/ - Pairwise alignments between all input genomes. clustering/ - Similarity and distance matrices for each AMR gene clustered via UPGMA, MCL and DBSCAN to identify similarities between their neighborhoods across all genomes. Gene Order is a subworkflow for bacterial gene order analysis, with outputs easily explorable through its partner visualization application Coeus . Pipeline information pipeline_info/ Reports generated by Nextflow: execution_report.html , execution_timeline.html , execution_trace.txt and pipeline_dag.dot / pipeline_dag.svg . Reports generated by the pipeline: pipeline_report.html , pipeline_report.txt and software_versions.csv . Reformatted samplesheet files used as input to the pipeline: samplesheet.valid.csv . Nextflow provides excellent functionality for generating various reports relevant to the running and execution of the pipeline. This will allow you to troubleshoot errors with the running of the pipeline, and also provide you with other information such as launch commands, run times and resource usage. MultiQC multiqc/ multiqc_report.html : a standalone HTML file that can be viewed in your web browser. multiqc_data/ : directory containing parsed statistics from the different tools used in the pipeline. multiqc_plots/ : directory containing static images from the report in various formats. MultiQC is a visualization tool that generates a single HTML report summarising all samples in your project. Most of the pipeline QC results are visualised in the report and further statistics are available in the report data directory. Results generated by MultiQC collate pipeline QC from supported tools e.g. FastQC. The pipeline has special steps which also allow the software versions to be reported in the MultiQC output for future traceability. For more information about how to use MultiQC reports, see http://multiqc.info .","title":"Output"},{"location":"output/#beiko-labarete-output","text":"","title":"beiko-lab/ARETE: Output"},{"location":"output/#introduction","text":"The directories listed below will be created in the results directory after the pipeline has finished. All paths are relative to the top-level results directory.","title":"Introduction"},{"location":"output/#pipeline-overview","text":"The pipeline is built using Nextflow and processes data using the following steps (steps in italics don't run by default): Short-read processing and assembly FastQC - Raw and trimmed read QC FastP - Read trimming Kraken2 - Taxonomic assignment Unicycler - Short read assembly Quast - Assembly quality score Annotation Bakta or Prokka - Gene detection and annotation MobRecon - Reconstruction and typing of plasmids RGI - Detection and annotation of AMR determinants IslandPath - Predicts genomic islands in bacterial and archaeal genomes. PhiSpy - Prediction of prophages from bacterial genomes IntegronFinder - Finds integrons in DNA sequences Diamond - Detection and annotation of genes using external databases. CAZy: Carbohydrate metabolism VFDB: Virulence factors BacMet: Metal resistance determinants ICEberg: Integrative and conjugative elements PopPUNK Subworkflow PopPUNK - Genome clustering Dynamics EvolCCM - Community Coevolution rSPR - rooted subtree-prune-and-regraft distances Recombination Verticall - Conduct pairwise assembly comparisons between genomes in a same PopPUNK cluster SKA2 - Generate a whole-genome FASTA alignment for each genome within a cluster. Gubbins - Detection of recombination events within genomes of the same cluster. Gene Order Phylogenomics and Pangenomics Panaroo or PPanGGoLiN - Pangenome alignment FastTree or IQTree - Maximum likelihood core genome phylogenetic tree SNPsites - Extracts SNPs from a multi-FASTA alignment Pipeline information Report metrics generated during the workflow execution MultiQC - Aggregate report describing results and QC from the whole pipeline","title":"Pipeline overview"},{"location":"output/#assembly","text":"","title":"Assembly"},{"location":"output/#fastqc","text":"read_processing/*_fastqc/ *_fastqc.html : FastQC report containing quality metrics for your untrimmed raw fastq files. *_fastqc.zip : Zip archive containing the FastQC report, tab-delimited data file and plot images. NB: The FastQC plots in this directory are generated relative to the raw, input reads. They may contain adapter sequence and regions of low quality. To see how your reads look after adapter and quality trimming please refer to the FastQC reports in the trimgalore/fastqc/ directory. FastQC gives general quality metrics about your sequenced reads. It provides information about the quality score distribution across your reads, per base sequence content (%A/T/G/C), adapter contamination and overrepresented sequences. For further reading and documentation see the FastQC help pages . NB: The FastQC plots displayed in the MultiQC report shows untrimmed reads. They may contain adapter sequence and potentially regions with low quality.","title":"FastQC"},{"location":"output/#fastp","text":"read_processing/fastp/ ${meta.id} : Trimmed files and trimming reports for each input sample. fastp is a all-in-one fastq preprocessor for read/adapter trimming and quality control. It is used in this pipeline for trimming adapter sequences and discard low-quality reads.","title":"fastp"},{"location":"output/#kraken2","text":"read_processing/kraken2/ *.kraken2.report.txt : Text file containing genome-wise information of Kraken2 findings. See here for details. *.classified(_(1|2))?.fastq.gz : Fasta file containing classified reads. If paired-end, one file per end. *.unclassified(_(1|2))?.fastq.gz : Fasta file containing unclassified reads. If paired-end, one file per end. Kraken2 is a read classification software which will assign taxonomy to each read comprising a sample. These results may be analyzed as an indicator of contamination.","title":"Kraken2"},{"location":"output/#unicycler","text":"assembly/unicycler/ *.assembly.gfa *.scaffolds.fa *.unicycler.log Short/hybrid read assembler. For now only handles short reads in ARETE.","title":"Unicycler"},{"location":"output/#quast","text":"assembly/quast/ report.tsv : A tab-seperated report compiling all QC metrics recorded over all genomes quast/ report.(html|tex|pdf|tsv|txt) : The Quast report in different file formats transposed_report.(tsv|txt) : Transpose of the Quast report quast.log : Log file of all Quast runs icarus_viewers/ contig_size_viewer.html basic_stats/ : Directory containing various summary plots generated by Quast.","title":"Quast"},{"location":"output/#annotation","text":"","title":"Annotation"},{"location":"output/#bakta","text":"annotation/bakta/ ${sample_id}/ : Bakta results will be in one directory per genome. ${sample_id}.tsv : annotations as simple human readble TSV ${sample_id}.gff3 : annotations & sequences in GFF3 format ${sample_id}.gbff : annotations & sequences in (multi) GenBank format ${sample_id}.embl : annotations & sequences in (multi) EMBL format ${sample_id}.fna : replicon/contig DNA sequences as FASTA ${sample_id}.ffn : feature nucleotide sequences as FASTA ${sample_id}.faa : CDS/sORF amino acid sequences as FASTA ${sample_id}.hypotheticals.tsv : further information on hypothetical protein CDS as simple human readble tab separated values ${sample_id}.hypotheticals.faa : hypothetical protein CDS amino acid sequences as FASTA ${sample_id}.txt : summary as TXT ${sample_id}.png : circular genome annotation plot as PNG ${sample_id}.svg : circular genome annotation plot as SVG Bakta is a tool for the rapid & standardized annotation of bacterial genomes and plasmids from both isolates and MAGs","title":"Bakta"},{"location":"output/#prokka","text":"annotation/prokka/ ${sample_id}/ : Prokka results will be in one directory per genome. ${sample_id}.err : Unacceptable annotations ${sample_id}.faa : Protein FASTA file of translated CDS sequences ${sample_id}.ffn : Nucleotide FASTA file of all the prediction transcripts (CDS, rRNA, tRNA, tmRNA, misc_RNA) ${sample_id}.fna : Nucleotide FASTA file of input contig sequences ${sample_id}.fsa : Nucleotide FASTA file of the input contig sequences, used by \"tbl2asn\" to create the .sqn file. It is mostly the same as the .fna file, but with extra Sequin tags in the sequence description lines. ${sample_id}.gff : This is the master annotation in GFF3 format, containing both sequences and annotations. ${sample_id}.gbk : This is a standard Genbank file derived from the master .gff. ${sample_id}.log : Contains all the output that Prokka produced during its run. This is a record of what settings used, even if the --quiet option was enabled. ${sample_id}.sqn : An ASN1 format \"Sequin\" file for submission to Genbank. It needs to be edited to set the correct taxonomy, authors, related publication etc. ${sample_id}.tbl : Feature Table file, used by \"tbl2asn\" to create the .sqn file. ${sample_id}.tsv : Tab-separated file of all features: locus_tag,ftype,len_bp,gene,EC_number,COG,product ${sample_id}.txt : Statistics relating to the annotated features found. Prokka is a software tool to annotate bacterial, archaeal and viral genomes quickly and produce standards-compliant output files.","title":"Prokka"},{"location":"output/#rgi","text":"annotation/rgi/ ${sample_id}_rgi.txt : A TSV report containing all AMR predictions for a given genome. For more info see here RGI predicts AMR determinants using the CARD ontology and various trained models.","title":"RGI"},{"location":"output/#mobrecon","text":"annotation/mob_recon ${sample_id}_mob_recon/ : MobRecon results will be in one directory per genome. contig_report.txt - This file describes the assignment of the contig to chromosome or a particular plasmid grouping. mge.report.txt - Blast HSP of detected MGE's/repetitive elements with contextual information. chromosome.fasta - Fasta file of all contigs found to belong to the chromosome. plasmid_*.fasta - Each plasmid group is written to an individual fasta file which contains the assigned contigs. mobtyper_results - Aggregate MOB-typer report files for all identified plasmid. MobRecon reconstructs individual plasmid sequences from draft genome assemblies using the clustered plasmid reference databases","title":"MobRecon"},{"location":"output/#diamond","text":"annotation/(vfdb|bacmet|cazy|iceberg2)/ ${sample_id}/${sample_id}_(VFDB|BACMET|CAZYDB|ICEberg2).txt : Blast6 formatted TSVs indicating BlastX results of the genes from each genome against VFDB, BacMet, and CAZy databases. (VFDB|BACMET|CAZYDB|ICEberg2).txt : Table with all hits to this database, with a column describing which genome the match originates from. Sorted and filtered by the match's coverage. Diamond is a sequence aligner for protein and translated DNA searches, designed for high performance analysis of big sequence data. We use DIAMOND to predict the presence of virulence factors, heavy metal resistance determinants, carbohydrate-active enzymes, and integrative and conjugative elements using VFDB , BacMet , CAZy , and ICEberg2 respectively.","title":"DIAMOND"},{"location":"output/#islandpath","text":"annotation/islandpath/ ${sample_id}/ : IslandPath results will be in one directory per genome. ${sample_id}.tsv : IslandPath results Dimob.log : IslandPath execution log IslandPath is a standalone software to predict genomic islands in bacterial and archaeal genomes based on the presence of dinucleotide biases and mobility genes.","title":"IslandPath"},{"location":"output/#integronfinder","text":"Disabled by default. Enable by adding --run_integronfinder to your command. annotation/integron_finder/ Results_Integron_Finder_${sample_id}/ : IntegronFinder results will be in one directory per genome. Integron Finder is a bioinformatics tool to find integrons in bacterial genomes.","title":"IntegronFinder"},{"location":"output/#phispy","text":"annotation/phispy/ ${sample_id}/ : PhiSpy results will be in one directory per genome. See the PhiSpy documentation for an extensive description of the output. PhiSpy is a tool for identification of prophages in Bacterial (and probably Archaeal) genomes. Given an annotated genome it will use several approaches to identify the most likely prophage regions.","title":"PhiSpy"},{"location":"output/#poppunk","text":"poppunk_results/ poppunk_db/ - Results from PopPUNK's create-db command poppunk_${poppunk_model}/ - Results from PopPUNK's fit-model command poppunk_visualizations/ - Results from the poppunk_visualise command PopPUNK is a tool for clustering genomes.","title":"PopPUNK"},{"location":"output/#phylogenomics-and-pangenomics","text":"","title":"Phylogenomics and Pangenomics"},{"location":"output/#panaroo","text":"pangenomics/panaroo/results/ See the panaroo documentation for an extensive description of output provided. Panaroo is a Bacterial Pangenome Analysis Pipeline.","title":"Panaroo"},{"location":"output/#ppanggolin","text":"pangenomics/ppanggolin/ See the PPanGGoLiN documentation for an extensive description of output provided. PPanGGoLiN is a tool to build a partitioned pangenome graph from microbial genomes","title":"PPanGGoLiN"},{"location":"output/#fasttree","text":"phylogenomics/fasttree/ *.tre : Newick formatted maximum likelihood tree of core-genome alignment. FastTree infers approximately-maximum-likelihood phylogenetic trees from alignments of nucleotide or protein sequences","title":"FastTree"},{"location":"output/#iqtree","text":"phylogenomics/iqtree/ *.treefile : Newick formatted maximum likelihood tree of core-genome alignment. IQTree is a fast and effective stochastic algorithm to infer phylogenetic trees by maximum likelihood.","title":"IQTree"},{"location":"output/#snpsites","text":"phylogenomics/snpsites/ filtered_alignment.fas : Variant fasta file. constant.sites.txt : Text file containing counts of constant sites. SNPsites is a tool to rapidly extract SNPs from a multi-FASTA alignment.","title":"SNPsites"},{"location":"output/#dynamics","text":"","title":"Dynamics"},{"location":"output/#evolccm","text":"dynamics/EvolCCM/ EvolCCM_*tsv EvolCCM_*pvals EvolCCM_*X2 EvolCCM_*tre EvolCCM is the R implementation for CCM (Community Coevolution Model)","title":"EvolCCM"},{"location":"output/#rspr","text":"The outputs are approximate and exact Subtree Prune and Regraft (rSPR) distances between pairs of rooted phylogenetic trees. Each CSV file contains these distances and the tree sizes. The PNG files are heatmaps of these distances and their respective tree sizes. dynamics/rSPR/ approx - Approximate rSPR distances exact - Exact rSPR distances rSPR is a software package for calculating rooted subtree-prune-and-regraft distances and rooted agreement forests.","title":"rSPR"},{"location":"output/#recombination","text":"","title":"Recombination"},{"location":"output/#verticall","text":"dynamics/recombination/verticall/ verticall_cluster*.tsv - Verticall results for the genomes within this PopPUNK cluster. Verticall is a tool to help produce bacterial genome phylogenies which are not influenced by horizontally acquired sequences such as recombination.","title":"Verticall"},{"location":"output/#ska2","text":"dynamics/recombination/ska2/ cluster_*.aln - SKA2 results for the genomes within this PopPUNK cluster. SKA2 (Split Kmer Analysis) is a toolkit for prokaryotic (and any other small, haploid) DNA sequence analysis using split kmers.","title":"SKA2"},{"location":"output/#gubbins","text":"dynamics/recombination/gubbins/ cluster_*/ - Gubbins results for the genomes within this PopPUNK cluster. Gubbins is an algorithm that iteratively identifies loci containing elevated densities of base substitutions while concurrently constructing a phylogeny based on the putative point mutations outside of these regions.","title":"Gubbins"},{"location":"output/#gene-order","text":"gene-order/ extraction/ - AMR genes of interest and their neighborhoods extracted from the assemblies. diamond/ - Pairwise alignments between all input genomes. clustering/ - Similarity and distance matrices for each AMR gene clustered via UPGMA, MCL and DBSCAN to identify similarities between their neighborhoods across all genomes. Gene Order is a subworkflow for bacterial gene order analysis, with outputs easily explorable through its partner visualization application Coeus .","title":"Gene Order"},{"location":"output/#pipeline-information","text":"pipeline_info/ Reports generated by Nextflow: execution_report.html , execution_timeline.html , execution_trace.txt and pipeline_dag.dot / pipeline_dag.svg . Reports generated by the pipeline: pipeline_report.html , pipeline_report.txt and software_versions.csv . Reformatted samplesheet files used as input to the pipeline: samplesheet.valid.csv . Nextflow provides excellent functionality for generating various reports relevant to the running and execution of the pipeline. This will allow you to troubleshoot errors with the running of the pipeline, and also provide you with other information such as launch commands, run times and resource usage.","title":"Pipeline information"},{"location":"output/#multiqc","text":"multiqc/ multiqc_report.html : a standalone HTML file that can be viewed in your web browser. multiqc_data/ : directory containing parsed statistics from the different tools used in the pipeline. multiqc_plots/ : directory containing static images from the report in various formats. MultiQC is a visualization tool that generates a single HTML report summarising all samples in your project. Most of the pipeline QC results are visualised in the report and further statistics are available in the report data directory. Results generated by MultiQC collate pipeline QC from supported tools e.g. FastQC. The pipeline has special steps which also allow the software versions to be reported in the MultiQC output for future traceability. For more information about how to use MultiQC reports, see http://multiqc.info .","title":"MultiQC"},{"location":"params/","text":"beiko-lab/ARETE pipeline parameters AMR/VF LGT-focused bacterial genomics workflow Input/output options Define where the pipeline should find input data and save output data. Parameter Description Type Default Required Hidden input_sample_table Path to comma-separated file containing information about the samples in the experiment. Help You will need to create a design file with information about the samples in your experiment before running the pipeline. Use this parameter to specify its location. It has to be a comma-separated file with 3 columns, and a header row. string outdir Path to the output directory where the results will be saved. string ./results db_cache Directory where the databases are located string email Email address for completion summary. Help Set this parameter to your e-mail address to get a summary e-mail with details of the run sent to you when the workflow exits. If set in your user config file ( ~/.nextflow/config ) then you don't need to specify this on the command line for every run. string multiqc_title MultiQC report title. Printed as page header, used for filename if not otherwise specified. string Reference genome options Reference and outgroup genome fasta files required for the workflow. Parameter Description Type Default Required Hidden reference_genome Path to FASTA reference genome file. string QC Parameter Description Type Default Required Hidden run_checkm Run CheckM QC software boolean apply_filtering Filter assemblies on QC results boolean skip_kraken Don't run Kraken2 taxonomic classification boolean min_n50 Minimum N50 for filtering integer 10000 min_contigs_1000_bp Minimum number of contigs with >1000bp integer 1 min_contig_length Minimum average contig length integer 1 Annotation Parameters for the annotation subworkflow Parameter Description Type Default Required Hidden annotation_tools Comma-separated list of annotation tools to run string mobsuite,rgi,cazy,vfdb,iceberg,bacmet,islandpath,phispy,report bakta_db Path to the BAKTA database string use_prokka Use Prokka (not Bakta) for annotating assemblies boolean min_pident Minimum match identity percentage for filtering integer 60 min_qcover Minimum coverage of each match for filtering number 0.6 skip_profile_creation Skip annotation feature profile creation boolean feature_profile_columns Columns to include in the feature profile string mobsuite,rgi,cazy,vfdb,iceberg,bacmet Phylogenomics Parameters for the phylogenomics subworkflow Parameter Description Type Default Required Hidden skip_phylo Skip Pangenomics and Phylogenomics subworkflow boolean use_ppanggolin Use ppanggolin for calculating the pangenome boolean use_full_alignment Use full alignment boolean use_fasttree Use FastTree boolean True PopPUNK Parameters for the lineage subworkflow Parameter Description Type Default Required Hidden skip_poppunk Skip PopPunk boolean poppunk_model Which PopPunk model to use (bgmm, dbscan, refine, threshold or lineage) string run_poppunk_qc Whether to run the QC step for PopPunk boolean enable_subsetting Enable subsetting workflow based on genome similarity boolean core_similarity Similarity threshold for core genomes number 99.99 accessory_similarity Similarity threshold for accessory genes number 99.0 Gene Order Parameters for the Gene Order Subworkflow Parameter Description Type Default Required Hidden run_gene_order Whether to run the Gene Order subworkflow boolean gene_order_percent_cutoff Cutoff percentage of genomes a gene should be present within to be included in extraction and subsequent analysis. Should a float between 0 and 1 (e.g., 0.25 means only genes present in a minimum of 25% of genomes are kept). number 0.25 gene_order_label_cols If using annotation files predicting features, list of space separated column names to be added to the gene names string None num_neighbors Neighborhood size to extract. Should be an even number N, such that N/2 neighbors upstream and N/2 neighbors downstream will be analyzed. integer 10 inflation Inflation hyperparameter value for Markov Clustering Algorithm. integer 2 epsilon Epsilon hyperparameter value for DBSCAN clustering. number 0.5 minpts Minpts hyperparameter value for DBSCAN clustering. integer 5 plot_clustering Create Clustering HTML Plots boolean Recombination Parameters for the recombination subworkflow Parameter Description Type Default Required Hidden run_recombination Run Recombination boolean run_verticall Run Verticall recombination tool boolean True run_gubbins Run Gubbins recombination tool boolean Dynamics Parameter Description Type Default Required Hidden run_evolccm Run the community coevolution model boolean run_rspr Run rSPR boolean min_rspr_distance Minimum rSPR distance used to define processing groups integer 10 min_branch_length Minimum rSPR branch length integer 0 max_support_threshold Maximum rSPR support threshold number 0.7 max_approx_rspr Maximum approximate rSPR distance for filtering integer -1 core_gene_tree Core (or reference) genome tree. Used in the rSPR and evolCCM entries. string concatenated_annotation TSV table of annotations for all genomes. Such as the ones generated by Bakta or Prokka in ARETE. string feature_profile Feature profile TSV (A presence-absence matrix). Used in the evolCCM entry. string Institutional config options Parameters used to describe centralised config profiles. These should not be edited. Parameter Description Type Default Required Hidden custom_config_version Git commit id for Institutional configs. string master True custom_config_base Base directory for Institutional configs. Help If you're running offline, Nextflow will not be able to fetch the institutional config files from the internet. If you don't need them, then this is not a problem. If you do need them, you should download the files from the repo and tell Nextflow where to find them with this parameter. string https://raw.githubusercontent.com/nf-core/configs/master True hostnames Institutional configs hostname. string True config_profile_name Institutional config name. string True config_profile_description Institutional config description. string True config_profile_contact Institutional config contact information. string True config_profile_url Institutional config URL link. string True Max job request options Set the top limit for requested resources for any single job. Parameter Description Type Default Required Hidden max_cpus Maximum number of CPUs that can be requested for any single job. Help Use to set an upper-limit for the CPU requirement for each process. Should be an integer e.g. --max_cpus 1 integer 16 True max_memory Maximum amount of memory that can be requested for any single job. Help Use to set an upper-limit for the memory requirement for each process. Should be a string in the format integer-unit e.g. --max_memory '8.GB' string 128.GB True max_time Maximum amount of time that can be requested for any single job. Help Use to set an upper-limit for the time requirement for each process. Should be a string in the format integer-unit e.g. --max_time '2.h' string 240.h True Generic options Less common options for the pipeline, typically set in a config file. Parameter Description Type Default Required Hidden help Display help text. boolean True publish_dir_mode Method used to save pipeline results to output directory. Help The Nextflow publishDir option specifies which intermediate files should be saved to the output directory. This option tells the pipeline what method should be used to move these files. See Nextflow docs for details. string copy True email_on_fail Email address for completion summary, only when pipeline fails. Help An email address to send a summary email to when the pipeline is