Make config_file_path optional on DeepLabCutInterface

[h-mayorquin]

Needed for the Cohen conversion where the files are (hopefully temporarily) missing. Plus, as described in #1030 we don't really extract any critical information from them.

[CodyCBakerPhD]

Can you add a small test case with the existing example data really quick?

[h-mayorquin]

Can you add a small test case with the existing example data really quick?

What kind of test are you envisioning? That it does not fail when you pass None instead of a file_path?

[CodyCBakerPhD]

A showcase of w/e new behavior you are adding in this PR insofar as it is different from the current test cases

[h-mayorquin]

The new behavior is that you don't require to pass a configuration file path. Would the test I mentioned cover it?

[CodyCBakerPhD]

Possibly - unless there is logic in DLC interface for 'searching' for a nearby config, in which case you'd have to making a tmpdir copy of example data w/o the config file

[h-mayorquin]

Possibly - unless there is logic in DLC interface for 'searching' for a nearby config, in which case you'd have to making a tmpdir copy of example data w/o the config file

There is no such a thing.

[h-mayorquin]

This was a good suggestion @CodyCBakerPhD. I found out another nested use of the configuration file deep inside the call stack of the utils.

[h-mayorquin]

Will need this:
#1047

[bendichter]

Suggested change

	expected_pose_estimation_series = ["ind1_leftear", "ind1_rightear", "ind1_snout", "ind1_tailbase"]
	expected_pose_estimation_series_are_in_nwb_file = [
	pose_estimation in pose_estimation_series_in_nwb for pose_estimation in expected_pose_estimation_series
	]
	assert all(expected_pose_estimation_series_are_in_nwb_file)
	assert all(
	pose_estimation in pose_estimation_series_in_nwb for pose_estimation in ["ind1_leftear", "ind1_rightear", "ind1_snout", "ind1_tailbase"]
	)

this would be clearer to me

[h-mayorquin]

This is a pet peeve of mine that I think makes debuggin easier so I would like to explain myself.

The reason that I like avoid definitions inside of other scopes is that it gets on the way of the standard debuggin workflow. With the current code If the test fails, I just set a debuggin point in line 390 (pdb.set_trace()) and then I can check immediatley what is the value of expected_pose_estimation_series_are_in_nwb_file and think about what failed. With the proposed change I would need to evaluate the expression again to inspect it.

Moreover, before python 3.12 list comprehension is debugged as a separate frame so if I use the automated tools of my IDE I can't be in a frame that has access to both values of interest: the values inside the list comprehension and the context from which is called.

All of that said I think the first definition in line 385 can be eliminated to make this more concise but I like having all the defintions before the assertions/checks so I can debug everything in the same frame.

[bendichter]

ok, I rarely use pdb.set_trace, so I mostly go by readability, but I can see how it could get in the way for you. Feel free to ignore

[codecov]

Codecov Report

Attention: Patch coverage is 75.00000% with 8 lines in your changes missing coverage. Please review.

Project coverage is 90.32%. Comparing base (a9b993a) to head (45c7b68).
Report is 63 commits behind head on main.

Files with missing lines	Patch %	Lines
...v/datainterfaces/behavior/deeplabcut/_dlc_utils.py	70.37%	8 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1031      +/-   ##
==========================================
- Coverage   90.36%   90.32%   -0.04%     
==========================================
  Files         129      129              
  Lines        7938     7988      +50     
==========================================
+ Hits         7173     7215      +42     
- Misses        765      773       +8     

Flag	Coverage Δ
unittests	90.32% <75.00%> (-0.04%)	⬇️

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

Files with missing lines	Coverage Δ
...ces/behavior/deeplabcut/deeplabcutdatainterface.py	93.18% <100.00%> (+0.68%)	⬆️
...v/datainterfaces/behavior/deeplabcut/_dlc_utils.py	55.15% <70.37%> (-3.24%)	⬇️

... and 1 file with indirect coverage changes