work in progress

[iancharest]

Porting the new single trial beta estimates of glmdenoise.

Things to do:

• [x] robustrange.py
• [x] picksubset.py
• [x] findtailthreshold.py
• [x] finish the first draft of the port.
• [x] get one subject from irsa to further test.
• [ ] think about splitting into distinct modules for the 4 different GLM flavours.
• [ ] further build tests for the new cases
• [ ] makeimagestack add range normalisation

[iancharest]

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
~/Documents/GitHub/pyGLMdenoise/example_singletrial.py in <module>
     81 gst = glm_estimatesingletrial(opt)
     82 
---> 83 gst.fit(
     84     design,
     85     data,

~/Documents/GitHub/pyGLMdenoise/glmdenoise/single_trial.py in fit(self, design, data, stimdur, tr, outputdir)
    691 
    692                 for p in range(params['hrflibrary'].shape[1]):
--> 693                     glm = glm_estimatemodel(optC)
    694                     results0 = glm.fit(
    695                         designSINGLE,

~/Documents/GitHub/pyGLMdenoise/glmdenoise/utils/glm_estimatemodel.py in __init__(self, params)
    114         params = params or dict()
    115         for key, _ in default_params.items():
--> 116             params[key] = params.get(key) or default_params[key]
    117 
    118         self.params = params

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

[iancharest]

something iffy about the way i use hrflibrary in this new class like build.

[iancharest]

I think this might be ready for review now.