Releases: stan-dev/stan
Releases · stan-dev/stan
v2.13.0
v2.13.0 (25 November 2016)
New Team Members
- Thel Seraphim (Columbia University) -- Stan and Math libraries
- Vincent Picaud (CEA, France) -- MathematicaStan
Bug Fixes
- generated code for lower truncation fixed for discrete variables (#2054)
- typo in error messages for RNGs only allowed in transformed data
block (#2124) - variables ending in _lpdf are now allowed (#2123)
- cov_exp_quad() not compiling in C++ (#2113)
- conditional operator in functions not compiling in C++ (#2101)
- off-by-one error in error message for integrate_ode_bdf() fixed (#2073)
New User-Facing Features
- vectorization of unary functions (#2119, #2037)
- bernoulli_logit_rng() added to language (#2084)
- Jacobian warning now suggests
target +=(#2066) - matrix_exp() function now available (#2043)
- compound declaration / definition statements (#1951)
New Internal Features
- user-defined functions can be declared and not defined (#2068)
Documentation
v2.12.0
v2.12.0 (6 September 2016)
New Team Members
- Charles Margossian (Metrum Research Group, LLC) -- diff eq models
Known, outstanding bugs
- Normalization for lower truncated discrete distributions is off
by one (#2054)
Bug Fixes
- updated the math library to deal with g++6 errors
(math#348, math#345, math#327) - fixed issue with 1:0 indexing in array slicing producing
huge size (#1960) - allow variables to be named the same as unary functions
(#1459) - distributions without cumulative distribution functions are now
prevented from being used in truncation (#2052) - allow user-defined lpdf functions to be used in
target +=
expressions (#1944) - fixed duplicate function warning message (#1946)
- fixed ternary operator generating C++ code that doesn't compile
(#2044)
New User-Facing Features
- added mdivide_left_spd() and mdivide_right_spd() to the language
(#2016) - added cov_exp_quad() to the language (#1957)
- added break and continue statements to the language (#1756)
- added append_row() and append_col() to the language (#1347)
- updated error message for missing
|when using probability functions
(#2003) - updated parser/warning errors for new syntax (#1940)
New Internal Features
v2.11.0
v2.11.0 (27 July 2016)
Bug Fixes
v2.10.0
v2.10.0 (17 June 2016)
New Team Members
- Aki Vehtari (Aalto Uni) --- GPs, LOO, statistical modeling,
MATLAB - Rayleigh Lei (U. Michigan) --- vectorizing functions
- Sebastian Weber (Novartis) --- diff eq models
New User-Facing Features
- introduction of = for assignment; deprecating <- (#1819)
- _lpdf, _lpmf, _lcdf, _lccdf and vertical bar notation
for probability mass, density, and cumulative distribution
functions; update user-defined function syntax;
deprecate older forms (#1811) - target += lp to replace increment_log_prob(lp); target() to rpelace
log_prob() function; remove lp__ variable access (#1800) - conditional (ternary) operator (#1125)
- stiff diff eq solver CVODES from Sundials
- add control parameters (tolerance, max iteraitons) to ODE solvers
- rename ODE solvers based on algorithm, integrate_ode_rk45
for existing non-stiff Runge-Kutta solver and integrate_ode_bdf
for the stiff backward differentiation form; deprecate
the unmarked integrate_ode function (#1886) - limiting diff eq iterations in solvers (Boost/CVODES)
- unit_vector as parameter (#1713) [it never worked in the past]
- rename multiply_log and log_binomial_coefficient
to lmultiply and lchoose (also part of #1811) - incomplete beta function as inc_beta (#1540)
New Internal Features
- exhaustive HMC (XHMC)
- multinomial variant of NUTS (#1846)
- simplified NUTS criterion (#1852)
- uniform static HMC (#1849)
- Riemannian HMC with SoftAbs (#304)
- refactor semantic actions to header/implementation (#1795)
- writer callback interface for algorithms (#1739)
- error writer callbacks (#1836)
