Fix Version/s: None
Sprint:AP S21-4 (March)
Allow the --dataset argument to ap_verify.py and related scripts to take the repository name. Raise a deprecation warning any time an old-style argument is processed.
DM-25629 Do release 22.0.0 of science pipelines
DM-29042 Remove support for old-style names in ap_verify
- has to be done before
DM-29040 Use repository name as ci_dataset tag in ap_verify SQuaSH uploads
- is triggered by
RFC-758 Deprecate ap_verify datasets' user-facing names
I'm partway through; so far the changes to pipeline_driver.py and ap_verify.yaml look fine, and I'm working on add_gen3_repo.py now. I plan to finish all today.
Nearly done, but I have one question on add_gen3_repo.py. I see you've hard-coded two names here:
Are these always guaranteed to be correct (maybe it's fine because it's ap_verify) or should they come from some config value?
I've pushed fixes to ap_verify_dataset_template and ap_verify. Can you take another look, and mark this ticket as reviewed if you're happy?
I believe that the special-casing for ap_verify_testdata can be removed on this ticket – since all references to the dataset should be internal anyway, I can update them to use the repository name immediately.
Edit: some references to the special name for ap_verify_testdata remain in test_dataset.py, in order to test the deprecation mechanism.