completed - ONLY sent if the pipeline does not exit successfully. string True plaintext_email Send plain-text email instead of HTML. boolean True max_multiqc_email_size File size limit when attaching MultiQC reports to summary emails. string 25.MB True monochrome_logs Do not use coloured log outputs. boolean True multiqc_config Custom config file to supply to MultiQC. string True tracedir Directory to keep pipeline Nextflow logs and reports. string ${params.outdir}/pipeline_info True validate_params Boolean whether to validate parameters against the schema at runtime boolean True True show_hidden_params Show all params when using --help Help By default, parameters set as hidden in the schema are not shown on the command line when a user runs with --help . Specifying this option will tell the pipeline to show all parameters. boolean True enable_conda Run this workflow with Conda. You can also use '-profile conda' instead of providing this parameter. boolean True singularity_pull_docker_container Instead of directly downloading Singularity images for use with Singularity, force the workflow to pull and convert Docker containers instead. Help This may be useful for example if you are unable to directly pull Singularity containers to run the pipeline due to http/https proxy issues. boolean True schema_ignore_params string genomes,modules multiqc_logo string True","title":"Parameters"},{"location":"params/#beiko-labarete-pipeline-parameters","text":"AMR/VF LGT-focused bacterial genomics workflow","title":"beiko-lab/ARETE pipeline parameters"},{"location":"params/#inputoutput-options","text":"Define where the pipeline should find input data and save output data. Parameter Description Type Default Required Hidden input_sample_table Path to comma-separated file containing information about the samples in the experiment. Help You will need to create a design file with information about the samples in your experiment before running the pipeline. Use this parameter to specify its location. It has to be a comma-separated file with 3 columns, and a header row. string outdir Path to the output directory where the results will be saved. string ./results db_cache Directory where the databases are located string email Email address for completion summary. Help Set this parameter to your e-mail address to get a summary e-mail with details of the run sent to you when the workflow exits. If set in your user config file ( ~/.nextflow/config ) then you don't need to specify this on the command line for every run. string multiqc_title MultiQC report title. Printed as page header, used for filename if not otherwise specified. string","title":"Input/output options"},{"location":"params/#reference-genome-options","text":"Reference and outgroup genome fasta files required for the workflow. Parameter Description Type Default Required Hidden reference_genome Path to FASTA reference genome file. string","title":"Reference genome options"},{"location":"params/#qc","text":"Parameter Description Type Default Required Hidden run_checkm Run CheckM QC software boolean apply_filtering Filter assemblies on QC results boolean skip_kraken Don't run Kraken2 taxonomic classification boolean min_n50 Minimum N50 for filtering integer 10000 min_contigs_1000_bp Minimum number of contigs with >1000bp integer 1 min_contig_length Minimum average contig length integer 1","title":"QC"},{"location":"params/#annotation","text":"Parameters for the annotation subworkflow Parameter Description Type Default Required Hidden annotation_tools Comma-separated list of annotation tools to run string mobsuite,rgi,cazy,vfdb,iceberg,bacmet,islandpath,phispy,report bakta_db Path to the BAKTA database string use_prokka Use Prokka (not Bakta) for annotating assemblies boolean min_pident Minimum match identity percentage for filtering integer 60 min_qcover Minimum coverage of each match for filtering number 0.6 skip_profile_creation Skip annotation feature profile creation boolean feature_profile_columns Columns to include in the feature profile string mobsuite,rgi,cazy,vfdb,iceberg,bacmet","title":"Annotation"},{"location":"params/#phylogenomics","text":"Parameters for the phylogenomics subworkflow Parameter Description Type Default Required Hidden skip_phylo Skip Pangenomics and Phylogenomics subworkflow boolean use_ppanggolin Use ppanggolin for calculating the pangenome boolean use_full_alignment Use full alignment boolean use_fasttree Use FastTree boolean True","title":"Phylogenomics"},{"location":"params/#poppunk","text":"Parameters for the lineage subworkflow Parameter Description Type Default Required Hidden skip_poppunk Skip PopPunk boolean poppunk_model Which PopPunk model to use (bgmm, dbscan, refine, threshold or lineage) string run_poppunk_qc Whether to run the QC step for PopPunk boolean enable_subsetting Enable subsetting workflow based on genome similarity boolean core_similarity Similarity threshold for core genomes number 99.99 accessory_similarity Similarity threshold for accessory genes number 99.0","title":"PopPUNK"},{"location":"params/#gene-order","text":"Parameters for the Gene Order Subworkflow Parameter Description Type Default Required Hidden run_gene_order Whether to run the Gene Order subworkflow boolean gene_order_percent_cutoff Cutoff percentage of genomes a gene should be present within to be included in extraction and subsequent analysis. Should a float between 0 and 1 (e.g., 0.25 means only genes present in a minimum of 25% of genomes are kept). number 0.25 gene_order_label_cols If using annotation files predicting features, list of space separated column names to be added to the gene names string None num_neighbors Neighborhood size to extract. Should be an even number N, such that N/2 neighbors upstream and N/2 neighbors downstream will be analyzed. integer 10 inflation Inflation hyperparameter value for Markov Clustering Algorithm. integer 2 epsilon Epsilon hyperparameter value for DBSCAN clustering. number 0.5 minpts Minpts hyperparameter value for DBSCAN clustering. integer 5 plot_clustering Create Clustering HTML Plots boolean","title":"Gene Order"},{"location":"params/#recombination","text":"Parameters for the recombination subworkflow Parameter Description Type Default Required Hidden run_recombination Run Recombination boolean run_verticall Run Verticall recombination tool boolean True run_gubbins Run Gubbins recombination tool boolean","title":"Recombination"},{"location":"params/#dynamics","text":"Parameter Description Type Default Required Hidden run_evolccm Run the community coevolution model boolean run_rspr Run rSPR boolean min_rspr_distance Minimum rSPR distance used to define processing groups integer 10 min_branch_length Minimum rSPR branch length integer 0 max_support_threshold Maximum rSPR support threshold number 0.7 max_approx_rspr Maximum approximate rSPR distance for filtering integer -1 core_gene_tree Core (or reference) genome tree. Used in the rSPR and evolCCM entries. string concatenated_annotation TSV table of annotations for all genomes. Such as the ones generated by Bakta or Prokka in ARETE. string feature_profile Feature profile TSV (A presence-absence matrix). Used in the evolCCM entry. string","title":"Dynamics"},{"location":"params/#institutional-config-options","text":"Parameters used to describe centralised config profiles. These should not be edited. Parameter Description Type Default Required Hidden custom_config_version Git commit id for Institutional configs. string master True custom_config_base Base directory for Institutional configs. Help If you're running offline, Nextflow will not be able to fetch the institutional config files from the internet. If you don't need them, then this is not a problem. If you do need them, you should download the files from the repo and tell Nextflow where to find them with this parameter. string https://raw.githubusercontent.com/nf-core/configs/master True hostnames Institutional configs hostname. string True config_profile_name Institutional config name. string True config_profile_description Institutional config description. string True config_profile_contact Institutional config contact information. string True config_profile_url Institutional config URL link. string True","title":"Institutional config options"},{"location":"params/#max-job-request-options","text":"Set the top limit for requested resources for any single job. Parameter Description Type Default Required Hidden max_cpus Maximum number of CPUs that can be requested for any single job. Help Use to set an upper-limit for the CPU requirement for each process. Should be an integer e.g. --max_cpus 1 integer 16 True max_memory Maximum amount of memory that can be requested for any single job. Help Use to set an upper-limit for the memory requirement for each process. Should be a string in the format integer-unit e.g. --max_memory '8.GB' string 128.GB True max_time Maximum amount of time that can be requested for any single job. Help Use to set an upper-limit for the time requirement for each process. Should be a string in the format integer-unit e.g. --max_time '2.h' string 240.h True","title":"Max job request options"},{"location":"params/#generic-options","text":"Less common options for the pipeline, typically set in a config file. Parameter Description Type Default Required Hidden help Display help text. boolean True publish_dir_mode Method used to save pipeline results to output directory. Help The Nextflow publishDir option specifies which intermediate files should be saved to the output directory. This option tells the pipeline what method should be used to move these files. See Nextflow docs for details. string copy True email_on_fail Email address for completion summary, only when pipeline fails. Help An email address to send a summary email to when the pipeline is completed - ONLY sent if the pipeline does not exit successfully. string True plaintext_email Send plain-text email instead of HTML. boolean True max_multiqc_email_size File size limit when attaching MultiQC reports to summary emails. string 25.MB True monochrome_logs Do not use coloured log outputs. boolean True multiqc_config Custom config file to supply to MultiQC. string True tracedir Directory to keep pipeline Nextflow logs and reports. string ${params.outdir}/pipeline_info True validate_params Boolean whether to validate parameters against the schema at runtime boolean True True show_hidden_params Show all params when using --help Help By default, parameters set as hidden in the schema are not shown on the command line when a user runs with --help . Specifying this option will tell the pipeline to show all parameters. boolean True enable_conda Run this workflow with Conda. You can also use '-profile conda' instead of providing this parameter. boolean True singularity_pull_docker_container Instead of directly downloading Singularity images for use with Singularity, force the workflow to pull and convert Docker containers instead. Help This may be useful for example if you are unable to directly pull Singularity containers to run the pipeline due to http/https proxy issues. boolean True schema_ignore_params string genomes,modules multiqc_logo string True","title":"Generic options"},{"location":"resource_profiles/","text":"ARETE and dataset size Currently ARETE has three distinct profiles that change the pipeline execution in some ways: The default profile (which we can call small ), the medium profile and the large profile. These three profiles were developed based on the size and diversity of the input dataset and change some parameter defaults based on tests we have performed on similar-sized datasets. If you want to first gauge the potential diversity of your dataset and have some input assemblies you can try the PopPUNK entry . One of the outputs will provide insight into how many clusters, or lineages, your dataset divides into. The sizes are: For the default or small profile, we expect datasets with 100 samples/assemblies or fewer. It runs on the default pipeline parameters, with no changes. For the medium profile, we expect datasets with >100 and <1000 samples. It increases the default resource requirements for most processes and also uses PPanGGoLiN for pangenome construction, instead of Panaroo . For the large profile, we expect datasets with >1000 samples. It also increases default resource requirements for some processes and uses PPanGGoLin. Additionally, it enables PopPUNK subsampling , with default parameters .","title":"Dataset Size"},{"location":"resource_profiles/#arete-and-dataset-size","text":"Currently ARETE has three distinct profiles that change the pipeline execution in some ways: The default profile (which we can call small ), the medium profile and the large profile. These three profiles were developed based on the size and diversity of the input dataset and change some parameter defaults based on tests we have performed on similar-sized datasets. If you want to first gauge the potential diversity of your dataset and have some input assemblies you can try the PopPUNK entry . One of the outputs will provide insight into how many clusters, or lineages, your dataset divides into. The sizes are: For the default or small profile, we expect datasets with 100 samples/assemblies or fewer. It runs on the default pipeline parameters, with no changes. For the medium profile, we expect datasets with >100 and <1000 samples. It increases the default resource requirements for most processes and also uses PPanGGoLiN for pangenome construction, instead of Panaroo . For the large profile, we expect datasets with >1000 samples. It also increases default resource requirements for some processes and uses PPanGGoLin. Additionally, it enables PopPUNK subsampling , with default parameters .","title":"ARETE and dataset size"},{"location":"subsampling/","text":"PopPUNK subsetting The subsampling subworkflow is executed if you want to reduce the number of genomes that get added to the phylogenomics subworkflow. By reducing the number of genomes, you can potentially reduce resource requirements for the pangenomics and phylogenomics tools. To enable this subworkflow, add --enable_subsetting when running beiko-lab/ARETE. This will subset genomes based on their core genome similarity and accessory genome similarity, as calculated via their PopPUNK distances. By default, the threshold is --core_similarity 99.9 and --accessory_similarity 99 . But these can be changed by adding these parameters to your execution. What happens then is if any pair of genomes is this similar, only one genome from this pair will be included in the phylogenomic section. All of the removed genome IDs will be present under poppunk_results/removed_genomes.txt . By adding --enable_subsetting , you'll be adding two processes to the execution DAG: POPPUNK_EXTRACT_DISTANCES: This process will extract pair-wise distances between all genomes, returning a table under poppunk_results/distances/ . This table will be used to perform the subsetting. MAKE_HEATMAP: This process will create a heatmap showing different similarity thresholds and the number of genomes that'd be present in each of the possible subsets. It'll also be under poppunk_results/distances/ . Example command The command below will execute the 'annotation' ARETE entry with subsetting enabled, with a core similarity threshold of 99% and an accessory similarity of 95%. nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --enable_subsetting \\ --core_similarity 99 \\ --accessory_similarity 95 \\ -profile docker \\ -entry annotation Be sure to not include --skip_poppunk in your command, because that will then disable all PopPUNK-related processes, including the subsetting subworkflow.","title":"Subsampling"},{"location":"subsampling/#poppunk-subsetting","text":"The subsampling subworkflow is executed if you want to reduce the number of genomes that get added to the phylogenomics subworkflow. By reducing the number of genomes, you can potentially reduce resource requirements for the pangenomics and phylogenomics tools. To enable this subworkflow, add --enable_subsetting when running beiko-lab/ARETE. This will subset genomes based on their core genome similarity and accessory genome similarity, as calculated via their PopPUNK distances. By default, the threshold is --core_similarity 99.9 and --accessory_similarity 99 . But these can be changed by adding these parameters to your execution. What happens then is if any pair of genomes is this similar, only one genome from this pair will be included in the phylogenomic section. All of the removed genome IDs will be present under poppunk_results/removed_genomes.txt . By adding --enable_subsetting , you'll be adding two processes to the execution DAG: POPPUNK_EXTRACT_DISTANCES: This process will extract pair-wise distances between all genomes, returning a table under poppunk_results/distances/ . This table will be used to perform the subsetting. MAKE_HEATMAP: This process will create a heatmap showing different similarity thresholds and the number of genomes that'd be present in each of the possible subsets. It'll also be under poppunk_results/distances/ .","title":"PopPUNK subsetting"},{"location":"subsampling/#example-command","text":"The command below will execute the 'annotation' ARETE entry with subsetting enabled, with a core similarity threshold of 99% and an accessory similarity of 95%. nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --enable_subsetting \\ --core_similarity 99 \\ --accessory_similarity 95 \\ -profile docker \\ -entry annotation Be sure to not include --skip_poppunk in your command, because that will then disable all PopPUNK-related processes, including the subsetting subworkflow.","title":"Example command"},{"location":"usage/","text":"beiko-lab/ARETE: Usage Introduction The ARETE pipeline can is designed as an end-to-end workflow manager for genome assembly, annotation, and phylogenetic analysis, beginning with read data. However, in some cases a user may wish to stop the pipeline prior to annotation or use the annotation features of the work flow with pre-existing assemblies. Therefore, ARETE allows users different use cases: Run the full pipeline end-to-end. Input a set of reads and stop after assembly. Input a set of assemblies and perform QC. Input a set of assemblies and perform annotation and taxonomic analyses. Input a set of assemblies and perform genome clustering with PopPUNK. Input a set of assemblies and perform phylogenomic and pangenomic analysis. This document will describe how to perform each workflow. \"Running the pipeline\" will show some example command on how to use these different entries to ARETE. Samplesheet input No matter your use case, you will need to create a samplesheet with information about the samples you would like to analyse before running the pipeline. Use this parameter to specify its location. For full runs and assembly, it has to be a comma-separated file with 3 columns, and a header row as shown in the examples below. --input_sample_table '[path to samplesheet file]' Full workflow or assembly samplesheet The pipeline will auto-detect whether a sample is single- or paired-end using the information provided in the samplesheet. The samplesheet can have as many columns as you desire, however, there is a strict requirement for the first 4 columns to match those defined in the table below. A final samplesheet file consisting of both single- and paired-end data may look something like the one below. This is for 6 samples, where TREATMENT_REP3 has been sequenced twice. sample,fastq_1,fastq_2 CONTROL_REP1,AEG588A1_S1_L002_R1_001.fastq.gz,AEG588A1_S1_L002_R2_001.fastq.gz CONTROL_REP2,AEG588A2_S2_L002_R1_001.fastq.gz,AEG588A2_S2_L002_R2_001.fastq.gz CONTROL_REP3,AEG588A3_S3_L002_R1_001.fastq.gz,AEG588A3_S3_L002_R2_001.fastq.gz TREATMENT_REP1,AEG588A4_S4_L003_R1_001.fastq.gz, TREATMENT_REP2,AEG588A5_S5_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L004_R1_001.fastq.gz, Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fastq_1 Full path to FastQ file for Illumina short reads 1. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". fastq_2 Full path to FastQ file for Illumina short reads 2. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". An example samplesheet has been provided with the pipeline. Annotation only samplesheet The ARETE pipeline allows users to provide pre-existing assemblies to make use of the annotation and reporting features of the workflow. Users may use the assembly_qc entry point to perform QC on the assemblies. Note that the QC workflow does not automatically filter low quality assemblies, it simply generates QC reports! annotation , assembly_qc and poppunk workflows accept the same format of sample sheet. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fna_file_path Full path to fna file for assembly or genome. File must have .fna file extension. An example samplesheet has been provided with the pipeline. Phylogenomics and Pangenomics only samplesheet The ARETE pipeline allows users to provide pre-existing assemblies to make use of the phylogenomic and pangenomic features of the workflow. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. gff_file_path Full path to GFF file for assembly or genome. File must have .gff or .gff3 file extension. These files can be the ones generated by Prokka or Bakta in ARETE's annotation subworkflow. Reference Genome For full workflow or assembly, users may provide a path to a reference genome in fasta format for use in assembly evaluation. --reference_genome ref.fasta Running the pipeline The typical command for running the pipeline is as follows: nextflow run beiko-lab/ARETE --input_sample_table samplesheet.csv --reference_genome ref.fasta --poppunk_model bgmm -profile docker This will launch the pipeline with the docker configuration profile. See below for more information about profiles. Note that the pipeline will create the following files in your working directory: work # Directory containing the nextflow working files results # Finished results (configurable, see below) .nextflow_log # Log file from Nextflow # Other nextflow hidden files, eg. history of pipeline runs and old logs. As written above, the pipeline also allows users to execute only assembly or only annotation. Assembly Entry To execute assembly (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker Assembly QC Entry To execute QC on pre-existing assemblies (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly_qc --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker Annotation Entry To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry annotation --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker PopPUNK Entry To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry poppunk --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker Phylogenomics and Pangenomics Entry To execute phylogenomic and pangenomics analysis on pre-existing assemblies: nextflow run beiko-lab/ARETE -entry phylogenomics --input_sample_table samplesheet.csv -profile docker rSPR Entry To execute the rSPR analysis on pre-existing trees: nextflow run beiko-lab/ARETE \\ -entry rspr \\ --input_sample_table samplesheet.csv \\ --core_gene_tree core_gene_alignment.tre \\ --concatenated_annotation BAKTA.txt \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --concatenated_annotation - The tabular annotation results (TSV) for all genomes, like the ones generated at the end of Prokka or Bakta in ARETE. Although useful, it's not necessary to execute the rSPR entry. --input_sample_table - A samplesheet containing all individual gene trees in the following format: gene_tree,path CDS_0000,/path/to/CDS_0000.tre CDS_0001,/path/to/CDS_0001.tre CDS_0002,/path/to/CDS_0002.tre CDS_0003,/path/to/CDS_0003.tre CDS_0004,/path/to/CDS_0004.tre evolCCM Entry To execute the evolCCM analysis on a pre-existing reference tree and feature profile: nextflow run beiko-lab/ARETE \\ -entry evolccm \\ --core_gene_tree core_gene_alignment.tre \\ --feature_profile feature_profile.tsv.gz \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --feature_profile - A presence/absence TSV matrix of features in genomes. Genome names should be the same in the core tree and should be contained to a 'genome_id' column, with all other columns represent features absent (0) or present (1) in each genome. I.e.: genome_id plasmid_AA155 plasmid_AA161 ED010 0 0 ED017 0 1 ED040 0 0 ED073 0 1 ED075 1 1 ED082 0 1 ED142 0 1 ED178 0 1 ED180 0 0 Recombination Entry To execute the recombination analysis on pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE \\ -entry recombination \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ -profile docker Updating the pipeline When you run the above command, Nextflow automatically pulls the pipeline code from GitHub and stores it as a cached version. When running the pipeline after this, it will always use the cached version if available - even if the pipeline has been updated since. To make sure that you're running the latest version of the pipeline, make sure that you regularly update the cached version of the pipeline: nextflow pull beiko-lab/ARETE Reproducibility It's a good idea to specify a pipeline version when running the pipeline on your data. This ensures that a specific version of the pipeline code and software are used when you run your pipeline. If you keep using the same tag, you'll be running the same version of the pipeline, even if there have been changes to the code since. First, go to the ARETE releases page and find the latest