replace opencv-python with pre-built CPU-only opencv-python-headless

[zacharyburnett]

Closes spacetelescope/stenv#103

This PR tests replacing opencv-python with opencv-python-headless, a pre-built CPU-only binary, to resolve the issue raised in spacetelescope/stenv#103

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

Patch and project coverage have no change.

Comparison is base (5fdad6f) 77.56% compared to head (d12b5f8) 77.56%.

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #7570   +/-   ##
=======================================
  Coverage   77.56%   77.56%           
=======================================
  Files         455      455           
  Lines       36340    36340           
=======================================
  Hits        28188    28188           
  Misses       8152     8152           

Flag	Coverage Δ		*Carryforward flag
nightly	77.62% <ø> (ø)		Carriedforward from 5fdad6f

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

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

[zacharyburnett]

regression test running at https://plwishmaster.stsci.edu:8081/blue/organizations/jenkins/RT%2FJWST-Developers-Pull-Requests/detail/JWST-Developers-Pull-Requests/696/pipeline/

[zacharyburnett]

upstream tests at spacetelescope/stcal#164

[hbushouse]

Looks like the regtest run was aborted. Will there be another?

[zacharyburnett]

Looks like the regtest run was aborted. Will there be another?

yes I ran one at https://plwishmaster.stsci.edu:8081/blue/organizations/jenkins/RT%2FJWST-Developers-Pull-Requests/detail/JWST-Developers-Pull-Requests/697/pipeline

It looks like there were quite a few failures:
https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/697/testReport/
with quite a few of them having the following:

../miniconda/envs/tmp_env0/lib/python3.9/site-packages/stdatamodels/jwst/transforms/converters/jwst_models.py:34: in to_yaml_tree_transform
    xll = [list(m) for m in model.xmodels]
E   TypeError: 'Polynomial1D' object is not iterable

[braingram]

Looks like the regtest run was aborted. Will there be another?

yes I ran one at https://plwishmaster.stsci.edu:8081/blue/organizations/jenkins/RT%2FJWST-Developers-Pull-Requests/detail/JWST-Developers-Pull-Requests/697/pipeline

It looks like there were quite a few failures: https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/697/testReport/ with quite a few of them having the following:

../miniconda/envs/tmp_env0/lib/python3.9/site-packages/stdatamodels/jwst/transforms/converters/jwst_models.py:34: in to_yaml_tree_transform
    xll = [list(m) for m in model.xmodels]
E   TypeError: 'Polynomial1D' object is not iterable

Those errors also appeared in CI runs for this PR for stdatamodels (where @nden mentioned they should be resolved by new reference files).

spacetelescope/stdatamodels#160 (comment)

and are likely related to: spacetelescope/stdatamodels#124

[hbushouse]

That's because you're letting stdatamodels float to the latest. Must pin to "stdatamodels==1.4.0".

[zacharyburnett]

That's because you're letting stdatamodels float to the latest. Must pin to "stdatamodels==1.4.0".

Ah right, I remember your announcement

[zacharyburnett]

regtest running at https://plwishmaster.stsci.edu:8081/blue/organizations/jenkins/RT%2FJWST-Developers-Pull-Requests/detail/JWST-Developers-Pull-Requests/702/pipeline with stdatamodels==1.4.0

[zacharyburnett]

regression test is all green now 🎉

[hbushouse]

regression test is all green now 🎉

I see that. One caveat though: the regtests only run on Linux. How do we know everything's OK with builds on a Mac?

[zacharyburnett]

Good point, I started a dev job at https://plwishmaster.stsci.edu:8081/blue/organizations/jenkins/RT%2FJWST-devdeps/detail/JWST-devdeps/656/pipeline/ as well with stdatamodels==1.4.0 to cover the MacOS tests

[zacharyburnett]

The MacOS tests are running into the ephemeral socket errors: https://plwishmaster.stsci.edu:8081/blue/organizations/jenkins/RT%2FJWST-devdeps/detail/JWST-devdeps/656/pipeline/194#step-196-log-29

However, it looks like there were no import errors (so far), so we'll see if the test failures are just socket errors and precision differences between platforms

[zacharyburnett]

aside from 24 failures in float differences, tests passed on MacOS: https://plwishmaster.stsci.edu:8081/blue/organizations/jenkins/RT%2FJWST-devdeps/detail/JWST-devdeps/656/pipeline/194#step-196-log-4418