[python] Add GHA workflow for minimal CI jobs

[atolopko-czi]

Issue and/or context:

#723

Changes:

• add python-ci-minimal.yml GHA workflow for minimal CI jobs, invoked by pushes to feature branches
• rename python-ci.yml to python-ci-full.yml, to differentiate from the "minimal" workflow; invoked by pushes to main and for releases
• add python-ci-single.yml, which is a reusable workflow called by the "minimal" and "full" workflows.
• factored out linting steps into a separate job from the build/tests steps to avoid having conditional logic on those steps

Notes for Reviewer:

Successful workflow runs:

• "minimal": https://github.com/single-cell-data/TileDB-SOMA/actions/runs/3961177942
• "full" https://github.com/single-cell-data/TileDB-SOMA/actions/runs/3961201306
  As these requires some temporary hacks to trigger, our final test of the "full" workflow can only be run upon merging this PR to main. 🤞

[codecov-commenter]

Codecov Report

Base: 78.76% // Head: 79.35% // Increases project coverage by +0.58% 🎉

Coverage data is based on head (5e2689d) compared to base (762fd68).
Patch has no changes to coverable lines.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #750      +/-   ##
==========================================
+ Coverage   78.76%   79.35%   +0.58%     
==========================================
  Files          30       30              
  Lines        1978     1981       +3     
==========================================
+ Hits         1558     1572      +14     
+ Misses        420      409      -11     

Flag	Coverage Δ
python	79.35% <ø> (+0.58%)	⬆️

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

Impacted Files	Coverage Δ
apis/python/src/tiledbsoma/collection.py	86.66% <0.00%> (-1.67%)	⬇️
apis/python/src/tiledbsoma/sparse_nd_array.py	94.06% <0.00%> (-0.85%)	⬇️
apis/python/src/tiledbsoma/dense_nd_array.py	93.10% <0.00%> (+0.08%)	⬆️
apis/python/src/tiledbsoma/util_arrow.py	92.85% <0.00%> (+0.12%)	⬆️
apis/python/src/tiledbsoma/query_condition.py	61.05% <0.00%> (+0.18%)	⬆️
apis/python/src/tiledbsoma/io.py	84.50% <0.00%> (+0.28%)	⬆️
apis/python/src/tiledbsoma/factory.py	89.79% <0.00%> (+2.04%)	⬆️
apis/python/src/tiledbsoma/eta.py	84.21% <0.00%> (+2.63%)	⬆️
apis/python/src/tiledbsoma/util_ann.py	15.56% <0.00%> (+4.79%)	⬆️
apis/python/src/tiledbsoma/logging.py	80.00% <0.00%> (+20.00%)	⬆️

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.

[atolopko-czi]

This now differs from r-ci.yml, which is not ideal. And the GHA console shows push- and pull_request-triggered workflow runs somewhat differently. I'm going to test once more with pull_request...

[atolopko-czi]

@johnkerl @thetorpedodog I'd say this is ready, if we want to move forward with it. I merged in the caching improvements that Paul made earlier today.

I never quite figured out why only push works to trigger Python workflows, while pull_request is working to trigger R workflows. Interestingly, the R workflow triggers even if just a apis/python/** file changes (it shouldn't), which makes me believe it's not the pull_request trigger that is working, but rather the push trigger that also exists in the r-ci.yml file. It must be that a transient merge into main is being done to test the PR against main, and that is what's triggering the R workflow. It's a bit mind-boggling and GHA docs are not clear on the matter at all. But they hint at such things as I've described...

Not going to spend more time on this, so we can take it or leave it as is. 😄

(I believe the R build failure is known and may be resolved with another merge from main.)

[johnkerl]

Yes @atolopko-czi re known-broken R CI pending #766 :)

[johnkerl]

@atolopko-czi thank you thank you thank you 🙏

I know this was a long slog, and a puzzler, with lots of revs trying to untangle various complexities

But I really do apprecate it -- the more people are actively contributing to SOMA right now, the more we're using CI -- and your work here will help us _all_move faster. :D

[johnkerl]

See also #693