testing code cleanup

[jGaboardi]

This PR:

• resolves better raises/warns control + DRY out testing #448
• resolves "Generate and publish the report" not working - remove #450
• resolves codecov acting up a bit – update action #451
• resolves numpy DeprecationWarning within pyproj #452
• eliminated CI warnings in all by oldest and dev

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 77.8%. Comparing base (785b396) to head (9707567).

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #449   +/-   ##
=====================================
  Coverage   77.8%   77.8%           
=====================================
  Files         27      27           
  Lines       2638    2638           
=====================================
  Hits        2053    2053           
  Misses       585     585           

Files	Coverage Δ
spopt/locate/util.py	100.0% <100.0%> (ø)

[gegen07]

Excellent work @jGaboardi! The PR smoothens the code a bit and adds more reusability.

• I noted some warnings were not moved to conftest.py. Is there a specific reason for deciding which test should use the conftest.py warning?
• About the numpy deprecationWarning, I think we should filter it, as it seems, it will disappear in future releases of pyproj.

[gegen07]

There is a similar infeasibility warning in line 86. Should we move all expected warnings to conftest.py?

[jGaboardi]

There is a similar infeasibility warning in line 86. Should we move all expected warnings to conftest.py?

It's similar, but not the same message. I think keeping where it is now would be best. However, this brings a larger discussion of it we should review and generalize the error messages. If we think that should be the case let's open a fresh issue for that.

[gegen07]

I agree with opening a fresh issue. I think we can keep these warnings without generalizing them first. To avoid over-engineering, we can do this transition step-by-step. Firstly, move the warnings in conftest. Then, if we see conftest has a lot of repeated code, we should generalize it.

[jGaboardi]

I noted some warnings were not moved to conftest.py. Is there a specific reason for deciding which test should use the conftest.py warning?

• Yeah, my general rule was that if identical warns/raises happen across testing file then put it in conftest.py. There is one raises that happens twice within test_knearest so I just kept that one in the file itself.
• After looking through again, I have missed a few cases. Good catch on that! I'll push up some changes.

[jGaboardi]

About the numpy deprecationWarning, I think we should filter it, as it seems, it will disappear in future releases of pyproj.

OK, I will see about adding that to this PR.

[jGaboardi]

@gegen07 Seems I missed quite few 😅.

Also, I am having trouble pinning down where the DeprecationWarning is being thrown. It happens when I run the tests through pytest like usual, but then nothing is thrown after I broke out the code line by line 🤔

[jGaboardi]

OK! I think I've found to offending line. Will push up the fix soon.

[jGaboardi]

@gegen07 OK. I think this is ready for a proper review.