Closed lidakanari closed 8 years ago
That probably makes more sense. Whatever it does, it should be documented clearly. i.e. the behaviour here would be different for populations and single neurons. So we'd need to document what it does in either case.
Maybe it would be better to specify clearly what each of the "feature" functions should return, both for single neurons and for populations.
Yes, sorry.. I though it was common sense!
The feature of "number_of_sections" should return:
The same goes for the "total_length" or other total measurements.
Is this a little bit more clear?
@lidakanari That is fine, thanks. @eleftherioszisis I think you implemented the "feture getters", right? If so, did you have a reason to implement the current functionality, or is it OK to change it?
@eleftherioszisis Also, I am not sure if it should be a list or a np.array (please confirm so that we make all the features consistent )!
@juanchopanza Yes, indeed I implemented them. I didn't have any particular reasons for the total numbers due to not knowing what was needed. I believe that the per neuron numbers are much more useful for analysis, and in fact I needed per neuron total numbers as well for astrocytes, which I calculated manually. :P
All the features I implemented in the features module returned iterators that were converted into numpy arrays. Keep also in mind that if a single value is returned instead of an array of one value, then respective checks must be added in automated analyses of features.
OK, thanks for the information. I can start working on this next week.
BTW, shouldn't the output be
array([42, 42, 84, 84, 84])
I can't see a reason for the inner arrays.
Yes, you are right :) That's what I got from my quick (and dirty) implementation, so your proposal is the correct one.
fst
was taken care of in #368. Someone else can do neuron.features
if required.
The "number_of_sections" should return a list of the total number of sections per neuron. Currently it adds up all the sections of the population. Here is an example:
The number of neurons in the population is 5:
So the correct result would be: