Closed braingram closed 9 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Comparison is base (
6884f83
) 64.70% compared to head (f0adb6c
) 64.70%.:exclamation: Current head f0adb6c differs from pull request most recent head d0f1ca5. Consider uploading reports for the commit d0f1ca5 to get more accurate results
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
The 8 regression tests failures are unrelated and caused by some new CRDS ref files back at the time the tests were run.
This PR changes what happens during
DataModel.clone
to avoid an extra validation of the asdf tree. The validation duringclone
is (without this PR) only performed if thedeepcopy
argument isTrue
.clone
withdeepcopy=True
is only called duringcopy
where we can assume that the instanced being copied is valid. https://github.com/spacetelescope/stdatamodels/blob/ed1c5f6ee6d361b6a0b564af66b766b661cb704d/src/stdatamodels/model_base.py#L421-L429Regression tests: https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/1021/ show 8 failures (the same as on jwst main).
While working an issue with cubebuild I noticed that 16% of the time spent during
test_cube_build_single_output
is validating asdf contents (in part duringDataModel.clone
). With this PR that number drops to 12% and the runtime of the test is sped up by ~10%.Checklist
CHANGES.rst
(either inBug Fixes
orChanges to API
)