Support datetime in conversion options

[h-mayorquin]

Hi, @alessandratrapani .

I need to add a test but this is the reason you are seeing this error. As you might know we use json schema to do validation of the conversion options, metadata and interface arguments (source data). The json schema validation requires arguments to be string and we use encoders to accomplish this. Check out the code. I still need to add a test for this but I wanted to share this with you as quick as possible.

A temporal solution in your code is to transform the datetime object to string (see the diff here for how) and then transform it back to datetime in add_to_nwb. But we can merge this quickly as well once I add a test.

[alessandratrapani]

THank you for taking care of this. PR looks good to me

[codecov]

Codecov Report

Attention: Patch coverage is 92.30769% with 2 lines in your changes missing coverage. Please review.

Project coverage is 90.60%. Comparing base (6960872) to head (862f42d).
Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
src/neuroconv/tools/testing/mock_interfaces.py	90.90%	1 Missing ⚠️
src/neuroconv/utils/json_schema.py	80.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1139      +/-   ##
==========================================
- Coverage   90.60%   90.60%   -0.01%     
==========================================
  Files         129      129              
  Lines        8174     8193      +19     
==========================================
+ Hits         7406     7423      +17     
- Misses        768      770       +2     

Flag	Coverage Δ
unittests	90.60% <92.30%> (-0.01%)	⬇️

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

Files with missing lines	Coverage Δ
src/neuroconv/basedatainterface.py	95.34% <100.00%> (ø)
src/neuroconv/nwbconverter.py	94.96% <100.00%> (+0.11%)	⬆️
src/neuroconv/tools/testing/mock_interfaces.py	99.14% <90.90%> (-0.86%)	⬇️
src/neuroconv/utils/json_schema.py	94.54% <80.00%> (-0.46%)	⬇️

[bendichter]

Why not add this to _NWBMetaDataEncoder instead of making a new class? I'm kind of surprised this isn't the default behavior for Path objs. What is the default behavior?

[h-mayorquin]

The default behavior is to throw an error.

Yeah, maybe we can use the same encoder for everything. Note that even before this PR we already had a _NWBSourceDataEncoder:

neuroconv/src/neuroconv/utils/json_schema.py

Lines 46 to 60 in dcd248b

	class _NWBSourceDataEncoder(_NWBMetaDataEncoder):
	"""
	Custom JSON encoder for data interface source data (i.e. kwargs).
	This encoder extends the default JSONEncoder class and provides custom serialization
	for certain data types commonly used in interface source data.
	"""
	def default(self, obj):
	# Over-write behaviors for Paths
	if isinstance(obj, Path):
	return str(obj)
	return super().default(obj)

I think that the original idea was that metadata, source data and the conversion options might require different behavior so it was created that way. Check out the inheritance pattern here that kind of implements this simplification:

https://github.com/catalystneuro/neuroconv/pull/1142/files#r1856981408