DM-49274: Stop sky objects appearing in NO_DATA regions

[leeskelvin]

No description provided.

[arunkannawadi]

It'd be great to split commit acee1a3 into two - one rearranging the names and one adding the nodata flag, but I am not insisting on it.

[mrawls]

Please also add base_PixelFlags_flag_nodataCenter and then @jrmullaney won't have to

[mrawls]

I suggest doing this in all the schemas where other flags have "center" variants

[jrmullaney]

I'm not sure adding it in this block will work. Aren't these pulling flags from the calexp, rather than the diff? My understanding is that the nodata flag that I want in the forcedSourceTable comes from the diff catalog, not the calexp. See here.
Having said that, if @leeskelvin wants to add that in a block above (as in that link-to PR), then I could delete my PR and have it all done in one go.

[mrawls]

That is probably true, but I still think we should add it here! I forgot this schema was for the merged forcedSource catalog, and not for the one upstream a step or two before the merge happens, which doesn't care what flavor of exposure it is run on.

[leeskelvin]

I don't mind adding it here, and the 4 other schema files at the same time. Troubled by this though:

I'm not sure adding it in this block will work.

If I add *nodataCenter into all 5 schemas, will something break?

[mrawls]

That feels like a "be sure to run ci_*" flavored question 😅

When @jrmullaney said he wasn't sure it would work, I think he meant it might not let him do the downstream selection/filtering on merged forced source tables to the extent we want, not that it would break anything.

[leeskelvin]

I've added base_PixelFlags_flag_nodataCenter to the five schema YAMLs in this repo. Currently running Jenkins with:

lsst_distrib lsst_ci lsst_middleware lsst_sitcom ci_cpp ci_hsc ci_imsim ci_middleware

[leeskelvin]

After discussion with @jrmullaney, I've removed this column. The column needed is different than a straight-forward propagation, and requires more thought and testing. @jrmullaney will follow this up on DM-49554.

[leeskelvin]

It wouldn't be an @arunkannawadi review without at least one request to split a commit into multiple commits 😉. Commit split! Just waiting on a reply from @jrmullaney / @mrawls Re: *nodataCenter flag additions into the schemas.