Fix Version/s: None
Component/s: obs_base, obs_cfht, obs_decam, obs_lsstSim, obs_sdss, obs_subaru
Sprint:DRP F16-2, DRP F16-3, DRP F16-4
Team:Data Release Production
RFC-204 by adding new entries for all patch/tract and config mapping definitions to .yaml files in daf_butlerUtils, and removing any such entries that are identical to the common ones from .paf files in obs* packages.
I think the "common" entry can usually be defined by consensus between any two of obs_cfht, obs_decam, and obs_subaru (and frequently all three). If there are any patch/tract or config datasets for which no two cameras agree, I think we should use obs_subaru's definitions (but I doubt there are any such cases).
Entries that are not identical to the common ones should not be removed on this issue (that should make this change entirely backwards compatible), but should be documented in new per-camera issues for later standardization.
- is blocked by
DM-6858 Mapper tests require modification when new datasets are added
- is triggered by
RFC-204 Move all config and tract/patch dataset definitions to base CameraMapper
- relates to
DM-7512 obs packages need a unified test framework
DM-4927 ci_hsc doesn't produce readable processCcd_metadata
RFC-233 Proposed change to datasets for LsstSim mapper
- mentioned in
I notice that deepCoadd: and related exposures have "level:" set in conflicting ways. Is this entry needed?
The fact that they all work despite having different definitions suggests that it's not needed, but I don't actually understand what it does. Could you ask this on community?
This came up because I was ending up with deepCoadd and deepCoad_calexp in two places for the non-HSC mappers, and I had to then choose whether to delete these entries from the other mappers (violating our previous assumption that we would not make any material changes to the other mappers on this ticket), or to add an override by those same names in the other mapper.paf files.
Since K-T is not willing to vouch that changing the level won't affect things (and in fact, he wants to remove this entirely in later code), I vote that we punt on this issue and retain compatibility with the previous paf files for now by moving these two entries from datasets: to exposures: in all the paf files.
The differerence logs for these changes will now be moved to individual RFC documents for each of the mappers.
I notice that deepCoadd: and related exposures have "level:"
set in conflicting ways. Is this entry needed?