desi_zcatalog backwards compatibility

[alxogm]

Hi , I should have catch this before, but we are just noticing that with the recent updates to desi_zcatalog we (Lya WG) can not longer use it in our existing quickquasars mocks, or new recent runs, since we have not updated our simulated spectra files. So I'm wondering if we should try to restore the compatibility in the desi_zcatalog, or we create our own script somewhere else?
A quick fix for now for us is to use an older tag of the code for existing mocks, and consider update the data model for new mocks, but I wanted to check if you discard to maintain the backwards compatibility.

Thanks!

[sbailey]

If I understand correctly, the issue is that quickquasars outputs the old zbest/redrock format (not splitting FIBERMAP vs. COADD_FIBERMAP), while desi_zcatalog now requires the new format. Is that correct?

For the future, I think you should update quickquasars to output the new format to match Everest/Fuji/SVDA.

For backwards compatibility, I see it would be helpful to update desi_zcatalog to support the old format to help with the transition period while using existing mock runs.

@alxogm could you help contribute both of those updates? Or otherwise at least provide an example pointing to files at NERSC for desi_zcatalog failing on quickquasar output? Thanks.

[alxogm]

Hi @sbailey,

If I understand correctly, the issue is that quickquasars outputs the old zbest/redrock format (not splitting FIBERMAP vs. COADD_FIBERMAP), while desi_zcatalog now requires the new format. Is that correct?

yes, exactly.

For the future, I think you should update quickquasars to output the new format to match Everest/Fuji/SVDA.

Yes, we'll update it for new runs.

For backwards compatibility, I see it would be helpful to update desi_zcatalog to support the old format to help with the transition period while using existing mock runs.

@alxogm could you help contribute both of those updates? Or otherwise at least provide an example pointing to files at NERSC for desi_zcatalog failing on quickquasar output? Thanks.

Yes, I'll do that and submit the PR. Thanks