Fix a bug in how a image group name is determined

[mcara]

It has been reported when an user used custom file names with the same common file name in several groups of images (i.e., with different group_id in the asn files). The algorithm for computing group names (_common_name()) may fail to produce a unique name for some groups resulting in these groups not being aligned.

This is not an issue for pipeline runs on original JWST file names but may affect grouping when user renames input models.

This PR is a further improvement on #8012

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 ✅

Project coverage is 56.30%. Comparing base (2fb073e) to head (71d91f1).
Report is 34 commits behind head on master.

❗ Current head 71d91f1 differs from pull request most recent head b4112c9. Consider uploading reports for the commit b4112c9 to get more accurate results

Additional details and impacted files

@@             Coverage Diff             @@
##           master    #8426       +/-   ##
===========================================
- Coverage   75.31%   56.30%   -19.02%     
===========================================
  Files         474      390       -84     
  Lines       38965    38886       -79     
===========================================
- Hits        29345    21893     -7452     
- Misses       9620    16993     +7373     

Flag	Coverage Δ
nightly	?

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

[braingram]

Is there an advantage to using a name different from the group_id? I see 2 uses of _common_name and in both cases the same _common_name is used for all images in a group (these images will also share the same group_id).

[mcara]

Is there an advantage to using a name different from the group_id?

I don't know for sure. One consideration was to have a relatively short name so that it is easily readable in the log file (a bad alternative would be to concatenate all file names of models in a group.) The main issue is that group_id is not a property of the image models and not even of the ModelContainer: all ModelContainer does is to compute group_id and use this information to return nested lists of models. Given that this code receives a nested list, _common_name() is the answer to that problem.

Ideally, I think group_id should be a property of DataModel whether group_id is stored in meta (I am not sure it should) or whether it is computed on demand. Until then, I don't see an alternative to _common_name().

[braingram]

Thanks! ModelContainer currently dynamically sets meta.group_id for the models returned from models_grouped (see the below code):

jwst/jwst/datamodels/container.py

Lines 461 to 480 in 9429dc6

	if (hasattr(model.meta, 'group_id') and
	model.meta.group_id not in [None, '']):
	group_id = model.meta.group_id
	else:
	for param in unique_exposure_parameters:
	params.append(getattr(model.meta.observation, param))
	try:
	group_id = (
	'jw' + '_'.join(
	[
	''.join(params[:3]),
	''.join(params[3:6]),
	params[6],
	]
	)
	)
	model.meta.group_id = group_id
	except TypeError:
	model.meta.group_id = 'exposure{0:04d}'.format(i + 1)

[mairanteodoro]

Looks good to me. Thanks, @mcara!

[braingram]

LGTM! Have regtests been run with these changes?

[mcara]

LGTM! Have regtests been run with these changes?

running now

[mcara]

https://plwishmaster.stsci.edu:8081/job/RT/job/JWST-Developers-Pull-Requests/1400/

[mcara]

Thanks @braingram & @mairanteodoro

[braingram]

Happy to help!
Looks like the regtests all passed 🎉