Fix EventedModel signatures with PySide2 imported

[tlambert03]

Description

fixes #2264. This fixes a strange and unfortunate clash resulting from decisions made by PySide2 and pydantic (neither of which I feel we'd have much luck changing upstream). I put a more thorough explanation in the code docstring itself, but here's the gist:

1. when PySide2 fixed this bug, they made the decision to make (all?) python __signature__ objects dynamically writable (as they should be), but only once (?!).
2. Pydantic, in attempting to fix signature conflicts in cases where a class defines a __call__ method, uses a descriptor object called ClassAttribute to wrap their custom inspect.Signature object during class creation. such that cls.__signature__ = ClassAttribute('__signature__', actual_signature)
3. when accessing obj.__signature__ PySide2 seems to just return that ClassAttribute descriptor rather than calling its __get__ method... so inspect.signature fails.

This PR temporarily monkeypatches pydantic's ClassAttribute object to just pass through the signature object directly when constructing our classes... then undoes the patch.

Type of change

• Bug-fix (non-breaking change which fixes an issue)

How has this been tested?

• added a test to make sure that signatures are working
• all tests pass with my change

Final checklist:

• My PR is the minimum possible work for the desired functionality
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works

[codecov]

Codecov Report

Merging #2265 (69a47a2) into master (dcfdb0c) will decrease coverage by 0.00%.
The diff coverage is 62.96%.

@@            Coverage Diff             @@
##           master    #2265      +/-   ##
==========================================
- Coverage   67.10%   67.09%   -0.01%     
==========================================
  Files         400      400              
  Lines       33652    33672      +20     
==========================================
+ Hits        22581    22593      +12     
- Misses      11071    11079       +8     

Impacted Files	Coverage Δ
napari/utils/events/evented_model.py	50.00% <44.44%> (+0.79%)	⬆️
napari/utils/events/_tests/test_evented_model.py	100.00% <100.00%> (ø)
napari/utils/events/event.py	49.56% <0.00%> (-0.88%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update dcfdb0c...69a47a2. Read the comment docs.

[sofroniewn]

LGTM! Great debugging and crazy situation!!

[sofroniewn]

Can you link to the issue where you raised #2264, maybe even this PR #2265 too, I can just imagine someone who reads this wanting to find there way back here too

[basnijholt]

@tlambert03 did you ever raise this issue with PySide2?

[tlambert03]

Hi @basnijholt, no I didn't create a new issue. I did find this issue: https://bugreports.qt.io/browse/PYSIDE-1004 ... so they're aware of it to some degree. But I haven't checked to see how more recent versions of PySide2/6 are handling this