elp pipeline saves the "cat" and "segm" files regardless of save_results setting

[stscieisenhamer]

If save_results=False, the "cat" and "segm" files should not be created.

[mairanteodoro]

If I recall correctly, the requirement was that we should always create both cat and segm after running SourceCatalogStep, which, by the way, combines save_results and return_updated_model to determine the logic. Just for the sake of reference, this is what we have for running the step alone:

• save_results=True + return_updated_model=True => return updated model and save updated model + cat + segm.
• save_results=True + return_updated_model=False => return source catalog and save cat + segm.
• save_results=False + return_updated_model=True => return updated model and save cat + segm.
• save_results=False + return_updated_model=False => return source catalog and save cat + segm.

I understand that the behavior should be independent when running the step alone versus running the step from within the ELP. Conceptually, when running SourceCatalogStep from the ELP, we should be able to determine whether or not to save the cat + segm files based on the save_results parameter passed to SourceCatalogStep by the ELP.

[schlafly]

It's worth discussing other designs, but the challenge here, @stscieisenhamer , is that if you don't save the cat and segm files when you run the source catalog step then you haven't done anything. Brett has suggested another design where the source catalog step returns the catalog and segmentation objects, and then the function that writes their results gets overridden to understand how to write out the files if that is the only step that's called. That sounds like a good approach to me. Then we'd have to update the ELP to do the metadata update that puts the file name in the metadata so that tweakreg can find it.