version number - numeric only (eg. 1.3.1 ). Then specify this when running the pipeline with -r (one hyphen) - eg. -r 1.3.1 . This version number will be logged in reports when you run the pipeline, so that you'll know what you used when you look back in the future. Core Nextflow arguments NB: These options are part of Nextflow and use a single hyphen (pipeline parameters use a double-hyphen). -profile Use this parameter to choose a configuration profile. Profiles can give configuration presets for different compute environments. Several generic profiles are bundled with the pipeline which instruct the pipeline to use software packaged using different methods (Docker, Singularity, Podman, Shifter, Charliecloud) - see below. We highly recommend the use of Docker or Singularity containers for full pipeline reproducibility. The pipeline also dynamically loads configurations from https://github.com/nf-core/configs when it runs, making multiple config profiles for various institutional clusters available at run time. For more information and to see if your system is available in these configs please see the nf-core/configs documentation . Note that multiple profiles can be loaded, for example: -profile test,docker - the order of arguments is important! They are loaded in sequence, so later profiles can overwrite earlier profiles. If -profile is not specified, the pipeline will run locally and expect all software to be installed and available on the PATH . This is not recommended. docker A generic configuration profile to be used with Docker singularity A generic configuration profile to be used with Singularity podman A generic configuration profile to be used with Podman shifter A generic configuration profile to be used with Shifter charliecloud A generic configuration profile to be used with Charliecloud conda Please only use Conda as a last resort i.e. when it's not possible to run the pipeline with Docker, Singularity, Podman, Shifter or Charliecloud. A generic configuration profile to be used with Conda Pulls most software from Bioconda test A profile with a complete configuration for automated testing Can run in personal computers with at least 6GB of RAM and 2 CPUs Includes links to test data so needs no other parameters -resume Specify this when restarting a pipeline. Nextflow will used cached results from any pipeline steps where the inputs are the same, continuing from where it got to previously. You can also supply a run name to resume a specific run: -resume [run-name] . Use the nextflow log command to show previous run names. -c Specify the path to a specific config file (this is a core Nextflow command). See the nf-core website documentation for more information. Custom resource requests Each step in the pipeline has a default set of requirements for number of CPUs, memory and time. For most of the steps in the pipeline, if the job exits with an error code of 143 (exceeded requested resources) it will automatically resubmit with higher requests (2 x original, then 3 x original). If it still fails after three times then the pipeline is stopped. Whilst these default requirements will hopefully work for most people with most data, you may find that you want to customise the compute resources that the pipeline requests. You can do this by creating a custom config file. For example, to give the workflow process UNICYCLER 32GB of memory, you could use the following config: process { withName: UNICYCLER { memory = 32.GB } } To find the exact name of a process you wish to modify the compute resources, check the live-status of a nextflow run displayed on your terminal or check the nextflow error for a line like so: Error executing process > 'bwa' . In this case the name to specify in the custom config file is bwa . See the main Nextflow documentation for more information. Running in the background Nextflow handles job submissions and supervises the running jobs. The Nextflow process must run until the pipeline is finished. The Nextflow -bg flag launches Nextflow in the background, detached from your terminal so that the workflow does not stop if you log out of your session. The logs are saved to a file. Alternatively, you can use screen / tmux or similar tool to create a detached session which you can log back into at a later time. Some HPC setups also allow you to run nextflow within a cluster job submitted your job scheduler (from where it submits more jobs). Nextflow memory requirements In some cases, the Nextflow Java virtual machines can start to request a large amount of memory. We recommend adding the following line to your environment to limit this (typically in ~/.bashrc or ~./bash_profile ): NXF_OPTS='-Xms1g -Xmx4g' Sometimes LevelDB, which is used by Nextflow to track execution metadata, can lead to memory-related issues, often showing as a SIGBUS error. This tends to happen when running Nextflow in SLURM environments . In this case, setting NXF_OPTS=\"-Dleveldb.mmap=false\" in your ~/.bashrc or immediately before executing nextflow run usually solves the issue.","title":"Usage"},{"location":"usage/#beiko-labarete-usage","text":"","title":"beiko-lab/ARETE: Usage"},{"location":"usage/#introduction","text":"The ARETE pipeline can is designed as an end-to-end workflow manager for genome assembly, annotation, and phylogenetic analysis, beginning with read data. However, in some cases a user may wish to stop the pipeline prior to annotation or use the annotation features of the work flow with pre-existing assemblies. Therefore, ARETE allows users different use cases: Run the full pipeline end-to-end. Input a set of reads and stop after assembly. Input a set of assemblies and perform QC. Input a set of assemblies and perform annotation and taxonomic analyses. Input a set of assemblies and perform genome clustering with PopPUNK. Input a set of assemblies and perform phylogenomic and pangenomic analysis. This document will describe how to perform each workflow. \"Running the pipeline\" will show some example command on how to use these different entries to ARETE.","title":"Introduction"},{"location":"usage/#samplesheet-input","text":"No matter your use case, you will need to create a samplesheet with information about the samples you would like to analyse before running the pipeline. Use this parameter to specify its location. For full runs and assembly, it has to be a comma-separated file with 3 columns, and a header row as shown in the examples below. --input_sample_table '[path to samplesheet file]'","title":"Samplesheet input"},{"location":"usage/#full-workflow-or-assembly-samplesheet","text":"The pipeline will auto-detect whether a sample is single- or paired-end using the information provided in the samplesheet. The samplesheet can have as many columns as you desire, however, there is a strict requirement for the first 4 columns to match those defined in the table below. A final samplesheet file consisting of both single- and paired-end data may look something like the one below. This is for 6 samples, where TREATMENT_REP3 has been sequenced twice. sample,fastq_1,fastq_2 CONTROL_REP1,AEG588A1_S1_L002_R1_001.fastq.gz,AEG588A1_S1_L002_R2_001.fastq.gz CONTROL_REP2,AEG588A2_S2_L002_R1_001.fastq.gz,AEG588A2_S2_L002_R2_001.fastq.gz CONTROL_REP3,AEG588A3_S3_L002_R1_001.fastq.gz,AEG588A3_S3_L002_R2_001.fastq.gz TREATMENT_REP1,AEG588A4_S4_L003_R1_001.fastq.gz, TREATMENT_REP2,AEG588A5_S5_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L004_R1_001.fastq.gz, Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fastq_1 Full path to FastQ file for Illumina short reads 1. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". fastq_2 Full path to FastQ file for Illumina short reads 2. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". An example samplesheet has been provided with the pipeline.","title":"Full workflow or assembly samplesheet"},{"location":"usage/#annotation-only-samplesheet","text":"The ARETE pipeline allows users to provide pre-existing assemblies to make use of the annotation and reporting features of the workflow. Users may use the assembly_qc entry point to perform QC on the assemblies. Note that the QC workflow does not automatically filter low quality assemblies, it simply generates QC reports! annotation , assembly_qc and poppunk workflows accept the same format of sample sheet. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fna_file_path Full path to fna file for assembly or genome. File must have .fna file extension. An example samplesheet has been provided with the pipeline.","title":"Annotation only samplesheet"},{"location":"usage/#phylogenomics-and-pangenomics-only-samplesheet","text":"The ARETE pipeline allows users to provide pre-existing assemblies to make use of the phylogenomic and pangenomic features of the workflow. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. gff_file_path Full path to GFF file for assembly or genome. File must have .gff or .gff3 file extension. These files can be the ones generated by Prokka or Bakta in ARETE's annotation subworkflow.","title":"Phylogenomics and Pangenomics only samplesheet"},{"location":"usage/#reference-genome","text":"For full workflow or assembly, users may provide a path to a reference genome in fasta format for use in assembly evaluation. --reference_genome ref.fasta","title":"Reference Genome"},{"location":"usage/#running-the-pipeline","text":"The typical command for running the pipeline is as follows: nextflow run beiko-lab/ARETE --input_sample_table samplesheet.csv --reference_genome ref.fasta --poppunk_model bgmm -profile docker This will launch the pipeline with the docker configuration profile. See below for more information about profiles. Note that the pipeline will create the following files in your working directory: work # Directory containing the nextflow working files results # Finished results (configurable, see below) .nextflow_log # Log file from Nextflow # Other nextflow hidden files, eg. history of pipeline runs and old logs. As written above, the pipeline also allows users to execute only assembly or only annotation.","title":"Running the pipeline"},{"location":"usage/#assembly-entry","text":"To execute assembly (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker","title":"Assembly Entry"},{"location":"usage/#assembly-qc-entry","text":"To execute QC on pre-existing assemblies (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly_qc --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker","title":"Assembly QC Entry"},{"location":"usage/#annotation-entry","text":"To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry annotation --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker","title":"Annotation Entry"},{"location":"usage/#poppunk-entry","text":"To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry poppunk --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker","title":"PopPUNK Entry"},{"location":"usage/#phylogenomics-and-pangenomics-entry","text":"To execute phylogenomic and pangenomics analysis on pre-existing assemblies: nextflow run beiko-lab/ARETE -entry phylogenomics --input_sample_table samplesheet.csv -profile docker","title":"Phylogenomics and Pangenomics Entry"},{"location":"usage/#rspr-entry","text":"To execute the rSPR analysis on pre-existing trees: nextflow run beiko-lab/ARETE \\ -entry rspr \\ --input_sample_table samplesheet.csv \\ --core_gene_tree core_gene_alignment.tre \\ --concatenated_annotation BAKTA.txt \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --concatenated_annotation - The tabular annotation results (TSV) for all genomes, like the ones generated at the end of Prokka or Bakta in ARETE. Although useful, it's not necessary to execute the rSPR entry. --input_sample_table - A samplesheet containing all individual gene trees in the following format: gene_tree,path CDS_0000,/path/to/CDS_0000.tre CDS_0001,/path/to/CDS_0001.tre CDS_0002,/path/to/CDS_0002.tre CDS_0003,/path/to/CDS_0003.tre CDS_0004,/path/to/CDS_0004.tre","title":"rSPR Entry"},{"location":"usage/#evolccm-entry","text":"To execute the evolCCM analysis on a pre-existing reference tree and feature profile: nextflow run beiko-lab/ARETE \\ -entry evolccm \\ --core_gene_tree core_gene_alignment.tre \\ --feature_profile feature_profile.tsv.gz \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --feature_profile - A presence/absence TSV matrix of features in genomes. Genome names should be the same in the core tree and should be contained to a 'genome_id' column, with all other columns represent features absent (0) or present (1) in each genome. I.e.: genome_id plasmid_AA155 plasmid_AA161 ED010 0 0 ED017 0 1 ED040 0 0 ED073 0 1 ED075 1 1 ED082 0 1 ED142 0 1 ED178 0 1 ED180 0 0","title":"evolCCM Entry"},{"location":"usage/#recombination-entry","text":"To execute the recombination analysis on pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE \\ -entry recombination \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ -profile docker","title":"Recombination Entry"},{"location":"usage/#updating-the-pipeline","text":"When you run the above command, Nextflow automatically pulls the pipeline code from GitHub and stores it as a cached version. When running the pipeline after this, it will always use the cached version if available - even if the pipeline has been updated since. To make sure that you're running the latest version of the pipeline, make sure that you regularly update the cached version of the pipeline: nextflow pull beiko-lab/ARETE","title":"Updating the pipeline"},{"location":"usage/#reproducibility","text":"It's a good idea to specify a pipeline version when running the pipeline on your data. This ensures that a specific version of the pipeline code and software are used when you run your pipeline. If you keep using the same tag, you'll be running the same version of the pipeline, even if there have been changes to the code since. First, go to the ARETE releases page and find the latest version number - numeric only (eg. 1.3.1 ). Then specify this when running the pipeline with -r (one hyphen) - eg. -r 1.3.1 . This version number will be logged in reports when you run the pipeline, so that you'll know what you used when you look back in the future.","title":"Reproducibility"},{"location":"usage/#core-nextflow-arguments","text":"NB: These options are part of Nextflow and use a single hyphen (pipeline parameters use a double-hyphen).","title":"Core Nextflow arguments"},{"location":"usage/#-profile","text":"Use this parameter to choose a configuration profile. Profiles can give configuration presets for different compute environments. Several generic profiles are bundled with the pipeline which instruct the pipeline to use software packaged using different methods (Docker, Singularity, Podman, Shifter, Charliecloud) - see below. We highly recommend the use of Docker or Singularity containers for full pipeline reproducibility. The pipeline also dynamically loads configurations from https://github.com/nf-core/configs when it runs, making multiple config profiles for various institutional clusters available at run time. For more information and to see if your system is available in these configs please see the nf-core/configs documentation . Note that multiple profiles can be loaded, for example: -profile test,docker - the order of arguments is important! They are loaded in sequence, so later profiles can overwrite earlier profiles. If -profile is not specified, the pipeline will run locally and expect all software to be installed and available on the PATH . This is not recommended. docker A generic configuration profile to be used with Docker singularity A generic configuration profile to be used with Singularity podman A generic configuration profile to be used with Podman shifter A generic configuration profile to be used with Shifter charliecloud A generic configuration profile to be used with Charliecloud conda Please only use Conda as a last resort i.e. when it's not possible to run the pipeline with Docker, Singularity, Podman, Shifter or Charliecloud. A generic configuration profile to be used with Conda Pulls most software from Bioconda test A profile with a complete configuration for automated testing Can run in personal computers with at least 6GB of RAM and 2 CPUs Includes links to test data so needs no other parameters","title":"-profile"},{"location":"usage/#-resume","text":"Specify this when restarting a pipeline. Nextflow will used cached results from any pipeline steps where the inputs are the same, continuing from where it got to previously. You can also supply a run name to resume a specific run: -resume [run-name] . Use the nextflow log command to show previous run names.","title":"-resume"},{"location":"usage/#-c","text":"Specify the path to a specific config file (this is a core Nextflow command). See the nf-core website documentation for more information.","title":"-c"},{"location":"usage/#custom-resource-requests","text":"Each step in the pipeline has a default set of requirements for number of CPUs, memory and time. For most of the steps in the pipeline, if the job exits with an error code of 143 (exceeded requested resources) it will automatically resubmit with higher requests (2 x original, then 3 x original). If it still fails after three times then the pipeline is stopped. Whilst these default requirements will hopefully work for most people with most data, you may find that you want to customise the compute resources that the pipeline requests. You can do this by creating a custom config file. For example, to give the workflow process UNICYCLER 32GB of memory, you could use the following config: process { withName: UNICYCLER { memory = 32.GB } } To find the exact name of a process you wish to modify the compute resources, check the live-status of a nextflow run displayed on your terminal or check the nextflow error for a line like so: Error executing process > 'bwa' . In this case the name to specify in the custom config file is bwa . See the main Nextflow documentation for more information.","title":"Custom resource requests"},{"location":"usage/#running-in-the-background","text":"Nextflow handles job submissions and supervises the running jobs. The Nextflow process must run until the pipeline is finished. The Nextflow -bg flag launches Nextflow in the background, detached from your terminal so that the workflow does not stop if you log out of your session. The logs are saved to a file. Alternatively, you can use screen / tmux or similar tool to create a detached session which you can log back into at a later time. Some HPC setups also allow you to run nextflow within a cluster job submitted your job scheduler (from where it submits more jobs).","title":"Running in the background"},{"location":"usage/#nextflow-memory-requirements","text":"In some cases, the Nextflow Java virtual machines can start to request a large amount of memory. We recommend adding the following line to your environment to limit this (typically in ~/.bashrc or ~./bash_profile ): NXF_OPTS='-Xms1g -Xmx4g' Sometimes LevelDB, which is used by Nextflow to track execution metadata, can lead to memory-related issues, often showing as a SIGBUS error. This tends to happen when running Nextflow in SLURM environments . In this case, setting NXF_OPTS=\"-Dleveldb.mmap=false\" in your ~/.bashrc or immediately before executing nextflow run usually solves the issue.","title":"Nextflow memory requirements"}]} \ No newline at end of file +{"config":{"indexing":"full","lang":["en"],"min_search_length":3,"prebuild_index":false,"separator":"[\\s\\-]+"},"docs":[{"location":"","text":"What is ARETE? ARETE (Antimicrobial Resistance: Emergence, Transmission, and Ecology) is a bioinformatics best-practice analysis pipeline for profiling the genomic repertoire and evolutionary dynamics of microorganisms with a particular focus on pathogens. We use ARETE to identify important genes (e.g., those that confer antimicrobial resistance or contribute to virulence) and mobile genetic elements such as plasmids and genomic islands, and infer important routes by which these are transmitted using evidence from recombination, cosegregation, coevolution, and phylogenetic trees comparisons. ARETE produces a range of useful outputs (see outputs ), including those generated by each tool integrated into the pipeline, as well as summaries across the entire dataset such as phylogenetic profiles. Outputs from ARETE can also be easily fed into packages such as Coeus and MicroReact for further analyses. Although ARETE was primarily developed with pathogens in mind, inference of pan-genomes, mobilomes, and phylogenomic histories can be performed for any set of microbial genomes, with the proviso that reference databases are much more complete for some taxonomic groups than others. In general, the tools in ARETE work best at the species and genus level of relatedness. A key design feature of ARETE is the versatility to find the right blend of software packages and parameter settings that best handle datasets of different sizes, introducing heuristics and swapping out tools as necessary. ARETE has been benchmarked on datasets from fewer than ten to over 10,000 genomes from a diversity of species and genera including Enterococcus faecium , Escherichia coli , Listeria , and Salmonella . Another key feature is enabling the user choice to run specific subsets of the pipeline; a user may already have assembled genomes, or they may not care about, say, recombination detection. There are also cases where it might be necessary to manually review the outputs from a particular step before moving on to the next one; ARETE makes this manual QC easy to do. Table of Contents About the pipeline Quick start A couple of examples Credits Contributing to ARETE Citing ARETE About the pipeline The pipeline is built using Nextflow , a workflow tool to run tasks across multiple compute infrastructures in a very portable manner. It uses Docker / Singularity containers making installation trivial and results highly reproducible. Like other workflow languages it provides useful features like -resume to only rerun tasks that haven't already been completed (e.g., allowing editing of inputs/tasks and recovery from crashes without a full re-run). The nf-core project provided overall project template, pre-written software modules when available, and general best-practice recommendations. ARETE is organized as a series of subworkflows, each of which executes a different conceptual step of the pipeline. The subworkflow orgnaization provides suitable entry and exit points for users who want to run only a portion of the full pipeline. Genome subsetting The user can optionally subdivide their set of genomes into related lineages identified by PopPUNK ( See documentation ). PopPUNK quickly assignes genomes to 'lineages' based on core and accessory genome identity. If this option is selected, all genomes will still be annotated, but cross-genome comparisons (e.g., pan-genome inference and phylogenomics) will use only a single representative genome from each lineage. The user can run PopPUNK with a spread of different thresholds and decide how to proceed based on the number of lineages produced and their own specific knowledge of the genetic population structure of the taxon being analyzed. Short-read processing and assembly Raw Read QC ( FastQC ) Read Trimming ( fastp ) Trimmed Read QC ( FastQC ) Taxonomic Profiling ( kraken2 ) Unicycler ( unicycler ) QUAST QC ( quast ) CheckM QC ( checkm ) Annotation Genome annotation with Bakta ( bakta ) or Prokka ( prokka ) Feature prediction: AMR genes with the Resistance Gene Identifier ( RGI ) Plasmids with MOB-Suite ( mob_suite ) Genomic Islands with IslandPath ( IslandPath ) Phages with PhiSpy ( PhiSpy ) ( optionally ) Integrons with IntegronFinder Specialized databases: CAZY, VFDB, BacMet and ICEberg2 using DIAMOND homology search ( diamond ) Phylogenomics ( optionally ) Genome subsetting with PopPUNK ( See documentation ) Pan-genome inference using PPanGGOLiN ( PPanGGOLiN ) or Panaroo ( panaroo ) Reference and gene tree inference using FastTree ( fasttree ) or IQTree ( iqtree ) ( optionally ) SNP-sites ( SNPsites ) Recombination detection Recombination detection is performed within lineages identified by PopPUNK ( poppunk ). Note that this application of PopPUNK is different from the subsetting described above. Genome alignment using SKA2 ( ska2 ) Recombination detection using Verticall ( verticall ) and/or Gubbins ( gubbins ) Coevolution Identification of coordinated gain and loss of features using EvolCCM ( EvolCCM ) Lateral gene transfer Phylogenetic inference of LGT using rSPR ( rSPR ) Gene order Comparison of genomic neighbourhoods using the Gene Order Workflow ( Gene Order Workflow ) See our roadmap for a full list of future development targets. Quick Start Install nextflow Install Docker , Singularity , or, as a last resort, Conda . Also ensure you have a working curl installed (should be present on almost all systems). 2.1. Note: this workflow should also support Podman , Shifter or Charliecloud execution for full pipeline reproducibility. We have minimized reliance on conda and suggest using it only as a last resort (see docs ). Configure mail on your system to send an email on workflow success/failure (without this you may get a small error at the end Failed to invoke workflow.onComplete event handler but this doesn't mean the workflow didn't finish successfully). Download the pipeline and test with a stub-run . The stub-run will ensure that the pipeline is able to download and use containers as well as execute in the proper logic. nextflow run beiko-lab/ARETE -profile test, -stub-run 3.1. Please check nf-core/configs to see if a custom config file to run nf-core pipelines already exists for your Institute. If so, you can simply use -profile in your command. This will enable either docker or singularity and set the appropriate execution settings for your local compute environment. 