Closed moustakas closed 7 months ago
@moustakas the pipeline hadn't tripped over it because we have had bad weather recently and haven't had any science data since the change.
I think we should proceed with modifying the mgii code in desispec given that Stephen's change was probably an intentional refactor.
I will submit a PR and self merge.
This is fixed via desispec in PR #2168
Ugh, apologies for breaking this with a cosmetic fix as part of the Redrock templates refactor. Thanks to the attention/debugging from @moustakas and @abhi0395 and for the weekend fix from @akremin . My penance shall be to write a unit test that would have caught the incompatibility in the first place.
@sbailey I'm surprised we haven't seen this crash in the daily pipeline yet, but I'm fairly certain that https://github.com/desihub/redrock/pull/271 broke the MgII after-burner by removing the
template_dir
argument ofredrock.find_templates
(see here), which is used bydesispec.mgii_afterburner.load_redrock_templates
here.I can easily fix this in a PR but there are a few different ways of doing so. First, do you want me to fix this in
redrock
ordesispec
? And if inredrock
, do we want to add backwards compatibility for thetemplate_dir
argument?