DM-32700: Add a task to compute noise correlations

[arunkannawadi]

No description provided.

[fred3m]

I might be misunderstanding the code, but I don't think that background subtraction is working properly (see comments in the PR). Either way, there should be a test to make sure that includes masked pixels and background subtraction to verify that it's working correctly and to prevent bitrot.

[fred3m]

Should this use the same mask planes as the variance maskPlanes?

[arunkannawadi]

Given that variance is estimated largely from the background, that seems reasonable to me, especially in the defaults. Of course, there's nothing in the code that requires them to be the same. I used the ones listed in ScaleVarianceTask as a starting point here.

[fred3m]

Is this the proper way to do this (I'm asking honestly here, as I don't know the proper way to set defaults for a subtask that might be different than the subtasks default config)?

[arunkannawadi]

It is certainly a proper way. The defaults here override the defaults set in SubtractBackgroundTask and they can be overridden by specifying different config at the runtime. As to whether the default values make sense, that's simply motivated by ScaleVarianceTask which has been used on HSC and DeCam analyses, so I'm running with that.