3.2. If you are using singularity then the pipeline will auto-detect this and attempt to download the Singularity images directly as opposed to performing a conversion from Docker images. If you are persistently observing issues downloading Singularity images directly due to timeout or network issues then please use the --singularity_pull_docker_container parameter to pull and convert the Docker image instead. In case of input datasets larger than 100 samples, check our resource profiles documentation , for optimal usage. Start running your own analysis (ideally using -profile docker or -profile singularity for stability)! nextflow run beiko-lab/ARETE \\ -profile \\ --input_sample_table samplesheet.csv \\ --poppunk_model bgmm samplesheet.csv must be formatted sample,fastq_1,fastq_2 Note : If you get this error at the end Failed to invoke `workflow.onComplete` event handler it isn't a problem, it just means you don't have an sendmail configured and it can't send an email report saying it finished correctly i.e., its not that the workflow failed. See usage docs for all of the available options when running the pipeline. See the parameter docs for a list of all params currently implemented in the pipeline and which ones are required. Testing To test the worklow on a minimal dataset you can use the test configuration (with either docker or singularity - replace docker below as appropriate): nextflow run beiko-lab/ARETE -profile test,docker To accelerate it you can download/cache the database files to a folder (e.g., test/db_cache ) and provide a database cache parameter. nextflow run beiko-lab/ARETE \\ -profile test,docker \\ --db_cache $PWD/test/db_cache \\ --bakta_db $PWD/baktadb/db-light We also provide a larger test dataset, under -profile test_full , for use in ARETE's annotation entry. This dataset is comprised of 8 bacterial genomes. As a note, this can take upwards of 20 minutes to complete on an average personal computer . Replace docker below as appropriate. nextflow run beiko-lab/ARETE -entry annotation -profile test_full,docker Examples The fine details of how to run ARETE are described in the command reference and documentation, but here are a couple of illustrative examples of how runs can be adjusted to accommodate genome sets of different sizes: Assembly, annotation, and pan-genome inference from a modestly sized dataset (50 or so genomes) from paired-end reads nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --annotation_tools 'mobsuite,rgi,vfdb,bacmet,islandpath,phispy,report' \\ --poppunk_model bgmm \\ -profile docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --annotation_tools - Select the annotation tools and modules to be executed (See the parameter documentation for defaults) --poppunk_model - Model to be used by PopPUNK -profile docker - Run tools in docker containers. Annotation to evolutionary dynamics on 300-ish genomes nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ --run_gubbins \\ -entry annotation \\ -profile medium,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --run_gubbins - Run Gubbins as part of the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile medium,docker - Run tools in docker containers. For -profile medium , check our resource requirements documentation . Annotation to evolutionary dynamics on 10,000 genomes nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ -entry annotation \\ -profile large,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. --enable_subsetting - Enable subsetting workflow based on genome similarity (See subsetting documentation ) -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile large,docker - Run tools in docker containers. For -profile large , check our resource requirements documentation . Credits The ARETE software was originally written and developed by Finlay Maguire and Alex Manuele , and is currently developed by Jo\u00e3o Cavalcante . Rob Beiko is the PI of the ARETE project. The project Co-PI is Fiona Brinkman. Other project leads include Andrew MacArthur, Cedric Chauve, Chris Whidden, Gary van Domselaar, John Nash, Rahat Zaheer, and Tim McAllister. Many students, postdocs, developers, and staff scientists have made invaluable contributions to the design and application of ARETE and its components, including Haley Sanderson, Kristen Gray, Julia Lewandowski, Chaoyue Liu, Kartik Kakadiya, Bryan Alcock, Amos Raphenya, Amjad Khan, Ryan Fink, Aniket Mane, Chandana Navanekere Rudrappa, Kyrylo Bessonov, James Robertson, Jee In Kim, and Nolan Woods. ARETE development has been supported from many sources, including Genome Canada, ResearchNS, Genome Atlantic, Genome British Columbia, The Canadian Institutes for Health Research, The Natural Sciences and Engineering Research Council of Canada, and Dalhousie University's Faculty of Computer Science. We have received tremendous support from federal agencies, most notably the Public Health Agency of Canada and Agriculture / Agri-Food Canada. Contributing to ARETE Thank you for your interest in contributing to ARETE. We are currently in the process of formalizing contribution guidelines. In the meantime, please feel free to open an issue describing your suggested changes. Citing ARETE Please cite the tools used in your ARETE run: A comprehensive list can be found in the CITATIONS.md file. An early version of ARETE was used for assembly and feature prediction in the following paper : Sanderson H, Gray KL, Manuele A, Maguire F, Khan A, Liu C, Navanekere Rudrappa C, Nash JHE, Robertson J, Bessonov K, Oloni M, Alcock BP, Raphenya AR, McAllister TA, Peacock SJ, Raven KE, Gouliouris T, McArthur AG, Brinkman FSL, Fink RC, Zaheer R, Beiko RG. Exploring the mobilome and resistome of Enterococcus faecium in a One Health context across two continents. Microb Genom. 2022 Sep;8(9):mgen000880. doi: 10.1099/mgen.0.000880. PMID: 36129737; PMCID: PMC9676038. This pipeline uses code and infrastructure developed and maintained by the nf-core initative, and reused here under the MIT license . The nf-core framework for community-curated bioinformatics pipelines. Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen. Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.","title":"Home"},{"location":"#what-is-arete","text":"ARETE (Antimicrobial Resistance: Emergence, Transmission, and Ecology) is a bioinformatics best-practice analysis pipeline for profiling the genomic repertoire and evolutionary dynamics of microorganisms with a particular focus on pathogens. We use ARETE to identify important genes (e.g., those that confer antimicrobial resistance or contribute to virulence) and mobile genetic elements such as plasmids and genomic islands, and infer important routes by which these are transmitted using evidence from recombination, cosegregation, coevolution, and phylogenetic trees comparisons. ARETE produces a range of useful outputs (see outputs ), including those generated by each tool integrated into the pipeline, as well as summaries across the entire dataset such as phylogenetic profiles. Outputs from ARETE can also be easily fed into packages such as Coeus and MicroReact for further analyses. Although ARETE was primarily developed with pathogens in mind, inference of pan-genomes, mobilomes, and phylogenomic histories can be performed for any set of microbial genomes, with the proviso that reference databases are much more complete for some taxonomic groups than others. In general, the tools in ARETE work best at the species and genus level of relatedness. A key design feature of ARETE is the versatility to find the right blend of software packages and parameter settings that best handle datasets of different sizes, introducing heuristics and swapping out tools as necessary. ARETE has been benchmarked on datasets from fewer than ten to over 10,000 genomes from a diversity of species and genera including Enterococcus faecium , Escherichia coli , Listeria , and Salmonella . Another key feature is enabling the user choice to run specific subsets of the pipeline; a user may already have assembled genomes, or they may not care about, say, recombination detection. There are also cases where it might be necessary to manually review the outputs from a particular step before moving on to the next one; ARETE makes this manual QC easy to do.","title":"What is ARETE?"},{"location":"#table-of-contents","text":"About the pipeline Quick start A couple of examples Credits Contributing to ARETE Citing ARETE","title":"Table of Contents"},{"location":"#about-the-pipeline","text":"The pipeline is built using Nextflow , a workflow tool to run tasks across multiple compute infrastructures in a very portable manner. It uses Docker / Singularity containers making installation trivial and results highly reproducible. Like other workflow languages it provides useful features like -resume to only rerun tasks that haven't already been completed (e.g., allowing editing of inputs/tasks and recovery from crashes without a full re-run). The nf-core project provided overall project template, pre-written software modules when available, and general best-practice recommendations. ARETE is organized as a series of subworkflows, each of which executes a different conceptual step of the pipeline. The subworkflow orgnaization provides suitable entry and exit points for users who want to run only a portion of the full pipeline.","title":"About the pipeline "},{"location":"#genome-subsetting","text":"The user can optionally subdivide their set of genomes into related lineages identified by PopPUNK ( See documentation ). PopPUNK quickly assignes genomes to 'lineages' based on core and accessory genome identity. If this option is selected, all genomes will still be annotated, but cross-genome comparisons (e.g., pan-genome inference and phylogenomics) will use only a single representative genome from each lineage. The user can run PopPUNK with a spread of different thresholds and decide how to proceed based on the number of lineages produced and their own specific knowledge of the genetic population structure of the taxon being analyzed.","title":"Genome subsetting"},{"location":"#short-read-processing-and-assembly","text":"Raw Read QC ( FastQC ) Read Trimming ( fastp ) Trimmed Read QC ( FastQC ) Taxonomic Profiling ( kraken2 ) Unicycler ( unicycler ) QUAST QC ( quast ) CheckM QC ( checkm )","title":"Short-read processing and assembly"},{"location":"#annotation","text":"Genome annotation with Bakta ( bakta ) or Prokka ( prokka ) Feature prediction: AMR genes with the Resistance Gene Identifier ( RGI ) Plasmids with MOB-Suite ( mob_suite ) Genomic Islands with IslandPath ( IslandPath ) Phages with PhiSpy ( PhiSpy ) ( optionally ) Integrons with IntegronFinder Specialized databases: CAZY, VFDB, BacMet and ICEberg2 using DIAMOND homology search ( diamond )","title":"Annotation"},{"location":"#phylogenomics","text":"( optionally ) Genome subsetting with PopPUNK ( See documentation ) Pan-genome inference using PPanGGOLiN ( PPanGGOLiN ) or Panaroo ( panaroo ) Reference and gene tree inference using FastTree ( fasttree ) or IQTree ( iqtree ) ( optionally ) SNP-sites ( SNPsites )","title":"Phylogenomics"},{"location":"#recombination-detection","text":"Recombination detection is performed within lineages identified by PopPUNK ( poppunk ). Note that this application of PopPUNK is different from the subsetting described above. Genome alignment using SKA2 ( ska2 ) Recombination detection using Verticall ( verticall ) and/or Gubbins ( gubbins )","title":"Recombination detection"},{"location":"#coevolution","text":"Identification of coordinated gain and loss of features using EvolCCM ( EvolCCM )","title":"Coevolution"},{"location":"#lateral-gene-transfer","text":"Phylogenetic inference of LGT using rSPR ( rSPR )","title":"Lateral gene transfer"},{"location":"#gene-order","text":"Comparison of genomic neighbourhoods using the Gene Order Workflow ( Gene Order Workflow ) See our roadmap for a full list of future development targets.","title":"Gene order"},{"location":"#quick-start","text":"Install nextflow Install Docker , Singularity , or, as a last resort, Conda . Also ensure you have a working curl installed (should be present on almost all systems). 2.1. Note: this workflow should also support Podman , Shifter or Charliecloud execution for full pipeline reproducibility. We have minimized reliance on conda and suggest using it only as a last resort (see docs ). Configure mail on your system to send an email on workflow success/failure (without this you may get a small error at the end Failed to invoke workflow.onComplete event handler but this doesn't mean the workflow didn't finish successfully). Download the pipeline and test with a stub-run . The stub-run will ensure that the pipeline is able to download and use containers as well as execute in the proper logic. nextflow run beiko-lab/ARETE -profile test, -stub-run 3.1. Please check nf-core/configs to see if a custom config file to run nf-core pipelines already exists for your Institute. If so, you can simply use -profile in your command. This will enable either docker or singularity and set the appropriate execution settings for your local compute environment. 3.2. If you are using singularity then the pipeline will auto-detect this and attempt to download the Singularity images directly as opposed to performing a conversion from Docker images. If you are persistently observing issues downloading Singularity images directly due to timeout or network issues then please use the --singularity_pull_docker_container parameter to pull and convert the Docker image instead. In case of input datasets larger than 100 samples, check our resource profiles documentation , for optimal usage. Start running your own analysis (ideally using -profile docker or -profile singularity for stability)! nextflow run beiko-lab/ARETE \\ -profile \\ --input_sample_table samplesheet.csv \\ --poppunk_model bgmm samplesheet.csv must be formatted sample,fastq_1,fastq_2 Note : If you get this error at the end Failed to invoke `workflow.onComplete` event handler it isn't a problem, it just means you don't have an sendmail configured and it can't send an email report saying it finished correctly i.e., its not that the workflow failed. See usage docs for all of the available options when running the pipeline. See the parameter docs for a list of all params currently implemented in the pipeline and which ones are required.","title":"Quick Start "},{"location":"#testing","text":"To test the worklow on a minimal dataset you can use the test configuration (with either docker or singularity - replace docker below as appropriate): nextflow run beiko-lab/ARETE -profile test,docker To accelerate it you can download/cache the database files to a folder (e.g., test/db_cache ) and provide a database cache parameter. nextflow run beiko-lab/ARETE \\ -profile test,docker \\ --db_cache $PWD/test/db_cache \\ --bakta_db $PWD/baktadb/db-light We also provide a larger test dataset, under -profile test_full , for use in ARETE's annotation entry. This dataset is comprised of 8 bacterial genomes. As a note, this can take upwards of 20 minutes to complete on an average personal computer . Replace docker below as appropriate. nextflow run beiko-lab/ARETE -entry annotation -profile test_full,docker","title":"Testing"},{"location":"#examples","text":"The fine details of how to run ARETE are described in the command reference and documentation, but here are a couple of illustrative examples of how runs can be adjusted to accommodate genome sets of different sizes:","title":"Examples "},{"location":"#assembly-annotation-and-pan-genome-inference-from-a-modestly-sized-dataset-50-or-so-genomes-from-paired-end-reads","text":"nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --annotation_tools 'mobsuite,rgi,vfdb,bacmet,islandpath,phispy,report' \\ --poppunk_model bgmm \\ -profile docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --annotation_tools - Select the annotation tools and modules to be executed (See the parameter documentation for defaults) --poppunk_model - Model to be used by PopPUNK -profile docker - Run tools in docker containers.","title":"Assembly, annotation, and pan-genome inference from a modestly sized dataset (50 or so genomes) from paired-end reads"},{"location":"#annotation-to-evolutionary-dynamics-on-300-ish-genomes","text":"nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ --run_gubbins \\ -entry annotation \\ -profile medium,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --run_gubbins - Run Gubbins as part of the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile medium,docker - Run tools in docker containers. For -profile medium , check our resource requirements documentation .","title":"Annotation to evolutionary dynamics on 300-ish genomes"},{"location":"#annotation-to-evolutionary-dynamics-on-10000-genomes","text":"nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ --run_recombination \\ -entry annotation \\ -profile large,docker Parameters used: --input_sample_table - Input dataset in samplesheet format (See usage ) --poppunk_model - Model to be used by PopPUNK . --run_recombination - Run the recombination subworkflow. --use_ppanggolin - Use PPanGGOLiN for calculating the pangenome. Tends to perform better on larger input sets. --enable_subsetting - Enable subsetting workflow based on genome similarity (See subsetting documentation ) -entry annotation - Run annotation subworkflow and further steps (See usage ). -profile large,docker - Run tools in docker containers. For -profile large , check our resource requirements documentation .","title":"Annotation to evolutionary dynamics on 10,000 genomes"},{"location":"#credits","text":"The ARETE software was originally written and developed by Finlay Maguire and Alex Manuele , and is currently developed by Jo\u00e3o Cavalcante . Rob Beiko is the PI of the ARETE project. The project Co-PI is Fiona Brinkman. Other project leads include Andrew MacArthur, Cedric Chauve, Chris Whidden, Gary van Domselaar, John Nash, Rahat Zaheer, and Tim McAllister. Many students, postdocs, developers, and staff scientists have made invaluable contributions to the design and application of ARETE and its components, including Haley Sanderson, Kristen Gray, Julia Lewandowski, Chaoyue Liu, Kartik Kakadiya, Bryan Alcock, Amos Raphenya, Amjad Khan, Ryan Fink, Aniket Mane, Chandana Navanekere Rudrappa, Kyrylo Bessonov, James Robertson, Jee In Kim, and Nolan Woods. ARETE development has been supported from many sources, including Genome Canada, ResearchNS, Genome Atlantic, Genome British Columbia, The Canadian Institutes for Health Research, The Natural Sciences and Engineering Research Council of Canada, and Dalhousie University's Faculty of Computer Science. We have received tremendous support from federal agencies, most notably the Public Health Agency of Canada and Agriculture / Agri-Food Canada.","title":"Credits "},{"location":"#contributing-to-arete","text":"Thank you for your interest in contributing to ARETE. We are currently in the process of formalizing contribution guidelines. In the meantime, please feel free to open an issue describing your suggested changes.","title":"Contributing to ARETE "},{"location":"#citing-arete","text":"Please cite the tools used in your ARETE run: A comprehensive list can be found in the CITATIONS.md file. An early version of ARETE was used for assembly and feature prediction in the following paper : Sanderson H, Gray KL, Manuele A, Maguire F, Khan A, Liu C, Navanekere Rudrappa C, Nash JHE, Robertson J, Bessonov K, Oloni M, Alcock BP, Raphenya AR, McAllister TA, Peacock SJ, Raven KE, Gouliouris T, McArthur AG, Brinkman FSL, Fink RC, Zaheer R, Beiko RG. Exploring the mobilome and resistome of Enterococcus faecium in a One Health context across two continents. Microb Genom. 2022 Sep;8(9):mgen000880. doi: 10.1099/mgen.0.000880. PMID: 36129737; PMCID: PMC9676038. This pipeline uses code and infrastructure developed and maintained by the nf-core initative, and reused here under the MIT license . The nf-core framework for community-curated bioinformatics pipelines. Philip Ewels, Alexander Peltzer, Sven Fillinger, Harshil Patel, Johannes Alneberg, Andreas Wilm, Maxime Ulysse Garcia, Paolo Di Tommaso & Sven Nahnsen. Nat Biotechnol. 2020 Feb 13. doi: 10.1038/s41587-020-0439-x.","title":"Citing ARETE "},{"location":"CITATIONS/","text":"beiko-lab/ARETE: Citations nf-core Ewels PA, Peltzer A, Fillinger S, Patel H, Alneberg J, Wilm A, Garcia MU, Di Tommaso P, Nahnsen S. The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol. 2020 Mar;38(3):276-278. doi: 10.1038/s41587-020-0439-x. PubMed PMID: 32055031. Nextflow Di Tommaso P, Chatzou M, Floden EW, Barja PP, Palumbo E, Notredame C. Nextflow enables reproducible computational workflows. Nat Biotechnol. 2017 Apr 11;35(4):316-319. doi: 10.1038/nbt.3820. PubMed PMID: 28398311. Pipeline tools CheckM Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. 2015. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes . Genome Research, 25: 1043\u20131055. DIAMOND Buchfink B, Xie C, Huson DH Fast and sensitive protein alignment using DIAMOND. Nat. Methods. 12, 59\u201360 (2015) FastQC FastP Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018 Sep 1;34(17):i884-i890. doi: 10.1093/bioinformatics/bty560. PubMed PMID: 30423086; PubMed Central PMCID: PMC6129281. FastTree Morgan N. Price, Paramvir S. Dehal, Adam P. Arkin, FastTree: Computing Large Minimum Evolution Trees with Profiles instead of a Distance Matrix, Molecular Biology and Evolution, Volume 26, Issue 7, July 2009, Pages 1641\u20131650, https://doi.org/10.1093/molbev/msp077 IQ-TREE2 Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, Lanfear R. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Mol Biol Evol. 2020 May 1;37(5):1530-1534. doi: 10.1093/molbev/msaa015. Erratum in: Mol Biol Evol. 2020 Aug 1;37(8):2461. PMID: 32011700; PMCID: PMC7182206. Kraken2 Wood, D et al., 2019. Improved metagenomic analysis with Kraken 2. Genome Biology volume 20, Article number: 257. doi: 10.1186/s13059-019-1891-0. MOB-SUITE Robertson, James, and John H E Nash. \u201cMOB-suite: software tools for clustering, reconstruction and typing of plasmids from draft assemblies.\u201d Microbial genomics vol. 4,8 (2018): e000206. doi:10.1099/mgen.0.000206 Robertson, James et al. \u201cUniversal whole-sequence-based plasmid typing and its utility to prediction of host range and epidemiological surveillance.\u201d Microbial genomics vol. 6,10 (2020): mgen000435. doi:10.1099/mgen.0.000435 MultiQC Ewels P, Magnusson M, Lundin S, K\u00e4ller M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. 2016 Oct 1;32(19):3047-8. doi: 10.1093/bioinformatics/btw354. Epub 2016 Jun 16. PubMed PMID: 27312411; PubMed Central PMCID: PMC5039924. Bakta Schwengers O., Jelonek L., Dieckmann M. A., Beyvers S., Blom J., Goesmann A. (2021). Bakta: rapid and standardized annotation of bacterial genomes via alignment-free sequence identification. Microbial Genomics, 7(11). https://doi.org/10.1099/mgen.0.000685 Prokka Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014 Jul 15;30(14):2068-9. doi: 10.1093/bioinformatics/btu153. Epub 2014 Mar 18. PMID: 24642063. QUAST Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013 Apr 15;29(8):1072-5. doi: 10.1093/bioinformatics/btt086. Epub 2013 Feb 19. PMID: 23422339; PMCID: PMC3624806. RGI Alcock et al. 2020. CARD 2020: antibiotic resistome surveillance with the comprehensive antibiotic resistance database. Nucleic Acids Research, Volume 48, Issue D1, Pages D517-525 [PMID 31665441] IntegronFinder N\u00e9ron, Bertrand, Eloi Littner, Matthieu Haudiquet, Amandine Perrin, Jean Cury, and Eduardo P.C. Rocha. 2022. IntegronFinder 2.0: Identification and Analysis of Integrons across Bacteria, with a Focus on Antibiotic Resistance in Klebsiella Microorganisms 10, no. 4: 700. https://doi.org/10.3390/microorganisms10040700 Panaroo Tonkin-Hill, G., MacAlasdair, N., Ruis, C. et al. Producing polished prokaryotic pangenomes with the Panaroo pipeline. Genome Biol 21, 180 (2020). https://doi.org/10.1186/s13059-020-02090-4 PPanGGoLiN Gautreau G et al. (2020) PPanGGOLiN: Depicting microbial diversity via a partitioned pangenome graph. PLOS Computational Biology 16(3): e1007732. https://doi.org/10.1371/journal.pcbi.1007732 PopPUNK Lees JA, Harris SR, Tonkin-Hill G, Gladstone RA, Lo SW, Weiser JN, Corander J, Bentley SD, Croucher NJ. Fast and flexible bacterial genomic epidemiology with PopPUNK. Genome Res. 2019 Feb;29(2):304-316. doi: 10.1101/gr.241455.118. Epub 2019 Jan 24. PMID: 30679308; PMCID: PMC6360808. SKA2 Harris SR. 2018. SKA: Split Kmer Analysis Toolkit for Bacterial Genomic Epidemiology. bioRxiv 453142 doi: https://doi.org/10.1101/453142 Gubbins Croucher N. J., Page A. J., Connor T. R., Delaney A. J., Keane J. A., Bentley S. D., Parkhill J., Harris S.R. \"Rapid phylogenetic analysis of large samples of recombinant bacterial whole genome sequences using Gubbins\". doi:10.1093/nar/gku1196, Nucleic Acids Research, 2014. Verticall SNP-sites Page AJ, Taylor B, Delaney AJ, Soares J, Seemann T, Keane JA, Harris SR. SNP-sites: rapid efficient extraction of SNPs from multi-FASTA alignments. Microb Genom. 