Fix slm requirement in Python API

[CodyCBakerPhD]

The spec specifies the SLM of a stimulus site is not required: https://github.com/catalystneuro/ndx-patterned-ogen/blob/main/spec/ndx-patterned-ogen.extensions.yaml#L124

But I was unable to use the API without one specified

This should fix the issue

[CodyCBakerPhD]

Also added spacing between type definitions here in the spec to make them easier to visually identify

[CodyCBakerPhD]

@rly Couple of things here

i) I get a build warning on file write

C:\Users\theac\anaconda3\envs\leifer_lab_to_nwb_created_5_13_2024\Lib\site-packages\hdmf\build\objectmapper.py:671: MissingRequiredBuildWarning: PatternedOptogeneticStimulusSite 'PatternedOptogeneticStimulusSite' is missing required value for attribute 'spatial_light_modulator'.
  warnings.warn(msg, MissingRequiredBuildWarning)

But file roundtrips just fine; also confused why it thinks it's required in the object mapper

Not too worried since the file reads fine, just curious / letting you know

ii) the testing suite for ndx-extension template seems to be failing due to numpy 2.0 issues; advice on how to resolve generally (besides hard pinning upper bound on all extensions?)

[rly]

@rly Couple of things here

i) I get a build warning on file write

C:\Users\theac\anaconda3\envs\leifer_lab_to_nwb_created_5_13_2024\Lib\site-packages\hdmf\build\objectmapper.py:671: MissingRequiredBuildWarning: PatternedOptogeneticStimulusSite 'PatternedOptogeneticStimulusSite' is missing required value for attribute 'spatial_light_modulator'.
  warnings.warn(msg, MissingRequiredBuildWarning)

required: false is not valid on links. I'll raise an issue to catch that. It's a common mistake. Use quantity: "?" instead. Strange that it reads fine; the default is quantity: 1 meaning it's required.

I'll look into the template.

[CodyCBakerPhD]

Though I guess, oddly, the PR on catalystneuro/ndx-microscopy#14 doesn't have that issue with the tests so maybe it's just this extension

[CodyCBakerPhD]

Yeah OK, pinning here fixed it but not sure about other extensions. Just wanted you to be aware

[rly]

Yeah OK, pinning here fixed it but not sure about other extensions. Just wanted you to be aware

Thanks. It turns out that the "upgraded" workflow was not actually upgrading the packages. That one should have passed. The others should fail because older versions of pynwb and hdmf, which are pinned for testing, did not set an upper limit on numpy, so numpy 2 ends up being installed. I will update the pinned versions in the template.

[CodyCBakerPhD]

How does this look @rly ?