Issue with running the example spike inference code

[vjlbym]

Hey guys,

I was trying to get the spike inference example working and it was giving me a bunch of errors. The version as is gave me a TypeError: object of type 'numpy.float64' has no len(). This seemed to be because the code was expecting a 2d array here. I tried putting the argument inside a list and that error went away but then a bunch of other errors popped up. I didn't want to post all the errors here since it is easily reproducible. I tried to work through them all but the final output I got made no sense as I just got no spikes detected. Could you please help me with it?

Thanks a lot!
Vijay

[pkaifosh]

Do these changes work for you?

 # Jointly estimate gamma on traces concatenated together
 [joint_gamma_est, joint_sigma_est] = spikes.estimate_parameters(
-    FLUORS.reshape(FLUORS.size), mode="correct")
+    FLUORS, mode="correct")

 for x in range(NTRACE):

     # Estimate noise and decay parameters
     [gamma_est, sigma_est] = spikes.estimate_parameters(
-        FLUORS[x, ], mode="correct", gamma=joint_gamma_est)
+        [FLUORS[x]], mode="correct", gamma=joint_gamma_est)
     print("tau = {tau},  sigma = {sigma}".format(
         tau=DELTAT / (1 - gamma_est), sigma=sigma_est))

[vjlbym]

Thanks @pkaifosh

That did solve the issue I raised. However, I now have a different problem, associated with MOSEK. I had installed it after I had the previous issues. So I don't know if this is a new issue; it likely is.

Line 67 now gives the following error that looks like some missing setting for MOSEK based on their error codes. Do you know how I can fix it? Thanks again!

---------------------------------------------------------------------------
Error                                     Traceback (most recent call last)
<ipython-input-1-6ea570844390> in <module>()
     69     # Run spike inference
     70     INFERENCE[:, x], FITS[:, x], params = spikes.spike_inference(
---> 71         FLUORS[x, ], sigma=sigma_est, gamma=joint_gamma_est, verbose=True)
     72 
     73 #########

/home/stuberlab/anaconda/lib/python2.7/site-packages/sima/spikes.pyc in spike_inference(fluor, sigma, gamma, mode, ar_order, psd_opts, verbose)
    263     start_time = time.time()
    264     try:
--> 265         prob.solve(solver='mosek', verbose=False)
    266     except ImportError:
    267         warn('MOSEK is not installed. Spike inference may be VERY slow!')

/home/stuberlab/anaconda/lib/python2.7/site-packages/picos/problem.pyc in solve(self, **options)
   4566                         or self.options['solver'] == 'mosek6'):
   4567 
-> 4568                     primals, duals, obj, sol = self._mosek_solve()
   4569 
   4570                 # for scip

/home/stuberlab/anaconda/lib/python2.7/site-packages/picos/problem.pyc in _mosek_solve(self)
   5938 
   5939         if version7:
-> 5940             solsta = task.getsolsta(soltype)
   5941         else:
   5942             [prosta, solsta] = task.getsolutionstatus(soltype)

/home/stuberlab/anaconda/lib/python2.7/site-packages/mosek/__init__.pyc in accept(*args)
    205                 raise TypeError('Expected %d argument(s) (%d given)' % (len(argtlst), len(args)))
    206             try:
--> 207                 return fun(*[ t(a) for (t,a) in zip(argtlst,args) ])
    208             except TypeAcceptError, e:
    209                 raise TypeError(e)

/home/stuberlab/anaconda/lib/python2.7/site-packages/mosek/__init__.pyc in getsolsta(self, whichsol_)
   5415     if res != 0:
   5416       _,msg = self.__getlasterror(res)
-> 5417       raise Error(rescode(res),msg)
   5418     _solsta_return_value = solsta(solsta_.value)
   5419     return (_solsta_return_value)

Error: (1265) The solution with code 0 is not defined.