Details
-
Type:
RFC
-
Status: Implemented
-
Resolution: Done
-
Component/s: DM
-
Labels:None
Description
This RFC proposes adding the obs_lsst package to the lsst_distrib meta-package by way of adding it as a dependency of lsst_obs.
obs_lsst provides camera descriptions for data that looks like LSST data, including that originating from test stands, ImSim, and PhoSim. As such, this package effectively renders obs_lsstSim obsolete. However, the future of obs_lsstSim is discussed on RFC-533: this RFC refers only to adding obs_lsst, and takes no position on deprecating or removing obs_lsstSim.
obs_lsst has been largely developed outside the regular DM development workflow. However, the code has recently been reviewed by Colin Slater, Tim Jenness and Wil O'Mullane. This review has resulted in a number of Jira tickets being filed requesting improvements to the obs_lsst codebase: here is a complete list. I propose that DM-17291 and DM-17299 are prerequisites for adopting obs_lsst (ie, they will block the implementation ticket for this RFC); the others will be added to the DM backlog and scheduled as regular work.
Attachments
Issue Links
- is triggering
-
- Done
-
- Done
- relates to
-
- Won't Fix
-
DM-17299 Clean up obs_lsst scripts
- Done
-
- Done
-
- Done
-
- Won't Fix
-
- Won't Fix
-
- Won't Fix
-
- Invalid
-
- Implemented
Activity
I'm a little confused. If you install lsst_distrib, even with lsstsw, you shouldn't get the test data.
The problem with making it smaller is that it has test data for four different instruments (effectively), so it is a different animal than the other test data repositories.
Ah! The name of the package confused me: I thought there was a setupOptional(testdata_obs_lsst) in obs_lsst. I see that there isn't, so I withdraw that complaint.
However, I now have to retract my slack comment about being happy with the name of testdata_obs_lsst. Since it's not a requirement of obs_lsst, the name is confusing, as I just demonstrated.
The name is confusing. I don't dispute it. Moving to #dm-naming.
I'm not seeing any fundamental issues here — work is still in progress to clean up obs_lsst, but it seems clear that we will be adopting it.
That size relates to ci_lsst and not obs_lsst so I don't think it has to be discussed as part of this RFC. ci_lsst is not being added to lsst_distrib.