2016 Apr 29;2(4):e000056. doi: 10.1099/mgen.0.000056. PMID: 28348851; PMCID: PMC5320690. Unicycler Wick RR, Judd LM, Gorrie CL, Holt KE. Unicycler: Resolving bacterial genome assemblies from short and long sequencing reads. PLoS Comput Biol. 2017 Jun 8;13(6):e1005595. doi: 10.1371/journal.pcbi.1005595. PMID: 28594827; PMCID: PMC5481147. IslandPath Claire Bertelli, Fiona S L Brinkman, Improved genomic island predictions with IslandPath-DIMOB, Bioinformatics, Volume 34, Issue 13, 01 July 2018, Pages 2161\u20132167, https://doi.org/10.1093/bioinformatics/bty095 PhiSpy Sajia Akhter, Ramy K. Aziz, Robert A. Edwards; PhiSpy: a novel algorithm for finding prophages in bacterial genomes that combines similarity- and composition-based strategies. Nucl Acids Res 2012; 40 (16): e126. doi: 10.1093/nar/gks406 EvolCCM Chaoyue Liu and others, The Community Coevolution Model with Application to the Study of Evolutionary Relationships between Genes Based on Phylogenetic Profiles, Systematic Biology, Volume 72, Issue 3, May 2023, Pages 559\u2013574, https://doi.org/10.1093/sysbio/syac052 rSPR Christopher Whidden, Norbert Zeh, Robert G. Beiko, Supertrees Based on the Subtree Prune-and-Regraft Distance, Systematic Biology, Volume 63, Issue 4, July 2014, Pages 566\u2013581, https://doi.org/10.1093/sysbio/syu023 Software packaging/containerisation tools Anaconda Anaconda Software Distribution. Computer software. Vers. 2-2.4.0. Anaconda, Nov. 2016. Web. Bioconda Gr\u00fcning B, Dale R, Sj\u00f6din A, Chapman BA, Rowe J, Tomkins-Tinch CH, Valieris R, K\u00f6ster J; Bioconda Team. Bioconda: sustainable and comprehensive software distribution for the life sciences. Nat Methods. 2018 Jul;15(7):475-476. doi: 10.1038/s41592-018-0046-7. PubMed PMID: 29967506. BioContainers da Veiga Leprevost F, Gr\u00fcning B, Aflitos SA, R\u00f6st HL, Uszkoreit J, Barsnes H, Vaudel M, Moreno P, Gatto L, Weber J, Bai M, Jimenez RC, Sachsenberg T, Pfeuffer J, Alvarez RV, Griss J, Nesvizhskii AI, Perez-Riverol Y. BioContainers: an open-source and community-driven framework for software standardization. Bioinformatics. 2017 Aug 15;33(16):2580-2582. doi: 10.1093/bioinformatics/btx192. PubMed PMID: 28379341; PubMed Central PMCID: PMC5870671. Docker Dirk Merkel. 2014. Docker: lightweight Linux containers for consistent development and deployment. Linux J. 2014, 239, Article 2 (March 2014). Singularity Kurtzer GM, Sochat V, Bauer MW. Singularity: Scientific containers for mobility of compute. PLoS One. 2017 May 11;12(5):e0177459. doi: 10.1371/journal.pone.0177459. eCollection 2017. PubMed PMID: 28494014; PubMed Central PMCID: PMC5426675.","title":"Citations"},{"location":"CITATIONS/#beiko-labarete-citations","text":"","title":"beiko-lab/ARETE: Citations"},{"location":"CITATIONS/#nf-core","text":"Ewels PA, Peltzer A, Fillinger S, Patel H, Alneberg J, Wilm A, Garcia MU, Di Tommaso P, Nahnsen S. The nf-core framework for community-curated bioinformatics pipelines. Nat Biotechnol. 2020 Mar;38(3):276-278. doi: 10.1038/s41587-020-0439-x. PubMed PMID: 32055031.","title":"nf-core"},{"location":"CITATIONS/#nextflow","text":"Di Tommaso P, Chatzou M, Floden EW, Barja PP, Palumbo E, Notredame C. Nextflow enables reproducible computational workflows. Nat Biotechnol. 2017 Apr 11;35(4):316-319. doi: 10.1038/nbt.3820. PubMed PMID: 28398311.","title":"Nextflow"},{"location":"CITATIONS/#pipeline-tools","text":"CheckM Parks DH, Imelfort M, Skennerton CT, Hugenholtz P, Tyson GW. 2015. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes . Genome Research, 25: 1043\u20131055. DIAMOND Buchfink B, Xie C, Huson DH Fast and sensitive protein alignment using DIAMOND. Nat. Methods. 12, 59\u201360 (2015) FastQC FastP Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018 Sep 1;34(17):i884-i890. doi: 10.1093/bioinformatics/bty560. PubMed PMID: 30423086; PubMed Central PMCID: PMC6129281. FastTree Morgan N. Price, Paramvir S. Dehal, Adam P. Arkin, FastTree: Computing Large Minimum Evolution Trees with Profiles instead of a Distance Matrix, Molecular Biology and Evolution, Volume 26, Issue 7, July 2009, Pages 1641\u20131650, https://doi.org/10.1093/molbev/msp077 IQ-TREE2 Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, Lanfear R. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Mol Biol Evol. 2020 May 1;37(5):1530-1534. doi: 10.1093/molbev/msaa015. Erratum in: Mol Biol Evol. 2020 Aug 1;37(8):2461. PMID: 32011700; PMCID: PMC7182206. Kraken2 Wood, D et al., 2019. Improved metagenomic analysis with Kraken 2. Genome Biology volume 20, Article number: 257. doi: 10.1186/s13059-019-1891-0. MOB-SUITE Robertson, James, and John H E Nash. \u201cMOB-suite: software tools for clustering, reconstruction and typing of plasmids from draft assemblies.\u201d Microbial genomics vol. 4,8 (2018): e000206. doi:10.1099/mgen.0.000206 Robertson, James et al. \u201cUniversal whole-sequence-based plasmid typing and its utility to prediction of host range and epidemiological surveillance.\u201d Microbial genomics vol. 6,10 (2020): mgen000435. doi:10.1099/mgen.0.000435 MultiQC Ewels P, Magnusson M, Lundin S, K\u00e4ller M. MultiQC: summarize analysis results for multiple tools and samples in a single report. Bioinformatics. 2016 Oct 1;32(19):3047-8. doi: 10.1093/bioinformatics/btw354. Epub 2016 Jun 16. PubMed PMID: 27312411; PubMed Central PMCID: PMC5039924. Bakta Schwengers O., Jelonek L., Dieckmann M. A., Beyvers S., Blom J., Goesmann A. (2021). Bakta: rapid and standardized annotation of bacterial genomes via alignment-free sequence identification. Microbial Genomics, 7(11). https://doi.org/10.1099/mgen.0.000685 Prokka Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014 Jul 15;30(14):2068-9. doi: 10.1093/bioinformatics/btu153. Epub 2014 Mar 18. PMID: 24642063. QUAST Gurevich A, Saveliev V, Vyahhi N, Tesler G. QUAST: quality assessment tool for genome assemblies. Bioinformatics. 2013 Apr 15;29(8):1072-5. doi: 10.1093/bioinformatics/btt086. Epub 2013 Feb 19. PMID: 23422339; PMCID: PMC3624806. RGI Alcock et al. 2020. CARD 2020: antibiotic resistome surveillance with the comprehensive antibiotic resistance database. Nucleic Acids Research, Volume 48, Issue D1, Pages D517-525 [PMID 31665441] IntegronFinder N\u00e9ron, Bertrand, Eloi Littner, Matthieu Haudiquet, Amandine Perrin, Jean Cury, and Eduardo P.C. Rocha. 2022. IntegronFinder 2.0: Identification and Analysis of Integrons across Bacteria, with a Focus on Antibiotic Resistance in Klebsiella Microorganisms 10, no. 4: 700. https://doi.org/10.3390/microorganisms10040700 Panaroo Tonkin-Hill, G., MacAlasdair, N., Ruis, C. et al. Producing polished prokaryotic pangenomes with the Panaroo pipeline. Genome Biol 21, 180 (2020). https://doi.org/10.1186/s13059-020-02090-4 PPanGGoLiN Gautreau G et al. (2020) PPanGGOLiN: Depicting microbial diversity via a partitioned pangenome graph. PLOS Computational Biology 16(3): e1007732. https://doi.org/10.1371/journal.pcbi.1007732 PopPUNK Lees JA, Harris SR, Tonkin-Hill G, Gladstone RA, Lo SW, Weiser JN, Corander J, Bentley SD, Croucher NJ. Fast and flexible bacterial genomic epidemiology with PopPUNK. Genome Res. 2019 Feb;29(2):304-316. doi: 10.1101/gr.241455.118. Epub 2019 Jan 24. PMID: 30679308; PMCID: PMC6360808. SKA2 Harris SR. 2018. SKA: Split Kmer Analysis Toolkit for Bacterial Genomic Epidemiology. bioRxiv 453142 doi: https://doi.org/10.1101/453142 Gubbins Croucher N. J., Page A. J., Connor T. R., Delaney A. J., Keane J. A., Bentley S. D., Parkhill J., Harris S.R. \"Rapid phylogenetic analysis of large samples of recombinant bacterial whole genome sequences using Gubbins\". doi:10.1093/nar/gku1196, Nucleic Acids Research, 2014. Verticall SNP-sites Page AJ, Taylor B, Delaney AJ, Soares J, Seemann T, Keane JA, Harris SR. SNP-sites: rapid efficient extraction of SNPs from multi-FASTA alignments. Microb Genom. 2016 Apr 29;2(4):e000056. doi: 10.1099/mgen.0.000056. PMID: 28348851; PMCID: PMC5320690. Unicycler Wick RR, Judd LM, Gorrie CL, Holt KE. Unicycler: Resolving bacterial genome assemblies from short and long sequencing reads. PLoS Comput Biol. 2017 Jun 8;13(6):e1005595. doi: 10.1371/journal.pcbi.1005595. PMID: 28594827; PMCID: PMC5481147. IslandPath Claire Bertelli, Fiona S L Brinkman, Improved genomic island predictions with IslandPath-DIMOB, Bioinformatics, Volume 34, Issue 13, 01 July 2018, Pages 2161\u20132167, https://doi.org/10.1093/bioinformatics/bty095 PhiSpy Sajia Akhter, Ramy K. Aziz, Robert A. Edwards; PhiSpy: a novel algorithm for finding prophages in bacterial genomes that combines similarity- and composition-based strategies. Nucl Acids Res 2012; 40 (16): e126. doi: 10.1093/nar/gks406 EvolCCM Chaoyue Liu and others, The Community Coevolution Model with Application to the Study of Evolutionary Relationships between Genes Based on Phylogenetic Profiles, Systematic Biology, Volume 72, Issue 3, May 2023, Pages 559\u2013574, https://doi.org/10.1093/sysbio/syac052 rSPR Christopher Whidden, Norbert Zeh, Robert G. Beiko, Supertrees Based on the Subtree Prune-and-Regraft Distance, Systematic Biology, Volume 63, Issue 4, July 2014, Pages 566\u2013581, https://doi.org/10.1093/sysbio/syu023","title":"Pipeline tools"},{"location":"CITATIONS/#software-packagingcontainerisation-tools","text":"Anaconda Anaconda Software Distribution. Computer software. Vers. 2-2.4.0. Anaconda, Nov. 2016. Web. Bioconda Gr\u00fcning B, Dale R, Sj\u00f6din A, Chapman BA, Rowe J, Tomkins-Tinch CH, Valieris R, K\u00f6ster J; Bioconda Team. Bioconda: sustainable and comprehensive software distribution for the life sciences. Nat Methods. 2018 Jul;15(7):475-476. doi: 10.1038/s41592-018-0046-7. PubMed PMID: 29967506. BioContainers da Veiga Leprevost F, Gr\u00fcning B, Aflitos SA, R\u00f6st HL, Uszkoreit J, Barsnes H, Vaudel M, Moreno P, Gatto L, Weber J, Bai M, Jimenez RC, Sachsenberg T, Pfeuffer J, Alvarez RV, Griss J, Nesvizhskii AI, Perez-Riverol Y. BioContainers: an open-source and community-driven framework for software standardization. Bioinformatics. 2017 Aug 15;33(16):2580-2582. doi: 10.1093/bioinformatics/btx192. PubMed PMID: 28379341; PubMed Central PMCID: PMC5870671. Docker Dirk Merkel. 2014. Docker: lightweight Linux containers for consistent development and deployment. Linux J. 2014, 239, Article 2 (March 2014). Singularity Kurtzer GM, Sochat V, Bauer MW. Singularity: Scientific containers for mobility of compute. PLoS One. 2017 May 11;12(5):e0177459. doi: 10.1371/journal.pone.0177459. eCollection 2017. PubMed PMID: 28494014; PubMed Central PMCID: PMC5426675.","title":"Software packaging/containerisation tools"},{"location":"ROADMAP/","text":"A list in no particular order of outstanding development features, both in-progress and planned: Integration of additional tools and scripts: Partner applications for analysis and visualization of phylogenetic distributions of genes and MGEs and gene-order clustering (For example, Coeus ).","title":"Roadmap"},{"location":"faq/","text":"Frequently Asked Questions How do I run ARETE in a Slurm HPC environment? Set a config file under ~/.nextflow/config to use the slurm executor: process { executor = 'slurm' pollInterval = '60 sec' submitRateLimit = '60/1min' queueSize = 100 // If an account is necessary: clusterOptions = '--account=' } See the Nextflow documentation for a description of these options. Now, when running ARETE, you'll need to set additional options if your compute nodes don't have network access - as is common for most Slurm clusters. The example below uses the default test data, i.e. the test profile, for demonstration purposes only. nextflow run beiko-lab/ARETE \\ --db_cache path/to/db_cache \\ --bakta_db path/to/baktadb \\ -profile test,singularity Apart from -profile singularity , which just makes ARETE use Singularity/Apptainer containers for running the tools, there are two additional parameters: --db_cache should be the location for the pre-downloaded databases used in the DIAMOND alignments (i.e. Bacmet, VFDB, ICEberg2 and CAZy FASTA files) and in the Kraken2 taxonomic read classification. Although these tools run by default, you can change the selection of annotation tools by changing --annotation_tools and skip Kraken2 by adding --skip_kraken . See the parameter documentation for a full list of parameters and their defaults. --bakta_db should be the location of the pre-downloaded Bakta database Alternatively, you can use Prokka for annotating your assemblies, since it doesn't require a downloaded database ( --use_prokka ). Do note that there could be memory-related issues when running Nextflow in SLURM environments. Can I use the ARETE outputs in MicroReact? Yes you can! In fact, ARETE provides many outputs that can be used in the MicroReact web app. Some of these files are: The PopPUNK lineages tree under poppunk_results/poppunk_visualizations/poppunk_visualizations.microreact . The reference tree built with FastTree under phylogenomics/reference_tree/core_gene_alignment.tre . The annotation feature profile annotation/feature_profile.tsv.gz . This file contains the annotation features in a presence/absence matrix format. Since MicroReact doesn't allow compressed files, just make sure to decompress it before-hand: gunzip feature_profile.tsv.gz Make sure to check our output documentation for a full list of outputs and the parameter documentation for a description of parameters to enable and disable these outputs.","title":"FAQ"},{"location":"faq/#frequently-asked-questions","text":"","title":"Frequently Asked Questions"},{"location":"faq/#how-do-i-run-arete-in-a-slurm-hpc-environment","text":"Set a config file under ~/.nextflow/config to use the slurm executor: process { executor = 'slurm' pollInterval = '60 sec' submitRateLimit = '60/1min' queueSize = 100 // If an account is necessary: clusterOptions = '--account=' } See the Nextflow documentation for a description of these options. Now, when running ARETE, you'll need to set additional options if your compute nodes don't have network access - as is common for most Slurm clusters. The example below uses the default test data, i.e. the test profile, for demonstration purposes only. nextflow run beiko-lab/ARETE \\ --db_cache path/to/db_cache \\ --bakta_db path/to/baktadb \\ -profile test,singularity Apart from -profile singularity , which just makes ARETE use Singularity/Apptainer containers for running the tools, there are two additional parameters: --db_cache should be the location for the pre-downloaded databases used in the DIAMOND alignments (i.e. Bacmet, VFDB, ICEberg2 and CAZy FASTA files) and in the Kraken2 taxonomic read classification. Although these tools run by default, you can change the selection of annotation tools by changing --annotation_tools and skip Kraken2 by adding --skip_kraken . See the parameter documentation for a full list of parameters and their defaults. --bakta_db should be the location of the pre-downloaded Bakta database Alternatively, you can use Prokka for annotating your assemblies, since it doesn't require a downloaded database ( --use_prokka ). Do note that there could be memory-related issues when running Nextflow in SLURM environments.","title":"How do I run ARETE in a Slurm HPC environment?"},{"location":"faq/#can-i-use-the-arete-outputs-in-microreact","text":"Yes you can! In fact, ARETE provides many outputs that can be used in the MicroReact web app. Some of these files are: The PopPUNK lineages tree under poppunk_results/poppunk_visualizations/poppunk_visualizations.microreact . The reference tree built with FastTree under phylogenomics/reference_tree/core_gene_alignment.tre . The annotation feature profile annotation/feature_profile.tsv.gz . This file contains the annotation features in a presence/absence matrix format. Since MicroReact doesn't allow compressed files, just make sure to decompress it before-hand: gunzip feature_profile.tsv.gz Make sure to check our output documentation for a full list of outputs and the parameter documentation for a description of parameters to enable and disable these outputs.","title":"Can I use the ARETE outputs in MicroReact?"},{"location":"output/","text":"beiko-lab/ARETE: Output Introduction The directories listed below will be created in the results directory after the pipeline has finished. All paths are relative to the top-level results directory. Pipeline overview The pipeline is built using Nextflow and processes data using the following steps (steps in italics don't run by default): Short-read processing and assembly FastQC - Raw and trimmed read QC FastP - Read trimming Kraken2 - Taxonomic assignment Unicycler - Short read assembly Quast - Assembly quality score Annotation Bakta or Prokka - Gene detection and annotation MobRecon - Reconstruction and typing of plasmids RGI - Detection and annotation of AMR determinants IslandPath - Predicts genomic islands in bacterial and archaeal genomes. PhiSpy - Prediction of prophages from bacterial genomes IntegronFinder - Finds integrons in DNA sequences Diamond - Detection and annotation of genes using external databases. CAZy: Carbohydrate metabolism VFDB: Virulence factors BacMet: Metal resistance determinants ICEberg: Integrative and conjugative elements PopPUNK Subworkflow PopPUNK - Genome clustering Dynamics EvolCCM - Community Coevolution rSPR - rooted subtree-prune-and-regraft distances Recombination Verticall - Conduct pairwise assembly comparisons between genomes in a same PopPUNK cluster SKA2 - Generate a whole-genome FASTA alignment for each genome within a cluster. Gubbins - Detection of recombination events within genomes of the same cluster. Gene Order Phylogenomics and Pangenomics Panaroo or PPanGGoLiN - Pangenome alignment FastTree or IQTree - Maximum likelihood core genome phylogenetic tree SNPsites - Extracts SNPs from a multi-FASTA alignment Pipeline information Report metrics generated during the workflow execution MultiQC - Aggregate report describing results and QC from the whole pipeline Assembly FastQC read_processing/*_fastqc/ *_fastqc.html : FastQC report containing quality metrics for your untrimmed raw fastq files. *_fastqc.zip : Zip archive containing the FastQC report, tab-delimited data file and plot images. NB: The FastQC plots in this directory are generated relative to the raw, input reads. They may contain adapter sequence and regions of low quality. To see how your reads look after adapter and quality trimming please refer to the FastQC reports in the trimgalore/fastqc/ directory. FastQC gives general quality metrics about your sequenced reads. It provides information about the quality score distribution across your reads, per base sequence content (%A/T/G/C), adapter contamination and overrepresented sequences. For further reading and documentation see the FastQC help pages . NB: The FastQC plots displayed in the MultiQC report shows untrimmed reads. They may contain adapter sequence and potentially regions with low quality. fastp read_processing/fastp/ ${meta.id} : Trimmed files and trimming reports for each input sample. fastp is a all-in-one fastq preprocessor for read/adapter trimming and quality control. It is used in this pipeline for trimming adapter sequences and discard low-quality reads. Kraken2 read_processing/kraken2/ *.kraken2.report.txt : Text file containing genome-wise information of Kraken2 findings. See here for details. *.classified(_(1|2))?.fastq.gz : Fasta file containing classified reads. If paired-end, one file per end. *.unclassified(_(1|2))?.fastq.gz : Fasta file containing unclassified reads. If paired-end, one file per end. Kraken2 is a read classification software which will assign taxonomy to each read comprising a sample. These results may be analyzed as an indicator of contamination. Unicycler assembly/unicycler/ *.assembly.gfa *.scaffolds.fa *.unicycler.log Short/hybrid read assembler. For now only handles short reads in ARETE. Quast assembly/quast/ report.tsv : A tab-seperated report compiling all QC metrics recorded over all genomes quast/ report.(html|tex|pdf|tsv|txt) : The Quast report in different file formats transposed_report.(tsv|txt) : Transpose of the Quast report quast.log : Log file of all Quast runs icarus_viewers/ contig_size_viewer.html basic_stats/ : Directory containing various summary plots generated by Quast. Annotation Bakta annotation/bakta/ ${sample_id}/ : Bakta results will be in one directory per genome. ${sample_id}.tsv : annotations as simple human readble TSV ${sample_id}.gff3 : annotations & sequences in GFF3 format ${sample_id}.gbff : annotations & sequences in (multi) GenBank format ${sample_id}.embl : annotations & sequences in (multi) EMBL format ${sample_id}.fna : replicon/contig DNA sequences as FASTA ${sample_id}.ffn : feature nucleotide sequences as FASTA ${sample_id}.faa : CDS/sORF amino acid sequences as FASTA ${sample_id}.hypotheticals.tsv : further information on hypothetical protein CDS as simple human readble tab separated values ${sample_id}.hypotheticals.faa : hypothetical protein CDS amino acid sequences as FASTA ${sample_id}.txt : summary as TXT ${sample_id}.png : circular genome annotation plot as PNG ${sample_id}.svg : circular genome annotation plot as SVG Bakta is a tool for the rapid & standardized annotation of bacterial genomes and plasmids from both isolates and MAGs Prokka annotation/prokka/ ${sample_id}/ : Prokka results will be in one directory per genome. ${sample_id}.err : Unacceptable annotations ${sample_id}.faa : Protein FASTA file of translated CDS sequences ${sample_id}.ffn : Nucleotide FASTA file of all the prediction transcripts (CDS, rRNA, tRNA, tmRNA, misc_RNA) ${sample_id}.fna : Nucleotide FASTA file of input contig sequences ${sample_id}.fsa : Nucleotide FASTA file of the input contig sequences, used by \"tbl2asn\" to create the .sqn file. It is mostly the same as the .fna file, but with extra Sequin tags in the sequence description lines. ${sample_id}.gff : This is the master annotation in GFF3 format, containing both sequences and annotations. ${sample_id}.gbk : This is a standard Genbank file derived from the master .gff. ${sample_id}.log : Contains all the output that Prokka produced during its run. This is a record of what settings used, even if the --quiet option was enabled. ${sample_id}.sqn : An ASN1 format \"Sequin\" file for submission to Genbank. It needs to be edited to set the correct taxonomy, authors, related publication etc. ${sample_id}.tbl : Feature Table file, used by \"tbl2asn\" to create the .sqn file. ${sample_id}.tsv : Tab-separated file of all features: locus_tag,ftype,len_bp,gene,EC_number,COG,product ${sample_id}.txt : Statistics relating to the annotated features found. Prokka is a software tool to annotate bacterial, archaeal and viral genomes quickly and produce standards-compliant output files. RGI annotation/rgi/ ${sample_id}_rgi.txt : A TSV report containing all AMR predictions for a given genome. For more info see here RGI predicts AMR determinants using the CARD ontology and various trained models. MobRecon annotation/mob_recon ${sample_id}_mob_recon/ : MobRecon results will be in one directory per genome. contig_report.txt - This file describes the assignment of the contig to chromosome or a particular plasmid grouping. mge.report.txt - Blast HSP of detected MGE's/repetitive elements with contextual information. chromosome.fasta - Fasta file of all contigs found to belong to the chromosome. plasmid_*.fasta - Each plasmid group is written to an individual fasta file which contains the assigned contigs. mobtyper_results - Aggregate MOB-typer report files for all identified plasmid. MobRecon reconstructs individual plasmid sequences from draft genome assemblies using the clustered plasmid reference databases DIAMOND annotation/(vfdb|bacmet|cazy|iceberg2)/ ${sample_id}/${sample_id}_(VFDB|BACMET|CAZYDB|ICEberg2).txt : Blast6 formatted TSVs indicating BlastX results of the genes from each genome against VFDB, BacMet, and CAZy databases. (VFDB|BACMET|CAZYDB|ICEberg2).txt : Table with all hits to this database, with a column describing which genome the match originates from. Sorted and filtered by the match's coverage. Diamond is a sequence aligner for protein and translated DNA searches, designed for high performance analysis of big sequence data. We use DIAMOND to predict the presence of virulence factors, heavy metal resistance determinants, carbohydrate-active enzymes, and integrative and conjugative elements using VFDB , BacMet , CAZy , and ICEberg2 respectively. IslandPath annotation/islandpath/ ${sample_id}/ : IslandPath results will be in one directory per genome. ${sample_id}.tsv : IslandPath results Dimob.log : IslandPath execution log IslandPath is a standalone software to predict genomic islands in bacterial and archaeal genomes based on the presence of dinucleotide biases and mobility genes. IntegronFinder Disabled by default. Enable by adding --run_integronfinder to your command. annotation/integron_finder/ Results_Integron_Finder_${sample_id}/ : IntegronFinder results will be in one directory per genome. Integron Finder is a bioinformatics tool to find integrons in bacterial genomes. PhiSpy annotation/phispy/ ${sample_id}/ : PhiSpy results will be in one directory per genome. See the PhiSpy documentation for an extensive description of the output. PhiSpy is a tool for identification of prophages in Bacterial (and probably Archaeal) genomes. Given an annotated genome it will use several approaches to identify the most likely prophage regions. PopPUNK poppunk_results/ poppunk_db/ - Results from PopPUNK's create-db command poppunk_${poppunk_model}/ - Results from PopPUNK's fit-model command poppunk_visualizations/ - Results from the poppunk_visualise command PopPUNK is a tool for clustering genomes. Phylogenomics and Pangenomics Panaroo pangenomics/panaroo/results/ See the panaroo documentation for an extensive description of output provided. Panaroo is a Bacterial Pangenome Analysis Pipeline. PPanGGoLiN pangenomics/ppanggolin/ See the PPanGGoLiN documentation for an extensive description of output provided. PPanGGoLiN is a tool to build a partitioned pangenome graph from microbial genomes FastTree phylogenomics/fasttree/ *.tre : Newick formatted maximum likelihood tree of core-genome alignment. FastTree infers approximately-maximum-likelihood phylogenetic trees from alignments of nucleotide or protein sequences IQTree phylogenomics/iqtree/ *.treefile : Newick formatted maximum likelihood tree of core-genome alignment. IQTree is a fast and effective stochastic algorithm to infer phylogenetic trees by maximum likelihood. SNPsites phylogenomics/snpsites/ filtered_alignment.fas : Variant fasta file. constant.sites.txt : Text file containing counts of constant sites. SNPsites is a tool to rapidly extract SNPs from a multi-FASTA alignment. Dynamics EvolCCM dynamics/EvolCCM/ EvolCCM_*tsv EvolCCM_*pvals EvolCCM_*X2 EvolCCM_*tre EvolCCM is the R implementation for CCM (Community Coevolution Model) rSPR The outputs are approximate and exact Subtree Prune and Regraft (rSPR) distances between pairs of rooted phylogenetic trees. Each CSV file contains these distances and the tree sizes. The PNG files are heatmaps of these distances and their respective tree sizes. dynamics/rSPR/ approx - Approximate rSPR distances exact - Exact rSPR distances rSPR is a software package for calculating rooted subtree-prune-and-regraft distances and rooted agreement forests. Recombination Verticall dynamics/recombination/verticall/ verticall_cluster*.tsv - Verticall results for the genomes within this PopPUNK cluster. Verticall is a tool to help produce bacterial genome phylogenies which are not influenced by horizontally acquired sequences such as recombination. SKA2 dynamics/recombination/ska2/ cluster_*.aln - SKA2 results for the genomes within this PopPUNK cluster. SKA2 (Split Kmer Analysis) is a toolkit for prokaryotic (and any other small, haploid) DNA sequence analysis using split kmers. Gubbins dynamics/recombination/gubbins/ cluster_*/ - Gubbins results for the genomes within this PopPUNK cluster. Gubbins is an algorithm that iteratively identifies loci containing elevated densities of base substitutions while concurrently constructing a phylogeny based on the putative point mutations outside of these regions. Gene Order gene-order/ extraction/ - AMR genes of interest and their neighborhoods extracted from the assemblies. diamond/ - Pairwise alignments between all input genomes. clustering/ - Similarity and distance matrices for each AMR gene clustered via UPGMA, MCL and DBSCAN to identify similarities between their neighborhoods across all genomes. Gene Order is a subworkflow for bacterial gene order analysis, with outputs easily explorable through its partner visualization application Coeus . Pipeline information pipeline_info/ Reports generated by Nextflow: execution_report.html , execution_timeline.html , execution_trace.txt and pipeline_dag.dot / pipeline_dag.svg . Reports generated by the pipeline: pipeline_report.html , pipeline_report.txt and software_versions.csv . Reformatted samplesheet files used as input to the pipeline: samplesheet.valid.csv . Nextflow provides excellent functionality for generating various reports relevant to the running and execution of the pipeline. This will allow you to troubleshoot errors with the running of the pipeline, and also provide you with other information such as launch commands, run times and resource usage. MultiQC multiqc/ multiqc_report.html : a standalone HTML file that can be viewed in your web browser. multiqc_data/ : directory containing parsed statistics from the different tools used in the pipeline. multiqc_plots/ : directory containing static images from the report in various formats. MultiQC is a visualization tool that generates a single HTML report summarising all samples in your project. Most of the pipeline QC results are visualised in the report and further statistics are available in the report data directory. Results generated by MultiQC collate pipeline QC from supported tools e.g. FastQC. The pipeline has special steps which also allow the software versions to be reported in the MultiQC output for future traceability. For more information about how to use MultiQC reports, see http://multiqc.info .","title":"Output"},{"location":"output/#beiko-labarete-output","text":"","title":"beiko-lab/ARETE: Output"},{"location":"output/#introduction","text":"The directories listed below will be created in the results directory after the pipeline has finished. All paths are relative to the top-level results directory.","title":"Introduction"},{"location":"output/#pipeline-overview","text":"The pipeline is built using Nextflow and processes data using the following steps (steps in italics don't run by default): Short-read processing and assembly FastQC - Raw and trimmed read QC FastP - Read trimming Kraken2 - Taxonomic assignment Unicycler - Short read assembly Quast - Assembly quality score Annotation Bakta or Prokka - Gene detection and annotation MobRecon - Reconstruction and typing of plasmids RGI - Detection and annotation of AMR determinants IslandPath - Predicts genomic islands in bacterial and archaeal genomes. PhiSpy - Prediction of prophages from bacterial genomes IntegronFinder - Finds integrons in DNA sequences Diamond - Detection and annotation of genes using external databases. CAZy: Carbohydrate metabolism VFDB: Virulence factors BacMet: Metal resistance determinants ICEberg: Integrative and conjugative elements PopPUNK Subworkflow PopPUNK - Genome clustering Dynamics EvolCCM - Community Coevolution rSPR - rooted subtree-prune-and-regraft distances Recombination Verticall - Conduct pairwise assembly comparisons between genomes in a same PopPUNK cluster SKA2 - Generate a whole-genome FASTA alignment for each genome within a cluster. Gubbins - Detection of recombination events within genomes of the same cluster. Gene Order Phylogenomics and Pangenomics Panaroo or PPanGGoLiN - Pangenome alignment FastTree or IQTree - Maximum likelihood core genome phylogenetic tree SNPsites - Extracts SNPs from a multi-FASTA alignment Pipeline information Report metrics generated during the workflow execution MultiQC - Aggregate report describing results and QC from the whole pipeline","title":"Pipeline overview"},{"location":"output/#assembly","text":"","title":"Assembly"},{"location":"output/#fastqc","text":"read_processing/*_fastqc/ *_fastqc.html : FastQC report containing quality metrics for your untrimmed raw fastq files. *_fastqc.zip : Zip archive containing the FastQC report, tab-delimited data file and plot images. NB: The FastQC plots in this directory are generated relative to the raw, input reads. They may contain adapter sequence and regions of low quality. To see how your reads look after adapter and quality trimming please refer to the FastQC reports in the trimgalore/fastqc/ directory. FastQC gives general quality metrics about your sequenced reads. It provides information about the quality score distribution across your reads, per base sequence content (%A/T/G/C), adapter contamination and overrepresented sequences. For further reading and documentation see the FastQC help pages . NB: The FastQC plots displayed in the MultiQC report shows untrimmed reads. They may contain adapter sequence and potentially regions with low quality.","title":"FastQC"},{"location":"output/#fastp","text":"read_processing/fastp/ ${meta.id} : Trimmed files and trimming reports for each input sample. fastp is a all-in-one fastq preprocessor for read/adapter trimming and quality control. It is used in this pipeline for trimming adapter sequences and discard low-quality reads.","title":"fastp"},{"location":"output/#kraken2","text":"read_processing/kraken2/ *.kraken2.report.txt : Text file containing genome-wise information of Kraken2 findings. See here for details. *.classified(_(1|2))?.fastq.gz : Fasta file containing classified reads. If paired-end, one file per end. *.unclassified(_(1|2))?.fastq.gz : Fasta file containing unclassified reads. If paired-end, one file per end. Kraken2 is a read classification software which will assign taxonomy to each read comprising a sample. These results may be analyzed as an indicator of contamination.","title":"Kraken2"},{"location":"output/#unicycler","text":"assembly/unicycler/ *.assembly.gfa *.scaffolds.fa *.unicycler.log Short/hybrid read assembler. For now only handles short reads in ARETE.","title":"Unicycler"},{"location":"output/#quast","text":"assembly/quast/ report.tsv : A tab-seperated report compiling all QC metrics recorded over all genomes quast/ report.(html|tex|pdf|tsv|txt) : The Quast report in different file formats transposed_report.(tsv|txt) : Transpose of the Quast report quast.log : Log file of all Quast runs icarus_viewers/ contig_size_viewer.html basic_stats/ : Directory containing various summary plots generated by Quast.","title":"Quast"},{"location":"output/#annotation","text":"","title":"Annotation"},{"location":"output/#bakta","text":"annotation/bakta/ ${sample_id}/ : Bakta results will be in one directory per genome. ${sample_id}.tsv : annotations as simple human readble TSV ${sample_id}.gff3 : annotations & sequences in GFF3 format ${sample_id}.gbff : annotations & sequences in (multi) GenBank format ${sample_id}.embl : annotations & sequences in (multi) EMBL format ${sample_id}.fna : replicon/contig DNA sequences as FASTA ${sample_id}.ffn : feature nucleotide sequences as FASTA ${sample_id}.faa : CDS/sORF amino acid sequences as FASTA ${sample_id}.hypotheticals.tsv : further information on hypothetical protein CDS as simple human readble tab separated values ${sample_id}.hypotheticals.faa : hypothetical protein CDS amino acid sequences as FASTA ${sample_id}.txt : summary as TXT ${sample_id}.png : circular genome annotation plot as PNG ${sample_id}.svg : circular genome annotation plot as SVG Bakta is a tool for the rapid & standardized annotation of bacterial genomes and plasmids from both isolates and MAGs","title":"Bakta"},{"location":"output/#prokka","text":"annotation/prokka/ ${sample_id}/ : Prokka results will be in one directory per genome. ${sample_id}.err : Unacceptable annotations ${sample_id}.faa : Protein FASTA file of translated CDS sequences ${sample_id}.ffn : Nucleotide FASTA file of all the prediction transcripts (CDS, rRNA, tRNA, tmRNA, misc_RNA) ${sample_id}.fna : Nucleotide FASTA file of input contig sequences ${sample_id}.fsa : Nucleotide FASTA file of the input contig sequences, used by \"tbl2asn\" to create the .sqn file. It is mostly the same as the .fna file, but with extra Sequin tags in the sequence description lines. ${sample_id}.gff : This is the master annotation in GFF3 format, containing both sequences and annotations. ${sample_id}.gbk : This is a standard Genbank file derived from the master .gff. ${sample_id}.log : Contains all the output that Prokka produced during its run. This is a record of what settings used, even if the --quiet option was enabled. ${sample_id}.sqn : An ASN1 format \"Sequin\" file for submission to Genbank. It needs to be edited to set the correct taxonomy, authors, related publication etc. ${sample_id}.tbl : Feature Table file, used by \"tbl2asn\" to create the .sqn file. ${sample_id}.tsv : Tab-separated file of all features: locus_tag,ftype,len_bp,gene,EC_number,COG,product ${sample_id}.txt : Statistics relating to the annotated features found. Prokka is a software tool to annotate bacterial, archaeal and viral genomes quickly and produce standards-compliant output files.","title":"Prokka"},{"location":"output/#rgi","text":"annotation/rgi/ ${sample_id}_rgi.txt : A TSV report containing all AMR predictions for a given genome. For more info see here RGI predicts AMR determinants using the CARD ontology and various trained models.","title":"RGI"},{"location":"output/#mobrecon","text":"annotation/mob_recon ${sample_id}_mob_recon/ : MobRecon results will be in one directory per genome. contig_report.txt - This file describes the assignment of the contig to chromosome or a particular plasmid grouping. mge.report.txt - Blast HSP of detected MGE's/repetitive elements with contextual information. chromosome.fasta - Fasta file of all contigs found to belong to the chromosome. plasmid_*.fasta - Each plasmid group is written to an individual fasta file which contains the assigned contigs. mobtyper_results - Aggregate MOB-typer report files for all identified plasmid. MobRecon reconstructs individual plasmid sequences from draft genome assemblies using the clustered plasmid reference databases","title":"MobRecon"},{"location":"output/#diamond","text":"annotation/(vfdb|bacmet|cazy|iceberg2)/ ${sample_id}/${sample_id}_(VFDB|BACMET|CAZYDB|ICEberg2).txt : Blast6 formatted TSVs indicating BlastX results of the genes from each genome against VFDB, BacMet, and CAZy databases. (VFDB|BACMET|CAZYDB|ICEberg2).txt : Table with all hits to this database, with a column describing which genome the match originates from. Sorted and filtered by the match's coverage. Diamond is a sequence aligner for protein and translated DNA searches, designed for high performance analysis of big sequence data. We use DIAMOND to predict the presence of virulence factors, heavy metal resistance determinants, carbohydrate-active enzymes, and integrative and conjugative elements using VFDB , BacMet , CAZy , and ICEberg2 respectively.","title":"DIAMOND"},{"location":"output/#islandpath","text":"annotation/islandpath/ ${sample_id}/ : IslandPath results will be in one directory per genome. ${sample_id}.tsv : IslandPath results Dimob.log : IslandPath execution log IslandPath is a standalone software to predict genomic islands in bacterial and archaeal genomes based on the presence of dinucleotide biases and mobility genes.","title":"IslandPath"},{"location":"output/#integronfinder","text":"Disabled by default. Enable by adding --run_integronfinder to your command. annotation/integron_finder/ Results_Integron_Finder_${sample_id}/ : IntegronFinder results will be in one directory per genome. Integron Finder is a bioinformatics tool to find integrons in bacterial genomes.","title":"IntegronFinder"},{"location":"output/#phispy","text":"annotation/phispy/ ${sample_id}/ : PhiSpy results will be in one directory per genome. See the PhiSpy documentation for an extensive description of the output. PhiSpy is a tool for identification of prophages in Bacterial (and probably Archaeal) genomes. Given an annotated genome it will use several approaches to identify the most likely prophage regions.","title":"PhiSpy"},{"location":"output/#poppunk","text":"poppunk_results/ poppunk_db/ - Results from PopPUNK's create-db command poppunk_${poppunk_model}/ - Results from PopPUNK's fit-model command poppunk_visualizations/ - Results from the poppunk_visualise command PopPUNK is a tool for clustering genomes.","title":"PopPUNK"},{"location":"output/#phylogenomics-and-pangenomics","text":"","title":"Phylogenomics and Pangenomics"},{"location":"output/#panaroo","text":"pangenomics/panaroo/results/ See the panaroo documentation for an extensive description of output provided. Panaroo is a Bacterial Pangenome Analysis Pipeline.","title":"Panaroo"},{"location":"output/#ppanggolin","text":"pangenomics/ppanggolin/ See the PPanGGoLiN documentation for an extensive description of output provided. PPanGGoLiN is a tool to build a partitioned pangenome graph from microbial genomes","title":"PPanGGoLiN"},{"location":"output/#fasttree","text":"phylogenomics/fasttree/ *.tre : Newick formatted maximum likelihood tree of core-genome alignment. FastTree infers approximately-maximum-likelihood phylogenetic trees from alignments of nucleotide or protein sequences","title":"FastTree"},{"location":"output/#iqtree","text":"phylogenomics/iqtree/ *.treefile : Newick formatted maximum likelihood tree of core-genome alignment. IQTree is a fast and effective stochastic algorithm to infer phylogenetic trees by maximum likelihood.","title":"IQTree"},{"location":"output/#snpsites","text":"phylogenomics/snpsites/ filtered_alignment.fas : Variant fasta file. constant.sites.txt : Text file containing counts of constant sites. SNPsites is a tool to rapidly extract SNPs from a multi-FASTA alignment.","title":"SNPsites"},{"location":"output/#dynamics","text":"","title":"Dynamics"},{"location":"output/#evolccm","text":"dynamics/EvolCCM/ EvolCCM_*tsv EvolCCM_*pvals EvolCCM_*X2 EvolCCM_*tre EvolCCM is the R implementation for CCM (Community Coevolution Model)","title":"EvolCCM"},{"location":"output/#rspr","text":"The outputs are approximate and exact Subtree Prune and Regraft (rSPR) distances between pairs of rooted phylogenetic trees. Each CSV file contains these distances and the tree sizes. The PNG files are heatmaps of these distances and their respective tree sizes. dynamics/rSPR/ approx - Approximate rSPR distances exact - Exact rSPR distances rSPR is a software package for calculating rooted subtree-prune-and-regraft distances and rooted agreement forests.","title":"rSPR"},{"location":"output/#recombination","text":"","title":"Recombination"},{"location":"output/#verticall","text":"dynamics/recombination/verticall/ verticall_cluster*.tsv - Verticall results for the genomes within this PopPUNK cluster. Verticall is a tool to help produce bacterial genome phylogenies which are not influenced by horizontally acquired sequences such as recombination.","title":"Verticall"},{"location":"output/#ska2","text":"dynamics/recombination/ska2/ cluster_*.aln - SKA2 results for the genomes within this PopPUNK cluster. SKA2 (Split Kmer Analysis) is a toolkit for prokaryotic (and any other small, haploid) DNA sequence analysis using split kmers.","title":"SKA2"},{"location":"output/#gubbins","text":"dynamics/recombination/gubbins/ cluster_*/ - Gubbins results for the genomes within this PopPUNK cluster. Gubbins is an algorithm that iteratively identifies loci containing elevated densities of base substitutions while concurrently constructing a phylogeny based on the putative point mutations outside of these regions.","title":"Gubbins"},{"location":"output/#gene-order","text":"gene-order/ extraction/ - AMR genes of interest and their neighborhoods extracted from the assemblies. diamond/ - Pairwise alignments between all input genomes. clustering/ - Similarity and distance matrices for each AMR gene clustered via UPGMA, MCL and DBSCAN to identify similarities between their neighborhoods across all genomes. Gene Order is a subworkflow for bacterial gene order analysis, with outputs easily explorable through its partner visualization application Coeus .","title":"Gene Order"},{"location":"output/#pipeline-information","text":"pipeline_info/ Reports generated by Nextflow: execution_report.html , execution_timeline.html , execution_trace.txt and pipeline_dag.dot / pipeline_dag.svg . Reports generated by the pipeline: pipeline_report.html , pipeline_report.txt and software_versions.csv . Reformatted samplesheet files used as input to the pipeline: samplesheet.valid.csv . Nextflow provides excellent functionality for generating various reports relevant to the running and execution of the pipeline. This will allow you to troubleshoot errors with the running of the pipeline, and also provide you with other information such as launch commands, run times and resource usage.","title":"Pipeline information"},{"location":"output/#multiqc","text":"multiqc/ multiqc_report.html : a standalone HTML file that can be viewed in your web browser. multiqc_data/ : directory containing parsed statistics from the different tools used in the pipeline. multiqc_plots/ : directory containing static images from the report in various formats. MultiQC is a visualization tool that generates a single HTML report summarising all samples in your project. Most of the pipeline QC results are visualised in the report and further statistics are available in the report data directory. Results generated by MultiQC collate pipeline QC from supported tools e.g. FastQC. The pipeline has special steps which also allow the software versions to be reported in the MultiQC output for future traceability. For more information about how to use MultiQC reports, see http://multiqc.info .","title":"MultiQC"},{"location":"params/","text":"beiko-lab/ARETE pipeline parameters AMR/VF LGT-focused bacterial genomics workflow Input/output options Define where the pipeline should find input data and save output data. Parameter Description Type Default Required Hidden input_sample_table Path to comma-separated file containing information about the samples in the experiment. Help You will need to create a design file with information about the samples in your experiment before running the pipeline. Use this parameter to specify its location. It has to be a comma-separated file with 3 columns, and a header row. string outdir Path to the output directory where the results will be saved. string ./results db_cache Directory where the databases are located string email Email address for completion summary. Help Set this parameter to your e-mail address to get a summary e-mail with details of the run sent to you when the workflow exits. If set in your user config file ( ~/.nextflow/config ) then you don't need to specify this on the command line for every run. string multiqc_title MultiQC report title. Printed as page header, used for filename if not otherwise specified. string Reference genome options Reference and outgroup genome fasta files required for the workflow. Parameter Description Type Default Required Hidden reference_genome Path to FASTA reference genome file. string QC Parameter Description Type Default Required Hidden run_checkm Run CheckM QC software boolean apply_filtering Filter assemblies on QC results boolean skip_kraken Don't run Kraken2 taxonomic classification boolean min_n50 Minimum N50 for filtering integer 10000 min_contigs_1000_bp Minimum number of contigs with >1000bp integer 1 min_contig_length Minimum average contig length integer 1 Annotation Parameters for the annotation subworkflow Parameter Description Type Default Required Hidden annotation_tools Comma-separated list of annotation tools to run string mobsuite,rgi,cazy,vfdb,iceberg,bacmet,islandpath,phispy,report bakta_db Path to the BAKTA database string use_prokka Use Prokka (not Bakta) for annotating assemblies boolean min_pident Minimum match identity percentage for filtering integer 60 min_qcover Minimum coverage of each match for filtering number 0.6 skip_profile_creation Skip annotation feature profile creation boolean feature_profile_columns Columns to include in the feature profile string mobsuite,rgi,cazy,vfdb,iceberg,bacmet Phylogenomics Parameters for the phylogenomics subworkflow Parameter Description Type Default Required Hidden skip_phylo Skip Pangenomics and Phylogenomics subworkflow boolean use_ppanggolin Use ppanggolin for calculating the pangenome boolean use_full_alignment Use full alignment boolean use_fasttree Use FastTree boolean True PopPUNK Parameters for the lineage subworkflow Parameter Description Type Default Required Hidden skip_poppunk Skip PopPunk boolean poppunk_model Which PopPunk model to use (bgmm, dbscan, refine, threshold or lineage) string run_poppunk_qc Whether to run the QC step for PopPunk boolean enable_subsetting Enable subsetting workflow based on genome similarity boolean core_similarity Similarity threshold for core genomes number 99.99 accessory_similarity Similarity threshold for accessory genes number 99.0 Gene Order Parameters for the Gene Order Subworkflow Parameter Description Type Default Required Hidden run_gene_order Whether to run the Gene Order subworkflow boolean gene_order_percent_cutoff Cutoff percentage of genomes a gene should be present within to be included in extraction and subsequent analysis. Should a float between 0 and 1 (e.g., 0.25 means only genes present in a minimum of 25% of genomes are kept). number 0.25 gene_order_label_cols If using annotation files predicting features, list of space separated column names to be added to the gene names string None num_neighbors Neighborhood size to extract. Should be an even number N, such that N/2 neighbors upstream and N/2 neighbors downstream will be analyzed. integer 10 inflation Inflation hyperparameter value for Markov Clustering Algorithm. integer 2 epsilon Epsilon hyperparameter value for DBSCAN clustering. number 0.5 minpts Minpts hyperparameter value for DBSCAN clustering. integer 5 plot_clustering Create Clustering HTML Plots boolean Recombination Parameters for the recombination subworkflow Parameter Description Type Default Required Hidden run_recombination Run Recombination boolean run_verticall Run Verticall recombination tool boolean True run_gubbins Run Gubbins recombination tool boolean Dynamics Parameter Description Type Default Required Hidden run_evolccm Run the community coevolution model boolean run_rspr Run rSPR boolean min_rspr_distance Minimum rSPR distance used to define processing groups integer 10 min_branch_length Minimum rSPR branch length integer 0 max_support_threshold