Make unit validators compatible with u.Quantity[u.T]-style annotations

[StanczakDominik]

Possibly closes #1413 and also closes #1414, maybe? Let's see if this blows up.

Essentially what happens when you do, say, def f(B: u.Quantity[u.T]) is that the annotation becomes a typing._AnnotatedAlias. Apparently the canonical way of reading type hints for a function is typing.get_type_hints, but since we didn't go that route for the existing code, I didn't want to start rewriting the whole thing just yet. Instead, the annotation has a __metadata__ field, an iterable of things within the [] parenthesis. So I just picked the first item in that for now.

I've also added this kind of annotation in frequencies for now, to see whether they'd explode.

• I have added a changelog entry for this pull request.

• If adding new functionality, I have added tests and
  docstrings.

• I have fixed any newly failing tests.

[StanczakDominik]

hey @rocco8773, I wanted to ping you for short feedback on what you think about this approach :)

[StanczakDominik]

I wonder whether we shouldn't deprecate the old, u.m style syntax.

[codecov]

Codecov Report

Base: 98.31% // Head: 97.19% // Decreases project coverage by -1.11% ⚠️

Coverage data is based on head (71371ff) compared to base (5e542ce).
Patch coverage: 75.00% of modified lines in pull request are covered.

❗ Current head 71371ff differs from pull request most recent head d65ae60. Consider uploading reports for the commit d65ae60 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1661      +/-   ##
==========================================
- Coverage   98.31%   97.19%   -1.12%     
==========================================
  Files          95       84      -11     
  Lines        8404     8020     -384     
==========================================
- Hits         8262     7795     -467     
- Misses        142      225      +83     

Impacted Files	Coverage Δ
plasmapy/utils/decorators/checks.py	99.12% <25.00%> (-0.58%)	⬇️
plasmapy/formulary/frequencies.py	100.00% <100.00%> (ø)
plasmapy/formulary/ionization.py	100.00% <100.00%> (ø)
plasmapy/formulary/lengths.py	100.00% <100.00%> (ø)
plasmapy/__init__.py	33.33% <0.00%> (-66.67%)	⬇️
plasmapy/formulary/quantum.py	68.67% <0.00%> (-31.33%)	⬇️
plasmapy/simulation/abstractions.py	76.04% <0.00%> (-23.96%)	⬇️
plasmapy/plasma/plasma_base.py	75.60% <0.00%> (-5.16%)	⬇️
plasmapy/analysis/fit_functions.py	98.55% <0.00%> (-1.45%)	⬇️
plasmapy/plasma/sources/openpmd_hdf5.py	81.52% <0.00%> (-0.90%)	⬇️
... and 58 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[StanczakDominik]

Oh yeah, here's the crux of the issue. typing._AnnotatedAlias, which is somehow related to typing.Annotated as I understand it, requires python 3.9. Without it, we can't even begin to annotate the functions themselves.

[StanczakDominik]

To answer potential questions about __metadata__, here's Guido:

It looks like __metadata__ was meant to be a public attribute, but somehow overseen when the PEP and docs were written. :-(

[StanczakDominik]

• no need to bump to 3.9 since we can do https://github.com/astropy/astropy/blob/58d45f2d4bc2147ea70c51fbf6abcf2f2cf7134a/astropy/units/_typing.py#L1
• try Quantity["length"] with mods to check_quantity

[github-actions]

Thank you for contributing to PlasmaPy! The project's future depends deeply on contributors like you, so we deeply appreciate it! 🌱 The following checklist will be used by the code reviewer to help guide the code review process.

• Overall
  • Does the PR do what it intends to do?
  • Except for very minor changes, is a changelog entry included and consistent with the changelog guide?
  • Are the continuous integration checks passing? (Most linter problems can be automagically fixed by commenting on this PR with pre-commit.ci autofix.)
• Code
  • Is new/updated code understandable and consistent with the coding guide?
  • Are there ways to greatly simplify the implementation?
  • Are there any large functions that should be split up into shorter functions?
• Tests
  • Are tests added/updated as required, and consistent with the testing guide?
  • Are the tests understandable?
  • Do the tests cover all important cases?
• Docs
  • Are docs added/updated as required, and consistent with the doc guide?
  • Are the docs understandable?
  • Do the docs show up correctly in the preview, including Jupyter notebooks?

[StanczakDominik]

Looks like this won't work without a proper rewrite of check_units to use the functionality from typing :(