JP-1793: Standardize wfs_combine Step #6376
Conversation
Codecov Report
@@ Coverage Diff @@
## master #6376 +/- ##
==========================================
+ Coverage 77.60% 77.65% +0.04%
==========================================
Files 408 408
Lines 34881 34941 +60
==========================================
+ Hits 27068 27132 +64
+ Misses 7813 7809 -4
*This pull request uses carry forward flags. Click here to find out more.
Continue to review full report at Codecov.
|
The stpipe issue does need to be in first, as this uses the new Step.output_use_format parameter. I've tried to dot my github i's and cross the JIRA t's, but I'm not sure the linking of the stpipe issue to JP-1793 has gone through - I may need another link somewhere? The stpipe issue is here and the PR here |
* flake8 * name the model not the container * Allow for ModelContainer as input * might as well output a ModelContainer instead of a list of models? not sure * stpipe changed new Step parameter name
Closes #5511
Resolves JP-1793
Description
This PR addresses issues with the Step/Pipeline identity crisis that calwebb_wfs-image3 - err, wfs_combine - found itself in.
The step now uses datamodels.open() instead of load_as_level3_asn() - this should allow the step to take a ModelContainer as input, though the formatting will need to follow the level 3 associations, i.e. presence of member pairs under products, expnames provided, etc. Documentation might be on the to-do list to make this clear...
The step now follows a more traditional structure for saving results, though this requires an stpipe PR to be merged that allows for Step instances to access the save_model(format=...) parameter.
Checklist