- cleaner HMC diagnostic ouptut (#1790)
- pull-request template on GitHub (#1794)
- remove unused filename argument in parser (#1752)
- remove unused lp_no_jacobian function (now controlled with
template params) - writer callback interface for HMC (#1680)
Bug Fixes
- index slicing not compiling for matrices (#1754, #1810)
- parsing too eager for reserved words (#1077)
- CSV reader whitespace sensitivity (#1907)
- delayed reject multinomial NUTS (#1906)
- PYTHON2 variable for cpplint (#1912)
- size zero vectors, row vectors, matrices (#1909)
- overflowing transition probs in multinomial NUTS (#1904)
- all the manual bug fixes (#1431, #1709, #1734, #1753,
#1826, #1874, #1879, #1890) - compiler fix for vector /. real to compile (#1883)
- terminate infinite loops in optimization (#1896)
- ADVI argument config failure (#1731)
- double struct creation removed for forward-declaration
of user-defined functions bug (#1079) - clarify deprecation of <- is only in Stan, not R (#1819)
- memory leaking with command getting in way of debug (#1293)
- makefile dependencies for stanc (#1682)
- remove NaN init for integers in place of min (negative)
value (#1766) - namespace pollution from log_mix removed (#1707)
- message buffering reduced in ADVI
- add pstream to user-defined CDFs supporting truncation (#1628)
- initialization failure message clarification (#1736)
v2.9.0
v2.9.0 (3 Dec 2015)
New Team Members
- Krzysztof Sakrejda, adding functions in Stan Math and Stan
New Features
- General R-like indexing for arrays and vectors! (#1057)
- Automatic differentation paper: http://arxiv.org/abs/1509.07164
- Changed unit_vector transformation (#1720)
- Automatically detect bitness for Windows 32-bit (#1614)
- Better adaptation for ADVI (#1550, #1559, #1587, #1591, #1593, #1710)
Bug Fixes
Maintenance
- Reduced Cpplint warnings to 0! (#1651, #1656, #1701)
- Better ADVI output (#1588)
- Updated logo for doxygen (#1631)
- Refactoring the use of output streams (#1602, 1653, #1662,
#1670, #1676, #1696, #1601) - Moved src/stan/services files and namespaces (#1616)
- Updated clang++ compiler options for Mac OS X (#1668)
- Additional testing on Travis CI (#1708)
- Removed unused python script (#1611)
- Updated makefile (#1618, #1655, #1659)
Manual
v2.8.0
v2.8.0 (8 Sep 2015)
New Team Members
- Dustin Tran, helps make variational inference a reality
New Features
- Inverse CDF for normal distribution (#1157)
- Compiler error messages are friendlier
- operator
- assignment (#1543, stan-dev/math#146)
- Sparse matrix by dense vector multiplication (#1565)
Bug Fixes
- Disallow user-defiend function overloading (#1547)
- Better behaved beta cdf (#1561, stan-dev/math#124)
Manual
- Corrected all known bug reports and typos; details in
issues #1526, #1548, #1577 - Thanks to @jonathan-g (Github handle), Damjan Vukcevic,
Daniel Chen, Guido Biele, Takahiro Kubo, Andy Choi
v2.7.0
v2.7.0 (9 July 2015)
New Team Members
- Alp Kucukelbir, who brings you variational inference
- Robert L. Grant, who brings you the StataStan interface
Major New Feature
- Black-box variational inference, mean field and full
rank (#1505)
New Features
- Line numbers reported for runtime errors (#1195)
- Wiener first passage time density (#765) (thanks to
Michael Schvartsman) - Partial initialization (#1069)
- NegBinomial2 RNG (#1471) and PoissonLog RNG (#1458) and extended
range for Dirichlet RNG (#1474) and fixed Poisson RNG for older
Mac compilers (#1472) - Error messages now use operator notation (#1401)
- More specific error messages for illegal assignments (#1100)
- More specific error messages for illegal sampling statement
signatures (#1425) - Extended range on ibeta derivatives with wide impact on CDFs (#1426)
- Display initialization error messages (#1403)
- Works with Intel compilers and GCC 4.4 (#1506, #1514, #1519)
Bug Fixes
- Allow functions ending in _lp to call functions ending in _lp (#1500)
- Update warnings to catch uses of illegal sampling functions like
CDFs and updated declared signatures (#1152) - Disallow constraints on local variables (#1295)
- Allow min() and max() in variable declaration bounds and remove
unnecessary use of math.h and top-level :: namespace (#1436) - Updated exponential lower bound check (#1179)
- Extended sum to work with zero size arrays (#1443)
- Positive definiteness checks fixed (were > 1e-8, now > 0) (#1386)
Code Reorganization and Back End Upgrades
- New static constants (#469, #765)
- Added major/minor/patch versions as properties (#1383)
- Pulled all math-like functionality into stan::math namespace
- Pulled the Stan Math Library out into its own repository (#1520)
- Included in Stan C++ repository as submodule
- Removed final usage of std::cout and std::cerr (#699) and
updated tests for null streams (#1239) - Removed over 1000 CppLint warnings
- Remove model write CSV methods (#445)
- Reduced generality of operators in fvar (#1198)
- Removed folder-level includes due to order issues (part of Math
reorg) and include math.hpp include (#1438) - Updated to Boost 1.58 (#1457)
- Travis continuous integration for Linux (#607)
- Add grad() method to math::var for autodiff to encapsulate math::vari
- Added finite diff functionals for testing (#1271)
- More configurable distribution unit tests (#1268)
- Clean up directory-level includes (#1511)
- Removed all lint from new math lib and add cpplint to build lib
(#1412) - Split out derivative functionals (#1389)
Manual and Documentation
- New Logo in Manual; remove old logos (#1023)
- Corrected all known bug reports and typos; details in
issues #1420, #1508, #1496 - Thanks to Sunil Nandihalli, Andy Choi, Sebastian Weber,
Heraa Hu, @jonathan-g (GitHub handle), M. B. Joseph, Damjan
Vukcevic, @tosh1ki (GitHub handle), Juan S. Casallas - Fix some parsing issues for index (#1498)
- Added chapter on variational inference
- Added strangely unrelated regressions and multivariate probit
examples - Discussion from Ben Goodrich about reject() and sampling
- Start to reorganize code with fast examples first, then
explanations - Added CONTRIBUTING.md file (#1408)
Note: If downloading source, use the green links to download a copy that includes the Stan Math Library.
v2.6.3
v2.6.3 (17 March 2015)
Bug Fixes
- removed extra semicolons in code; they are techinically a C++11
extension [#1379] - removed use of variable length array; this is a C99 feature
[#1381]
Features
- error messages in initialization during sampling get printed out
to the message stream [#1326]
Test Enhancements
- added tests for is_vector_like [#1360]
v2.6.2
v2.6.2 (14 March 2015)
Bug Fixes
- fixed append_row and append_col to return the corect types [#1297]
- documentation fix: fixed generation of stan-functions.txt [#1327]
- changed constructor of stan::agrad::var so it can compile without
C++11 [#1337, 1364] - fixed bug in matrix double-indexing [#1342]
- fixed calculation of derivative in normal_cdf_log, normal_ccdf_log
[#1357]
API Enhancements
v2.6.1
v2.6.1 (3 March 2015)
Language Enhancements
- improved warning messages from the parser [#1244]
API/Build/Test Enhancements
- the C++ API now throws 3 types of exceptions [#685]
- parts of the C++ API were rearranged (gm, common) [#942]
- replace promote_args
- allow reading from a local makefile outside of the code repository [#1277]
- added gradient of hessian functional in C++ [#1272]
- supressed printing of adjoints when printing stan::agrad::var [#330]
- removed some cpplint warnings [#1282]
- removed doxygen warnings [#1290]
- reimplemented fmix and fmax [#1311]
Bug Fixes
- allow nested indexes to compile with integrate_ode [#1317]
Documentation
- fixed typos in MCMC sampling