Maximum rSPR support threshold number 0.7 max_approx_rspr Maximum approximate rSPR distance for filtering integer -1 core_gene_tree Core (or reference) genome tree. Used in the rSPR and evolCCM entries. string concatenated_annotation TSV table of annotations for all genomes. Such as the ones generated by Bakta or Prokka in ARETE. string feature_profile Feature profile TSV (A presence-absence matrix). Used in the evolCCM entry. string Institutional config options Parameters used to describe centralised config profiles. These should not be edited. Parameter Description Type Default Required Hidden custom_config_version Git commit id for Institutional configs. string master True custom_config_base Base directory for Institutional configs. Help If you're running offline, Nextflow will not be able to fetch the institutional config files from the internet. If you don't need them, then this is not a problem. If you do need them, you should download the files from the repo and tell Nextflow where to find them with this parameter. string https://raw.githubusercontent.com/nf-core/configs/master True hostnames Institutional configs hostname. string True config_profile_name Institutional config name. string True config_profile_description Institutional config description. string True config_profile_contact Institutional config contact information. string True config_profile_url Institutional config URL link. string True Max job request options Set the top limit for requested resources for any single job. Parameter Description Type Default Required Hidden max_cpus Maximum number of CPUs that can be requested for any single job. Help Use to set an upper-limit for the CPU requirement for each process. Should be an integer e.g. --max_cpus 1 integer 16 True max_memory Maximum amount of memory that can be requested for any single job. Help Use to set an upper-limit for the memory requirement for each process. Should be a string in the format integer-unit e.g. --max_memory '8.GB' string 128.GB True max_time Maximum amount of time that can be requested for any single job. Help Use to set an upper-limit for the time requirement for each process. Should be a string in the format integer-unit e.g. --max_time '2.h' string 240.h True Generic options Less common options for the pipeline, typically set in a config file. Parameter Description Type Default Required Hidden help Display help text. boolean True publish_dir_mode Method used to save pipeline results to output directory. Help The Nextflow publishDir option specifies which intermediate files should be saved to the output directory. This option tells the pipeline what method should be used to move these files. See Nextflow docs for details. string copy True email_on_fail Email address for completion summary, only when pipeline fails. Help An email address to send a summary email to when the pipeline is completed - ONLY sent if the pipeline does not exit successfully. string True plaintext_email Send plain-text email instead of HTML. boolean True max_multiqc_email_size File size limit when attaching MultiQC reports to summary emails. string 25.MB True monochrome_logs Do not use coloured log outputs. boolean True multiqc_config Custom config file to supply to MultiQC. string True tracedir Directory to keep pipeline Nextflow logs and reports. string ${params.outdir}/pipeline_info True validate_params Boolean whether to validate parameters against the schema at runtime boolean True True show_hidden_params Show all params when using --help Help By default, parameters set as hidden in the schema are not shown on the command line when a user runs with --help . Specifying this option will tell the pipeline to show all parameters. boolean True enable_conda Run this workflow with Conda. You can also use '-profile conda' instead of providing this parameter. boolean True singularity_pull_docker_container Instead of directly downloading Singularity images for use with Singularity, force the workflow to pull and convert Docker containers instead. Help This may be useful for example if you are unable to directly pull Singularity containers to run the pipeline due to http/https proxy issues. boolean True schema_ignore_params string genomes,modules multiqc_logo string True","title":"Parameters"},{"location":"params/#beiko-labarete-pipeline-parameters","text":"AMR/VF LGT-focused bacterial genomics workflow","title":"beiko-lab/ARETE pipeline parameters"},{"location":"params/#inputoutput-options","text":"Define where the pipeline should find input data and save output data. Parameter Description Type Default Required Hidden input_sample_table Path to comma-separated file containing information about the samples in the experiment. Help You will need to create a design file with information about the samples in your experiment before running the pipeline. Use this parameter to specify its location. It has to be a comma-separated file with 3 columns, and a header row. string outdir Path to the output directory where the results will be saved. string ./results db_cache Directory where the databases are located string email Email address for completion summary. Help Set this parameter to your e-mail address to get a summary e-mail with details of the run sent to you when the workflow exits. If set in your user config file ( ~/.nextflow/config ) then you don't need to specify this on the command line for every run. string multiqc_title MultiQC report title. Printed as page header, used for filename if not otherwise specified. string","title":"Input/output options"},{"location":"params/#reference-genome-options","text":"Reference and outgroup genome fasta files required for the workflow. Parameter Description Type Default Required Hidden reference_genome Path to FASTA reference genome file. string","title":"Reference genome options"},{"location":"params/#qc","text":"Parameter Description Type Default Required Hidden run_checkm Run CheckM QC software boolean apply_filtering Filter assemblies on QC results boolean skip_kraken Don't run Kraken2 taxonomic classification boolean min_n50 Minimum N50 for filtering integer 10000 min_contigs_1000_bp Minimum number of contigs with >1000bp integer 1 min_contig_length Minimum average contig length integer 1","title":"QC"},{"location":"params/#annotation","text":"Parameters for the annotation subworkflow Parameter Description Type Default Required Hidden annotation_tools Comma-separated list of annotation tools to run string mobsuite,rgi,cazy,vfdb,iceberg,bacmet,islandpath,phispy,report bakta_db Path to the BAKTA database string use_prokka Use Prokka (not Bakta) for annotating assemblies boolean min_pident Minimum match identity percentage for filtering integer 60 min_qcover Minimum coverage of each match for filtering number 0.6 skip_profile_creation Skip annotation feature profile creation boolean feature_profile_columns Columns to include in the feature profile string mobsuite,rgi,cazy,vfdb,iceberg,bacmet","title":"Annotation"},{"location":"params/#phylogenomics","text":"Parameters for the phylogenomics subworkflow Parameter Description Type Default Required Hidden skip_phylo Skip Pangenomics and Phylogenomics subworkflow boolean use_ppanggolin Use ppanggolin for calculating the pangenome boolean use_full_alignment Use full alignment boolean use_fasttree Use FastTree boolean True","title":"Phylogenomics"},{"location":"params/#poppunk","text":"Parameters for the lineage subworkflow Parameter Description Type Default Required Hidden skip_poppunk Skip PopPunk boolean poppunk_model Which PopPunk model to use (bgmm, dbscan, refine, threshold or lineage) string run_poppunk_qc Whether to run the QC step for PopPunk boolean enable_subsetting Enable subsetting workflow based on genome similarity boolean core_similarity Similarity threshold for core genomes number 99.99 accessory_similarity Similarity threshold for accessory genes number 99.0","title":"PopPUNK"},{"location":"params/#gene-order","text":"Parameters for the Gene Order Subworkflow Parameter Description Type Default Required Hidden run_gene_order Whether to run the Gene Order subworkflow boolean gene_order_percent_cutoff Cutoff percentage of genomes a gene should be present within to be included in extraction and subsequent analysis. Should a float between 0 and 1 (e.g., 0.25 means only genes present in a minimum of 25% of genomes are kept). number 0.25 gene_order_label_cols If using annotation files predicting features, list of space separated column names to be added to the gene names string None num_neighbors Neighborhood size to extract. Should be an even number N, such that N/2 neighbors upstream and N/2 neighbors downstream will be analyzed. integer 10 inflation Inflation hyperparameter value for Markov Clustering Algorithm. integer 2 epsilon Epsilon hyperparameter value for DBSCAN clustering. number 0.5 minpts Minpts hyperparameter value for DBSCAN clustering. integer 5 plot_clustering Create Clustering HTML Plots boolean","title":"Gene Order"},{"location":"params/#recombination","text":"Parameters for the recombination subworkflow Parameter Description Type Default Required Hidden run_recombination Run Recombination boolean run_verticall Run Verticall recombination tool boolean True run_gubbins Run Gubbins recombination tool boolean","title":"Recombination"},{"location":"params/#dynamics","text":"Parameter Description Type Default Required Hidden run_evolccm Run the community coevolution model boolean run_rspr Run rSPR boolean min_rspr_distance Minimum rSPR distance used to define processing groups integer 10 min_branch_length Minimum rSPR branch length integer 0 max_support_threshold Maximum rSPR support threshold number 0.7 max_approx_rspr Maximum approximate rSPR distance for filtering integer -1 core_gene_tree Core (or reference) genome tree. Used in the rSPR and evolCCM entries. string concatenated_annotation TSV table of annotations for all genomes. Such as the ones generated by Bakta or Prokka in ARETE. string feature_profile Feature profile TSV (A presence-absence matrix). Used in the evolCCM entry. string","title":"Dynamics"},{"location":"params/#institutional-config-options","text":"Parameters used to describe centralised config profiles. These should not be edited. Parameter Description Type Default Required Hidden custom_config_version Git commit id for Institutional configs. string master True custom_config_base Base directory for Institutional configs. Help If you're running offline, Nextflow will not be able to fetch the institutional config files from the internet. If you don't need them, then this is not a problem. If you do need them, you should download the files from the repo and tell Nextflow where to find them with this parameter. string https://raw.githubusercontent.com/nf-core/configs/master True hostnames Institutional configs hostname. string True config_profile_name Institutional config name. string True config_profile_description Institutional config description. string True config_profile_contact Institutional config contact information. string True config_profile_url Institutional config URL link. string True","title":"Institutional config options"},{"location":"params/#max-job-request-options","text":"Set the top limit for requested resources for any single job. Parameter Description Type Default Required Hidden max_cpus Maximum number of CPUs that can be requested for any single job. Help Use to set an upper-limit for the CPU requirement for each process. Should be an integer e.g. --max_cpus 1 integer 16 True max_memory Maximum amount of memory that can be requested for any single job. Help Use to set an upper-limit for the memory requirement for each process. Should be a string in the format integer-unit e.g. --max_memory '8.GB' string 128.GB True max_time Maximum amount of time that can be requested for any single job. Help Use to set an upper-limit for the time requirement for each process. Should be a string in the format integer-unit e.g. --max_time '2.h' string 240.h True","title":"Max job request options"},{"location":"params/#generic-options","text":"Less common options for the pipeline, typically set in a config file. Parameter Description Type Default Required Hidden help Display help text. boolean True publish_dir_mode Method used to save pipeline results to output directory. Help The Nextflow publishDir option specifies which intermediate files should be saved to the output directory. This option tells the pipeline what method should be used to move these files. See Nextflow docs for details. string copy True email_on_fail Email address for completion summary, only when pipeline fails. Help An email address to send a summary email to when the pipeline is completed - ONLY sent if the pipeline does not exit successfully. string True plaintext_email Send plain-text email instead of HTML. boolean True max_multiqc_email_size File size limit when attaching MultiQC reports to summary emails. string 25.MB True monochrome_logs Do not use coloured log outputs. boolean True multiqc_config Custom config file to supply to MultiQC. string True tracedir Directory to keep pipeline Nextflow logs and reports. string ${params.outdir}/pipeline_info True validate_params Boolean whether to validate parameters against the schema at runtime boolean True True show_hidden_params Show all params when using --help Help By default, parameters set as hidden in the schema are not shown on the command line when a user runs with --help . Specifying this option will tell the pipeline to show all parameters. boolean True enable_conda Run this workflow with Conda. You can also use '-profile conda' instead of providing this parameter. boolean True singularity_pull_docker_container Instead of directly downloading Singularity images for use with Singularity, force the workflow to pull and convert Docker containers instead. Help This may be useful for example if you are unable to directly pull Singularity containers to run the pipeline due to http/https proxy issues. boolean True schema_ignore_params string genomes,modules multiqc_logo string True","title":"Generic options"},{"location":"resource_profiles/","text":"ARETE and dataset size Currently ARETE has three distinct profiles that change the pipeline execution in some ways: The default profile (which we can call small ), the medium profile and the large profile. These three profiles were developed based on the size and diversity of the input dataset and change some parameter defaults based on tests we have performed on similar-sized datasets. If you want to first gauge the potential diversity of your dataset and have some input assemblies you can try the PopPUNK entry . One of the outputs will provide insight into how many clusters, or lineages, your dataset divides into. The sizes are: For the default or small profile, we expect datasets with 100 samples/assemblies or fewer. It runs on the default pipeline parameters, with no changes. For the medium profile, we expect datasets with >100 and <1000 samples. It increases the default resource requirements for most processes and also uses PPanGGoLiN for pangenome construction, instead of Panaroo . For the large profile, we expect datasets with >1000 samples. It also increases default resource requirements for some processes and uses PPanGGoLin. Additionally, it enables PopPUNK subsampling , with default parameters .","title":"Dataset Size"},{"location":"resource_profiles/#arete-and-dataset-size","text":"Currently ARETE has three distinct profiles that change the pipeline execution in some ways: The default profile (which we can call small ), the medium profile and the large profile. These three profiles were developed based on the size and diversity of the input dataset and change some parameter defaults based on tests we have performed on similar-sized datasets. If you want to first gauge the potential diversity of your dataset and have some input assemblies you can try the PopPUNK entry . One of the outputs will provide insight into how many clusters, or lineages, your dataset divides into. The sizes are: For the default or small profile, we expect datasets with 100 samples/assemblies or fewer. It runs on the default pipeline parameters, with no changes. For the medium profile, we expect datasets with >100 and <1000 samples. It increases the default resource requirements for most processes and also uses PPanGGoLiN for pangenome construction, instead of Panaroo . For the large profile, we expect datasets with >1000 samples. It also increases default resource requirements for some processes and uses PPanGGoLin. Additionally, it enables PopPUNK subsampling , with default parameters .","title":"ARETE and dataset size"},{"location":"subsampling/","text":"PopPUNK subsetting The subsampling subworkflow is executed if you want to reduce the number of genomes that get added to the phylogenomics subworkflow. By reducing the number of genomes, you can potentially reduce resource requirements for the pangenomics and phylogenomics tools. To enable this subworkflow, add --enable_subsetting when running beiko-lab/ARETE. This will subset genomes based on their core genome similarity and accessory genome similarity, as calculated via their PopPUNK distances. By default, the threshold is --core_similarity 99.9 and --accessory_similarity 99 . But these can be changed by adding these parameters to your execution. What happens then is if any pair of genomes is this similar, only one genome from this pair will be included in the phylogenomic section. All of the removed genome IDs will be present under poppunk_results/removed_genomes.txt . By adding --enable_subsetting , you'll be adding two processes to the execution DAG: POPPUNK_EXTRACT_DISTANCES: This process will extract pair-wise distances between all genomes, returning a table under poppunk_results/distances/ . This table will be used to perform the subsetting. MAKE_HEATMAP: This process will create a heatmap showing different similarity thresholds and the number of genomes that'd be present in each of the possible subsets. It'll also be under poppunk_results/distances/ . Example command The command below will execute the 'annotation' ARETE entry with subsetting enabled, with a core similarity threshold of 99% and an accessory similarity of 95%. nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --enable_subsetting \\ --core_similarity 99 \\ --accessory_similarity 95 \\ -profile docker \\ -entry annotation Be sure to not include --skip_poppunk in your command, because that will then disable all PopPUNK-related processes, including the subsetting subworkflow.","title":"Subsampling"},{"location":"subsampling/#poppunk-subsetting","text":"The subsampling subworkflow is executed if you want to reduce the number of genomes that get added to the phylogenomics subworkflow. By reducing the number of genomes, you can potentially reduce resource requirements for the pangenomics and phylogenomics tools. To enable this subworkflow, add --enable_subsetting when running beiko-lab/ARETE. This will subset genomes based on their core genome similarity and accessory genome similarity, as calculated via their PopPUNK distances. By default, the threshold is --core_similarity 99.9 and --accessory_similarity 99 . But these can be changed by adding these parameters to your execution. What happens then is if any pair of genomes is this similar, only one genome from this pair will be included in the phylogenomic section. All of the removed genome IDs will be present under poppunk_results/removed_genomes.txt . By adding --enable_subsetting , you'll be adding two processes to the execution DAG: POPPUNK_EXTRACT_DISTANCES: This process will extract pair-wise distances between all genomes, returning a table under poppunk_results/distances/ . This table will be used to perform the subsetting. MAKE_HEATMAP: This process will create a heatmap showing different similarity thresholds and the number of genomes that'd be present in each of the possible subsets. It'll also be under poppunk_results/distances/ .","title":"PopPUNK subsetting"},{"location":"subsampling/#example-command","text":"The command below will execute the 'annotation' ARETE entry with subsetting enabled, with a core similarity threshold of 99% and an accessory similarity of 95%. nextflow run beiko-lab/ARETE \\ --input_sample_table samplesheet.csv \\ --enable_subsetting \\ --core_similarity 99 \\ --accessory_similarity 95 \\ -profile docker \\ -entry annotation Be sure to not include --skip_poppunk in your command, because that will then disable all PopPUNK-related processes, including the subsetting subworkflow.","title":"Example command"},{"location":"usage/","text":"beiko-lab/ARETE: Usage Introduction The ARETE pipeline can is designed as an end-to-end workflow manager for genome assembly, annotation, and phylogenetic analysis, beginning with read data. However, in some cases a user may wish to stop the pipeline prior to annotation or use the annotation features of the work flow with pre-existing assemblies. Therefore, ARETE allows users different use cases: Run the full pipeline end-to-end. Input a set of reads and stop after assembly. Input a set of assemblies and perform QC. Input a set of assemblies and perform annotation and taxonomic analyses. Input a set of assemblies and perform genome clustering with PopPUNK. Input a set of assemblies and perform phylogenomic and pangenomic analysis. This document will describe how to perform each workflow. \"Running the pipeline\" will show some example command on how to use these different entries to ARETE. Samplesheet input No matter your use case, you will need to create a samplesheet with information about the samples you would like to analyse before running the pipeline. Use this parameter to specify its location. For full runs and assembly, it has to be a comma-separated file with 3 columns, and a header row as shown in the examples below. --input_sample_table '[path to samplesheet file]' Full workflow or assembly samplesheet The pipeline will auto-detect whether a sample is single- or paired-end using the information provided in the samplesheet. The samplesheet can have as many columns as you desire, however, there is a strict requirement for the first 4 columns to match those defined in the table below. A final samplesheet file consisting of both single- and paired-end data may look something like the one below. This is for 6 samples, where TREATMENT_REP3 has been sequenced twice. sample,fastq_1,fastq_2 CONTROL_REP1,AEG588A1_S1_L002_R1_001.fastq.gz,AEG588A1_S1_L002_R2_001.fastq.gz CONTROL_REP2,AEG588A2_S2_L002_R1_001.fastq.gz,AEG588A2_S2_L002_R2_001.fastq.gz CONTROL_REP3,AEG588A3_S3_L002_R1_001.fastq.gz,AEG588A3_S3_L002_R2_001.fastq.gz TREATMENT_REP1,AEG588A4_S4_L003_R1_001.fastq.gz, TREATMENT_REP2,AEG588A5_S5_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L004_R1_001.fastq.gz, Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fastq_1 Full path to FastQ file for Illumina short reads 1. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". fastq_2 Full path to FastQ file for Illumina short reads 2. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". An example samplesheet has been provided with the pipeline. Annotation only samplesheet The ARETE pipeline allows users to provide pre-existing assemblies to make use of the annotation and reporting features of the workflow. Users may use the assembly_qc entry point to perform QC on the assemblies. Note that the QC workflow does not automatically filter low quality assemblies, it simply generates QC reports! annotation , assembly_qc and poppunk workflows accept the same format of sample sheet. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fna_file_path Full path to fna file for assembly or genome. File must have .fna file extension. An example samplesheet has been provided with the pipeline. Phylogenomics and Pangenomics only samplesheet The ARETE pipeline allows users to provide pre-existing assemblies to make use of the phylogenomic and pangenomic features of the workflow. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. gff_file_path Full path to GFF file for assembly or genome. File must have .gff or .gff3 file extension. These files can be the ones generated by Prokka or Bakta in ARETE's annotation subworkflow. Reference Genome For full workflow or assembly, users may provide a path to a reference genome in fasta format for use in assembly evaluation. --reference_genome ref.fasta Running the pipeline The typical command for running the pipeline is as follows: nextflow run beiko-lab/ARETE --input_sample_table samplesheet.csv --reference_genome ref.fasta --poppunk_model bgmm -profile docker This will launch the pipeline with the docker configuration profile. See below for more information about profiles. Note that the pipeline will create the following files in your working directory: work # Directory containing the nextflow working files results # Finished results (configurable, see below) .nextflow_log # Log file from Nextflow # Other nextflow hidden files, eg. history of pipeline runs and old logs. As written above, the pipeline also allows users to execute only assembly or only annotation. Assembly Entry To execute assembly (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker Assembly QC Entry To execute QC on pre-existing assemblies (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly_qc --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker Annotation Entry To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry annotation --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker PopPUNK Entry To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry poppunk --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker Phylogenomics and Pangenomics Entry To execute phylogenomic and pangenomics analysis on pre-existing assemblies: nextflow run beiko-lab/ARETE -entry phylogenomics --input_sample_table samplesheet.csv -profile docker rSPR Entry To execute the rSPR analysis on pre-existing trees: nextflow run beiko-lab/ARETE \\ -entry rspr \\ --input_sample_table samplesheet.csv \\ --core_gene_tree core_gene_alignment.tre \\ --concatenated_annotation BAKTA.txt \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --concatenated_annotation - The tabular annotation results (TSV) for all genomes, like the ones generated at the end of Prokka or Bakta in ARETE. Although useful, it's not necessary to execute the rSPR entry. --input_sample_table - A samplesheet containing all individual gene trees in the following format: gene_tree,path CDS_0000,/path/to/CDS_0000.tre CDS_0001,/path/to/CDS_0001.tre CDS_0002,/path/to/CDS_0002.tre CDS_0003,/path/to/CDS_0003.tre CDS_0004,/path/to/CDS_0004.tre evolCCM Entry To execute the evolCCM analysis on a pre-existing reference tree and feature profile: nextflow run beiko-lab/ARETE \\ -entry evolccm \\ --core_gene_tree core_gene_alignment.tre \\ --feature_profile feature_profile.tsv.gz \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --feature_profile - A presence/absence TSV matrix of features in genomes. Genome names should be the same in the core tree and should be contained to a 'genome_id' column, with all other columns represent features absent (0) or present (1) in each genome. I.e.: genome_id plasmid_AA155 plasmid_AA161 ED010 0 0 ED017 0 1 ED040 0 0 ED073 0 1 ED075 1 1 ED082 0 1 ED142 0 1 ED178 0 1 ED180 0 0 Recombination Entry To execute the recombination analysis on pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE \\ -entry recombination \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ -profile docker Gene Order Entry To execute the Gene Order analysis on pre-existing assemblies and RGI annotations: nextflow run beiko-lab/ARETE \\ -entry gene_order \\ --input_sample_table gene_order_samplesheet.csv \\ -profile docker --input_sample_table - A samplesheet containing a fasta file, a genbank file and an RGI output file for each assembly: sample,fna_file_path,gbk,rgi SAMD00052607,SAMD00052607.faa,SAMD00052607.gbk,SAMD00052607_rgi.txt SAMEA1466699,SAMEA1466699.faa,SAMEA1466699.gbk,SAMEA1466699_rgi.txt SAMEA1486355,SAMEA1486355.faa,SAMEA1486355.gbk,SAMEA1486355_rgi.txt Updating the pipeline When you run the above command, Nextflow automatically pulls the pipeline code from GitHub and stores it as a cached version. When running the pipeline after this, it will always use the cached version if available - even if the pipeline has been updated since. To make sure that you're running the latest version of the pipeline, make sure that you regularly update the cached version of the pipeline: nextflow pull beiko-lab/ARETE Reproducibility It's a good idea to specify a pipeline version when running the pipeline on your data. This ensures that a specific version of the pipeline code and software are used when you run your pipeline. If you keep using the same tag, you'll be running the same version of the pipeline, even if there have been changes to the code since. First, go to the ARETE releases page and find the latest version number - numeric only (eg. 1.3.1 ). Then specify this when running the pipeline with -r (one hyphen) - eg. -r 1.3.1 . This version number will be logged in reports when you run the pipeline, so that you'll know what you used when you look back in the future. Core Nextflow arguments NB: These options are part of Nextflow and use a single hyphen (pipeline parameters use a double-hyphen). -profile Use this parameter to choose a configuration profile. Profiles can give configuration presets for different compute environments. Several generic profiles are bundled with the pipeline which instruct the pipeline to use software packaged using different methods (Docker, Singularity, Podman, Shifter, Charliecloud) - see below. We highly recommend the use of Docker or Singularity containers for full pipeline reproducibility. The pipeline also dynamically loads configurations from https://github.com/nf-core/configs when it runs, making multiple config profiles for various institutional clusters available at run time. For more information and to see if your system is available in these configs please see the nf-core/configs documentation . Note that multiple profiles can be loaded, for example: -profile test,docker - the order of arguments is important! They are loaded in sequence, so later profiles can overwrite earlier profiles. If -profile is not specified, the pipeline will run locally and expect all software to be installed and available on the PATH . This is not recommended. docker A generic configuration profile to be used with Docker singularity A generic configuration profile to be used with Singularity podman A generic configuration profile to be used with Podman shifter A generic configuration profile to be used with Shifter charliecloud A generic configuration profile to be used with Charliecloud conda Please only use Conda as a last resort i.e. when it's not possible to run the pipeline with Docker, Singularity, Podman, Shifter or Charliecloud. A generic configuration profile to be used with Conda Pulls most software from Bioconda test A profile with a complete configuration for automated testing Can run in personal computers with at least 6GB of RAM and 2 CPUs Includes links to test data so needs no other parameters -resume Specify this when restarting a pipeline. Nextflow will used cached results from any pipeline steps where the inputs are the same, continuing from where it got to previously. You can also supply a run name to resume a specific run: -resume [run-name] . Use the nextflow log command to show previous run names. -c Specify the path to a specific config file (this is a core Nextflow command). See the nf-core website documentation for more information. Custom resource requests Each step in the pipeline has a default set of requirements for number of CPUs, memory and time. For most of the steps in the pipeline, if the job exits with an error code of 143 (exceeded requested resources) it will automatically resubmit with higher requests (2 x original, then 3 x original). If it still fails after three times then the pipeline is stopped. Whilst these default requirements will hopefully work for most people with most data, you may find that you want to customise the compute resources that the pipeline requests. You can do this by creating a custom config file. For example, to give the workflow process UNICYCLER 32GB of memory, you could use the following config: process { withName: UNICYCLER { memory = 32.GB } } To find the exact name of a process you wish to modify the compute resources, check the live-status of a nextflow run displayed on your terminal or check the nextflow error for a line like so: Error executing process > 'bwa' . In this case the name to specify in the custom config file is bwa . See the main Nextflow documentation for more information. Running in the background Nextflow handles job submissions and supervises the running jobs. The Nextflow process must run until the pipeline is finished. The Nextflow -bg flag launches Nextflow in the background, detached from your terminal so that the workflow does not stop if you log out of your session. The logs are saved to a file. Alternatively, you can use screen / tmux or similar tool to create a detached session which you can log back into at a later time. Some HPC setups also allow you to run nextflow within a cluster job submitted your job scheduler (from where it submits more jobs). Nextflow memory requirements In some cases, the Nextflow Java virtual machines can start to request a large amount of memory. We recommend adding the following line to your environment to limit this (typically in ~/.bashrc or ~./bash_profile ): NXF_OPTS='-Xms1g -Xmx4g' Sometimes LevelDB, which is used by Nextflow to track execution metadata, can lead to memory-related issues, often showing as a SIGBUS error. This tends to happen when running Nextflow in SLURM environments . In this case, setting NXF_OPTS=\"-Dleveldb.mmap=false\" in your ~/.bashrc or immediately before executing nextflow run usually solves the issue.","title":"Usage"},{"location":"usage/#beiko-labarete-usage","text":"","title":"beiko-lab/ARETE: Usage"},{"location":"usage/#introduction","text":"The ARETE pipeline can is designed as an end-to-end workflow manager for genome assembly, annotation, and phylogenetic analysis, beginning with read data. However, in some cases a user may wish to stop the pipeline prior to annotation or use the annotation features of the work flow with pre-existing assemblies. Therefore, ARETE allows users different use cases: Run the full pipeline end-to-end. Input a set of reads and stop after assembly. Input a set of assemblies and perform QC. Input a set of assemblies and perform annotation and taxonomic analyses. Input a set of assemblies and perform genome clustering with PopPUNK. Input a set of assemblies and perform phylogenomic and pangenomic analysis. This document will describe how to perform each workflow. \"Running the pipeline\" will show some example command on how to use these different entries to ARETE.","title":"Introduction"},{"location":"usage/#samplesheet-input","text":"No matter your use case, you will need to create a samplesheet with information about the samples you would like to analyse before running the pipeline. Use this parameter to specify its location. For full runs and assembly, it has to be a comma-separated file with 3 columns, and a header row as shown in the examples below. --input_sample_table '[path to samplesheet file]'","title":"Samplesheet input"},{"location":"usage/#full-workflow-or-assembly-samplesheet","text":"The pipeline will auto-detect whether a sample is single- or paired-end using the information provided in the samplesheet. The samplesheet can have as many columns as you desire, however, there is a strict requirement for the first 4 columns to match those defined in the table below. A final samplesheet file consisting of both single- and paired-end data may look something like the one below. This is for 6 samples, where TREATMENT_REP3 has been sequenced twice. sample,fastq_1,fastq_2 CONTROL_REP1,AEG588A1_S1_L002_R1_001.fastq.gz,AEG588A1_S1_L002_R2_001.fastq.gz CONTROL_REP2,AEG588A2_S2_L002_R1_001.fastq.gz,AEG588A2_S2_L002_R2_001.fastq.gz CONTROL_REP3,AEG588A3_S3_L002_R1_001.fastq.gz,AEG588A3_S3_L002_R2_001.fastq.gz TREATMENT_REP1,AEG588A4_S4_L003_R1_001.fastq.gz, TREATMENT_REP2,AEG588A5_S5_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L003_R1_001.fastq.gz, TREATMENT_REP3,AEG588A6_S6_L004_R1_001.fastq.gz, Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fastq_1 Full path to FastQ file for Illumina short reads 1. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". fastq_2 Full path to FastQ file for Illumina short reads 2. File has to be gzipped and have the extension \".fastq.gz\" or \".fq.gz\". An example samplesheet has been provided with the pipeline.","title":"Full workflow or assembly samplesheet"},{"location":"usage/#annotation-only-samplesheet","text":"The ARETE pipeline allows users to provide pre-existing assemblies to make use of the annotation and reporting features of the workflow. Users may use the assembly_qc entry point to perform QC on the assemblies. Note that the QC workflow does not automatically filter low quality assemblies, it simply generates QC reports! annotation , assembly_qc and poppunk workflows accept the same format of sample sheet. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. fna_file_path Full path to fna file for assembly or genome. File must have .fna file extension. An example samplesheet has been provided with the pipeline.","title":"Annotation only samplesheet"},{"location":"usage/#phylogenomics-and-pangenomics-only-samplesheet","text":"The ARETE pipeline allows users to provide pre-existing assemblies to make use of the phylogenomic and pangenomic features of the workflow. The sample sheet must be a 2 column, comma-seperated CSV file with header. Column Description sample Custom sample name. This entry will be identical for multiple sequencing libraries/runs from the same sample. gff_file_path Full path to GFF file for assembly or genome. File must have .gff or .gff3 file extension. These files can be the ones generated by Prokka or Bakta in ARETE's annotation subworkflow.","title":"Phylogenomics and Pangenomics only samplesheet"},{"location":"usage/#reference-genome","text":"For full workflow or assembly, users may provide a path to a reference genome in fasta format for use in assembly evaluation. --reference_genome ref.fasta","title":"Reference Genome"},{"location":"usage/#running-the-pipeline","text":"The typical command for running the pipeline is as follows: nextflow run beiko-lab/ARETE --input_sample_table samplesheet.csv --reference_genome ref.fasta --poppunk_model bgmm -profile docker This will launch the pipeline with the docker configuration profile. See below for more information about profiles. Note that the pipeline will create the following files in your working directory: work # Directory containing the nextflow working files results # Finished results (configurable, see below) .nextflow_log # Log file from Nextflow # Other nextflow hidden files, eg. history of pipeline runs and old logs. As written above, the pipeline also allows users to execute only assembly or only annotation.","title":"Running the pipeline"},{"location":"usage/#assembly-entry","text":"To execute assembly (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker","title":"Assembly Entry"},{"location":"usage/#assembly-qc-entry","text":"To execute QC on pre-existing assemblies (reference genome optional): nextflow run beiko-lab/ARETE -entry assembly_qc --input_sample_table samplesheet.csv --reference_genome ref.fasta -profile docker","title":"Assembly QC Entry"},{"location":"usage/#annotation-entry","text":"To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry annotation --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker","title":"Annotation Entry"},{"location":"usage/#poppunk-entry","text":"To execute annotation of pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE -entry poppunk --input_sample_table samplesheet.csv --poppunk_model bgmm -profile docker","title":"PopPUNK Entry"},{"location":"usage/#phylogenomics-and-pangenomics-entry","text":"To execute phylogenomic and pangenomics analysis on pre-existing assemblies: nextflow run beiko-lab/ARETE -entry phylogenomics --input_sample_table samplesheet.csv -profile docker","title":"Phylogenomics and Pangenomics Entry"},{"location":"usage/#rspr-entry","text":"To execute the rSPR analysis on pre-existing trees: nextflow run beiko-lab/ARETE \\ -entry rspr \\ --input_sample_table samplesheet.csv \\ --core_gene_tree core_gene_alignment.tre \\ --concatenated_annotation BAKTA.txt \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --concatenated_annotation - The tabular annotation results (TSV) for all genomes, like the ones generated at the end of Prokka or Bakta in ARETE. Although useful, it's not necessary to execute the rSPR entry. --input_sample_table - A samplesheet containing all individual gene trees in the following format: gene_tree,path CDS_0000,/path/to/CDS_0000.tre CDS_0001,/path/to/CDS_0001.tre CDS_0002,/path/to/CDS_0002.tre CDS_0003,/path/to/CDS_0003.tre CDS_0004,/path/to/CDS_0004.tre","title":"rSPR Entry"},{"location":"usage/#evolccm-entry","text":"To execute the evolCCM analysis on a pre-existing reference tree and feature profile: nextflow run beiko-lab/ARETE \\ -entry evolccm \\ --core_gene_tree core_gene_alignment.tre \\ --feature_profile feature_profile.tsv.gz \\ -profile docker The parameters being: --core_gene_tree - The reference tree, coming from a core genome alignment, like the one generated by panaroo in ARETE. --feature_profile - A presence/absence TSV matrix of features in genomes. Genome names should be the same in the core tree and should be contained to a 'genome_id' column, with all other columns represent features absent (0) or present (1) in each genome. I.e.: genome_id plasmid_AA155 plasmid_AA161 ED010 0 0 ED017 0 1 ED040 0 0 ED073 0 1 ED075 1 1 ED082 0 1 ED142 0 1 ED178 0 1 ED180 0 0","title":"evolCCM Entry"},{"location":"usage/#recombination-entry","text":"To execute the recombination analysis on pre-existing assemblies (PopPUNK model can be either bgmm, dbscan, refine, threshold or lineage): nextflow run beiko-lab/ARETE \\ -entry recombination \\ --input_sample_table samplesheet.csv \\ --poppunk_model dbscan \\ -profile docker","title":"Recombination Entry"},{"location":"usage/#gene-order-entry","text":"To execute the Gene Order analysis on pre-existing assemblies and RGI annotations: nextflow run beiko-lab/ARETE \\ -entry gene_order \\ --input_sample_table gene_order_samplesheet.csv \\ -profile docker --input_sample_table - A samplesheet containing a fasta file, a genbank file and an RGI output file for each assembly: sample,fna_file_path,gbk,rgi SAMD00052607,SAMD00052607.faa,SAMD00052607.gbk,SAMD00052607_rgi.txt SAMEA1466699,SAMEA1466699.faa,SAMEA1466699.gbk,SAMEA1466699_rgi.txt SAMEA1486355,SAMEA1486355.faa,SAMEA1486355.gbk,SAMEA1486355_rgi.txt","title":"Gene Order Entry"},{"location":"usage/#updating-the-pipeline","text":"When you run the above command, Nextflow automatically pulls the pipeline code from GitHub and stores it as a cached version. When running the pipeline after this, it will always use the cached version if available - even if the pipeline has been updated since. To make sure that you're running the latest version of the pipeline, make sure that you regularly update the cached version of the pipeline: nextflow pull beiko-lab/ARETE","title":"Updating the pipeline"},{"location":"usage/#reproducibility","text":"It's a good idea to specify a pipeline version when running the pipeline on your data. This ensures that a specific version of the pipeline code and software are used when you run your pipeline. If you keep using the same tag, you'll be running the same version of the pipeline, even if there have been changes to the code since. First, go to the ARETE releases page and find the latest version number - numeric only (eg. 1.3.1 ). Then specify this when running the pipeline with -r (one hyphen) - eg. -r 1.3.1 . This version number will be logged in reports when you run the pipeline, so that you'll know what you used when you look back in the future.","title":"Reproducibility"},{"location":"usage/#core-nextflow-arguments","text":"NB: These options are part of Nextflow and use a single hyphen (pipeline parameters use a double-hyphen).","title":"Core Nextflow arguments"},{"location":"usage/#-profile","text":"Use this parameter to choose a configuration profile. Profiles can give configuration presets for different compute environments. Several generic profiles are bundled with the pipeline which instruct the pipeline to use software packaged using different methods (Docker, Singularity, Podman, Shifter, Charliecloud) - see below. We highly recommend the use of Docker or Singularity containers for full pipeline reproducibility. The pipeline also dynamically loads configurations from https://github.com/nf-core/configs when it runs, making multiple config profiles for various institutional clusters available at run time. For more information and to see if your system is available in these configs please see the nf-core/configs documentation . Note that multiple profiles can be loaded, for example: -profile test,docker - the order of arguments is important! They are loaded in sequence, so later profiles can overwrite earlier profiles. If -profile is not specified, the pipeline will run locally and expect all software to be installed and available on the PATH . This is not recommended. docker A generic configuration profile to be used with Docker singularity A generic configuration profile to be used with Singularity podman A generic configuration profile to be used with Podman shifter A generic configuration profile to be used with Shifter charliecloud A generic configuration profile to be used with Charliecloud conda Please only use Conda as a last resort i.e. when it's not possible to run the pipeline with Docker, Singularity, Podman, Shifter or Charliecloud. A generic configuration profile to be used with Conda Pulls most software from Bioconda test A profile with a complete configuration for automated testing Can run in personal computers with at least 6GB of RAM and 2 CPUs Includes links to test data so needs no other parameters","title":"-profile"},{"location":"usage/#-resume","text":"Specify this when restarting a pipeline. Nextflow will used cached results from any pipeline steps where the inputs are the same, continuing from where it got to previously. You can also supply a run name to resume a specific run: -resume [run-name] . Use the nextflow log command to show previous run names.","title":"-resume"},{"location":"usage/#-c","text":"Specify the path to a specific config file (this is a core Nextflow command). See the nf-core website documentation for more information.","title":"-c"},{"location":"usage/#custom-resource-requests","text":"Each step in the pipeline has a default set of requirements for number of CPUs, memory and time. For most of the steps in the pipeline, if the job exits with an error code of 143 (exceeded requested resources) it will automatically resubmit with higher requests (2 x original, then 3 x original). If it still fails after three times then the pipeline is stopped. Whilst these default requirements will hopefully work for most people with most data, you may find that you want to customise the compute resources that the pipeline requests. You can do this by creating a custom config file. For example, to give the workflow process UNICYCLER 32GB of memory, you could use the following config: process { withName: UNICYCLER { memory = 32.GB } } To find the exact name of a process you wish to modify the compute resources, check the live-status of a nextflow run displayed on your terminal or check the nextflow error for a line like so: Error executing process > 'bwa' . In this case the name to specify in the custom config file is bwa . See the main Nextflow documentation for more information.","title":"Custom resource requests"},{"location":"usage/#running-in-the-background","text":"Nextflow handles job submissions and supervises the running jobs. The Nextflow process must run until the pipeline is finished. The Nextflow -bg flag launches Nextflow in the background, detached from your terminal so that the workflow does not stop if you log out of your session. The logs are saved to a file. Alternatively, you can use screen / tmux or similar tool to create a detached session which you can log back into at a later time. Some HPC setups also allow you to run nextflow within a cluster job submitted your job scheduler (from where it submits more jobs).","title":"Running in the background"},{"location":"usage/#nextflow-memory-requirements","text":"In some cases, the Nextflow Java virtual machines can start to request a large amount of memory. We recommend adding the following line to your environment to limit this (typically in ~/.bashrc or ~./bash_profile ): NXF_OPTS='-Xms1g -Xmx4g' Sometimes LevelDB, which is used by Nextflow to track execution metadata, can lead to memory-related issues, often showing as a SIGBUS error. This tends to happen when running Nextflow in SLURM environments . In this case, setting NXF_OPTS=\"-Dleveldb.mmap=false\" in your ~/.bashrc or immediately before executing nextflow run usually solves the issue.","title":"Nextflow memory requirements"}]} \ No newline at end of file diff --git a/sitemap.xml.gz b/sitemap.xml.gz index 5bbd9eea6f89d73d294b0b5ad849a402c51ab776..08044c1220e76700432f51571f5c5325a6545049 100644 GIT binary patch delta 13 Ucmb=gXP58h;5fD=Vj_D503QevolCCM Entry
  • Recombination Entry +
  • +
  • Gene Order Entry
  • @@ -347,6 +349,21 @@

    Recombination Entry

    --poppunk_model dbscan \ -profile docker +

    Gene Order Entry

    +

    To execute the Gene Order analysis on pre-existing assemblies and RGI annotations:

    +
    nextflow run beiko-lab/ARETE \
    +  -entry gene_order \
    +  --input_sample_table gene_order_samplesheet.csv \
    +  -profile docker
    +
    +
      +
    • --input_sample_table - A samplesheet containing a fasta file, a genbank file and an RGI output file for each assembly:
    • +
    +
    sample,fna_file_path,gbk,rgi
    +SAMD00052607,SAMD00052607.faa,SAMD00052607.gbk,SAMD00052607_rgi.txt
    +SAMEA1466699,SAMEA1466699.faa,SAMEA1466699.gbk,SAMEA1466699_rgi.txt
    +SAMEA1486355,SAMEA1486355.faa,SAMEA1486355.gbk,SAMEA1486355_rgi.txt
    +

    Updating the pipeline

    When you run the above command, Nextflow automatically pulls the pipeline code from GitHub and stores it as a cached version. When running the pipeline after this, it will always use the cached version if available - even if the pipeline has been updated since. To make sure that you're running the latest version of the pipeline, make sure that you regularly update the cached version of the pipeline:

    nextflow pull beiko-lab/ARETE