[Merge after #200]: Subframe segmentation

[CodyCBakerPhD]

Relies on trace orientation fix/standardization in #200.

Introduces subframe selection capabilities to SegmentationExtractors as well as tests for this.

Necessary to introduce frame stubbing into the SegmentationDataInterface on NeuroConv.

[h-mayorquin]

Some comments here and there.

Can you add a test where the segmentation extractor does not have one of the traces? there is an argument in generate_dummy_segmentation_extractor to control that. The sliced segmentation extractor traces_dict should not have more keys than the father.

[CodyCBakerPhD]

Can you add a test where the segmentation extractor does not have one of the traces? there is an argument in generate_dummy_segmentation_extractor to control that. The sliced segmentation extractor traces_dict should not have more keys than the father.

Done

[codecov]

Codecov Report

Merging #201 (5cff4ac) into master (6f241df) will increase coverage by 0.55%.
The diff coverage is 96.36%.

@@            Coverage Diff             @@
##           master     #201      +/-   ##
==========================================
+ Coverage   72.73%   73.28%   +0.55%     
==========================================
  Files          33       33              
  Lines        2226     2276      +50     
==========================================
+ Hits         1619     1668      +49     
- Misses        607      608       +1     

Flag	Coverage Δ
unittests	73.28% <96.36%> (+0.55%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/roiextractors/testing.py	99.31% <ø> (ø)
src/roiextractors/segmentationextractor.py	91.03% <96.29%> (+3.66%)	⬆️
...tors/extractors/numpyextractors/numpyextractors.py	60.94% <100.00%> (ø)

[h-mayorquin]

Looks good to me, one minor suggestion if you want to add.

[h-mayorquin]

I did not think on this pattern, it is neat.

[CodyCBakerPhD]

Multiplies the total number of tests run when doing it, but going for maximum safety here.

[h-mayorquin]

I see that you have some tests for rejected and accepted list. Up to your but the generate_dummy_segmentation_extractor has a rejected list argument to generate that:

roiextractors/src/roiextractors/testing.py

Lines 54 to 65 in b71f17d

	def generate_dummy_segmentation_extractor(
	num_rois: int = 10,
	num_frames: int = 30,
	num_rows: int = 25,
	num_columns: int = 25,
	sampling_frequency: float = 30.0,
	has_summary_images: bool = True,
	has_raw_signal: bool = True,
	has_dff_signal: bool = True,
	has_deconvolved_signal: bool = True,
	has_neuropil_signal: bool = True,
	rejected_list: Optional[list] = None,

So you can test the non-default case when rejected_list != [ ]

[CodyCBakerPhD]

Yeah, I could. But I actually have yet to see any format in practice that bothers to save ROI info that has not been accepted. (so far, every single ROI written as series data or image mask is 'accepted' so 'rejected' is always empty). Maybe something to think about and revisit in the future