reorganize tests, make devdeps testing optional, unpin latest Python testing, and prepare for possibility of data caching

[zacharyburnett]

Resolves SCSB-79

This PR

• reorganizes CI tests and moves devdeps tests to a new optional workflow that can be started with the label run devdeps tests
• similarly, the label run scheduled tests can be used as well
• pins the version of the OpenAstronomy reusable test workflow to the v1 tag to minimize surprises with API changes
• abstracts retrieval of the CRDS context (spacetelescope/crds/.github/workflows/contexts.yml@master) to improve maintainability of this workflow going forward
• caches CRDS data to /tmp/data/ like in other repos' data.yml workflows (can potentially be packaged into a combined cache)
• run tests on latest Python version in devdeps (py3)

Checklist for maintainers

• added entry in CHANGES.rst within the relevant release section
• updated or added relevant tests
• updated relevant documentation
• added relevant milestone
• added relevant label(s)
• ran regression tests, post a link to the Jenkins job below.
  How to run regression tests on a PR
• Make sure the JIRA ticket is resolved properly

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (f77b5d6) 75.42% compared to head (4f7f614) 75.42%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7910   +/-   ##
=======================================
  Coverage   75.42%   75.42%           
=======================================
  Files         464      464           
  Lines       37938    37938           
=======================================
  Hits        28615    28615           
  Misses       9323     9323           

Flag	Coverage Δ		*Carryforward flag
nightly	77.37% <ø> (ø)		Carriedforward from 4be1691

*This pull request uses carry forward flags. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[hbushouse]

Why all the CI failures? Looks like something to do with our pinning of the python version <3.12.

[zacharyburnett]

in the regular non-devdeps tests it looks like an issue with timecoeff (I'm not sure what that comes from)
https://github.com/spacetelescope/jwst/actions/runs/7062856370/job/19227602452#step:10:410

=========================== short test summary info ============================
FAILED jwst/photom/tests/test_photom.py::test_miri_image - AttributeError: No attribute 'timecoeff'
= 1 failed, 2705 passed, 532 skipped, 4 xfailed, 110 warnings in 952.53s (0:15:52) =

In the py311-devdeps tests (with 3rd party dev dependencies) the error is with using Numpy 1.x at build time and Numpy 2.x at run time; I'll look into how to resolve that I think we'll just have to wait for a Numpy 2.0 release to resolve this
https://github.com/spacetelescope/jwst/actions/runs/7062856371/job/19227604346#step:10:275

A module that was compiled using NumPy 1.x cannot be run in
NumPy 2.0.0.dev0 as it may crash. To support both 1.x and 2.x
versions of NumPy, modules must be compiled against NumPy 2.0.

Then for the latest Python ones (py3) the error you mentioned:
https://github.com/spacetelescope/jwst/actions/runs/7062856371/job/19227604975#step:10:28

ERROR: Package 'jwst' requires a different Python: 3.12.0 not in '<3.12,>=3.9'

[braingram]

Re the numpy build vs runtime:
If it's useful, I made a brief attempt to fix this by installing the pyerfa nightly but the run still failed due to drizzle (which might also need to be built against numpy 2.0): #8094

It looks like stdatamodels has the same devdeps issues :-/

[hbushouse]

Ah yes, the "timecoeff" errors are due to an update in the jwst master that is not yet available in the latest stdatamodels release, which is used by the CI tests. It is available, however, in stdatamodels/master. I guess we need an stdatamodels release ASAP.

[zacharyburnett]

If the regular tests pass here, we can likely merge this and then deal with the Numpy 2.0 devdeps issues in another PR

[hbushouse]

So the remaining CI errors all seem to be due to numpy.core.multiarray import errors. Is that what you expect and hence is this ready to merge as-is?

[zacharyburnett]

So the remaining CI errors all seem to be due to numpy.core.multiarray import errors. Is that what you expect and hence is this ready to merge as-is?

yes, I think that's expected. This is ready to merge, and then we can fix the devdeps