Add mypy and black to CI

[WardBrian]

As discussed in #16, this will run mypy and black against PRs. At the moment it will fail because main doesn't pass these checks, but after main does I will rebase and merge this

[codecov-commenter]

Codecov Report

Merging #26 (90bcde3) into main (3157c8f) will decrease coverage by 0.09%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main      #26      +/-   ##
==========================================
- Coverage   98.10%   98.01%   -0.09%     
==========================================
  Files          11       11              
  Lines         316      302      -14     
==========================================
- Hits          310      296      -14     
  Misses          6        6              

Impacted Files	Coverage Δ
bayes_kit/__init__.py	100.00% <100.00%> (ø)
bayes_kit/autocorr.py	100.00% <100.00%> (ø)
bayes_kit/ensemble.py	100.00% <100.00%> (ø)
bayes_kit/ess.py	100.00% <100.00%> (ø)
bayes_kit/hmc.py	94.59% <100.00%> (-0.28%)	⬇️
bayes_kit/iat.py	100.00% <100.00%> (ø)
bayes_kit/mala.py	93.54% <100.00%> (-0.40%)	⬇️
bayes_kit/metropolis.py	100.00% <100.00%> (ø)
bayes_kit/rhat.py	100.00% <100.00%> (ø)
bayes_kit/smc.py	95.83% <100.00%> (-0.09%)	⬇️
... and 1 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[jsoules]

Couple minor things, mostly in tests.

mypy is complaining about three lines when I run it on command line; am I missing a config file, or is there a version inconsistency, or something? Those lines were changed to remove the ignore-type annotations, so I've noted in the review.

Also: I am getting some pylance typing warnings that aren't picked up by mypy. Dunno if we care at all, or want to address them in this PR, but just in case:

• test/models/binomial.py:
  • l 48 warns that the call to scipy.stats.beta.logpdf returns an NDArray not a float
• test/test_rhat.py:
  • l 24 warns that subtraction isn't supported between psij_bar and psi_bar since the former is list[floating[Any]] and the latter is floating[Any]. The test still passes, so it's either a spurious error in pylance or something is doing some implicit munging
• tests/test_metropolis.py:
  • l 202, warning about inability to locate numpy.typing within numpy, which may just be PyLance being weird; however this is fixed by updating the import from numpy.typing to include ArrayLike as well as NDArray and then referring to them directly without the package qualifiers
  • l 150, complains that sst.skewnorm.rvs call might be returning an int (maybe just needs a cast)
  • l 212 same issue as with l. 150, maybe needs a cast

[WardBrian]

l 48 warns that the call to scipy.stats.beta.logpdf returns an NDArray not a float

Pylance is lying to you

>>> isinstance(scipy.stats.beta.logpdf(.1, 3, 2), float)
True

(it does return an array if the first argument is an array, but it isn't in this model)

I think I resolved the other pylance warnings, though I'm not sure if they were also spurious

[jsoules]

This all looks good to me--nice work!