Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: meas_mosaic, obs_subaru, pipe_drivers
-
Labels:None
-
Story Points:10
-
Epic Link:
-
Sprint:Per-CCD Sprint 1
-
Team:Data Release Production
Description
Following the work done on DM-20289, make the switch from the current "unity" scaling for the HSC camera focal plane to the actual physical focal plane units of millimeters (i.e. the representation preferred by the LSST stack). There are a few other packages (see list in components + fgcmcal, which doesn't seem to be recognized) that will need adjusting. The ultimate test for updates will be to run at least one full tract from the RC2 dataset (i.e. running through singleFrameDriver, skyCorrection, jointcal, coaddDriver, and multibandDriver) to make sure the results are consistent with that of the latest weekly reprocessing run.
Merging of this work will also be contingent on acceptance of an RFC to make the switch (this will be the triggered ticket for that RFC once it is written).
Attachments
Issue Links
- blocks
-
- Done
- is contained by
-
- Done
- is triggered by
-
- Implemented
- relates to
-
- Done
-
DM-19943 Per-CCD Distorted Astrometry
- To Do
-
DM-17862 HSC camera in obs_subaru has the incorrect pixel scale
- Won't Fix
-
- Done
Activity
Paul, would you be willing to have a look at the obs_subaru, pipe_drivers, and meas_mosaic PRs?.
Eli, would you have a look at the one in fgcmcal to make sure you're happy with the changes:
https://github.com/lsst/fgcmcal/pull/15
A recent Jenkins run passed and, just to be sure, I just kicked off another one.
Everything looks good to me.
Paul Price
added a comment - Everything looks good to me. And me too. I removed myself from the review list, so last one out turns off the lights...
Good to go!
Paul Price
added a comment - Good to go! Thanks both! Merged to master.
The implementation to convert HSC's camera geometry in obs_subaru to use units of mm is accomplished here. There are four packages affected: obs_subaru, pipe_drivers, fgcmcal, and meas_mosaic. Having setup the ticket branches along with weekly w_2019_26, I ran the full COSMOS tract from the RC2 dataset through singleFrameDriver, skyCorrection, jointcal, coaddDriver, and multibandDriver. The results of this run were compared with the w_2019_26 reprocessing run (i.e.
DM-19560). All qa scripts in pipe_analysis were run to compare the results. The differences are effectively trivial at the visit level (e.g. maximum differences at the 10^-11 mmag level). Slightly larger differences (rms at the 1 mmag level) occur at the coadd level. A direct comparison of the coadds reveals maximum differences at the ~0.003% level:All plots for the ticket run and the comparisons with the w_2019_26 reprocessing can be perused at https://lsst-web.ncsa.illinois.edu/~lauren/lauren/DM-20548/ and the individual w_2019_26 reprocessing plots are at: https://lsst-web.ncsa.illinois.edu/~lauren/lauren/DM-20548/w_2019_26/
Finally, here is an example of the output of the un-bitrotted bin.src/showCamera.py in obs_subaru:
