Step 1, rebuild calibs from scratch following Lee's Merian docs, https://hackmd.io/@lsk/merian

Step 2, update hits2014 and hits2015 chained collections, ideally with new skymap (DM-34516) and definitely with new calibs

Step 3, run ApTemplate

Step 4, run ApPipe

Step 5, make some plots

For rebuilding calibs from scratch, I also note that that cp_pipe documentation is now quite extensive: https://pipelines.lsst.io/v/daily/modules/lsst.cp.pipe/constructing-calibrations.html

Quick update, calibs are now rebuilt and are the new "top level" default in DECam/calib in /repo/main. The commands I used to rebuild the calibs and the resulting logs are in /project/mrawls/hits-calibs-rebuild. See specifically bias.sh and flat.sh, and note that these are not intended to be run as standalone bash scripts, but instead each line was pasted into a Terminal manually.

Template building is underway, and didn't take eons for the processing to begin starting this time!

Next up will be ApPipeWithFakes.

Logs and bps submit files for pipeline runs are in /project/mrawls/repo-main-logs/DM-34710.

Ran into an issue where I didn't initially specify which exposures (visits) to use in template-building vs in the main ApPipe processing, and the good seeing selector chose nothing for one of the three HiTS fields. In addition, there was a known-trash-WCS 2014 visit, 289777, that apparently has great seeing and showed up as both a template and science in a totally wrong spot on the sky (-2 dec instead of +2 dec, cute). It has been entirely removed from DECam/raw/all, finally.

I re-ran both the ApTemplate step (took about 20 minutes to build the graph) and the ApPipeWithFakes step (took about 24 hours!!! to build the qgraph), with these inputs and data queries -

ApTemplate:

  payloadName: DM-34710/templates
  butlerConfig: /repo/main/butler.yaml
  inCollection: DECam/defaults/hits2014,u/mrawls/DM-34710-hits-crosstalk-sources
  dataQuery: "instrument='DECam' AND skymap='decam_rings_v1' AND detector NOT IN (2, 61) AND exposure<300000"

ApPipeWithFakes:

  payloadName: DM-34710/hits2015-fakes
  butlerConfig: /repo/main/butler.yaml
  inCollection: DECam/defaults/hits2015,u/mrawls/DM-34710-hits-crosstalk-sources,u/mrawls/DM-34710/templates
  dataQuery: "instrument='DECam' AND skymap='decam_rings_v1' AND detector NOT IN (2, 61) AND exposure>400000"

It is a miracle that the reprocessing was not bitten by the NCSA filesystem having a bad day. At a glance, the output all looks good now, finally, and an analysis notebook is forthcoming.

Will you please review this, Ian Sullivan? It's just the one PR. The main thing to look at is the big analysis notebook, and there is also a trivial plotting script update that was helpful when I was troubleshooting my first attempt at this reprocessing.

This looks very good. I think it would be very informative to include one more section at the end (before the troubleshooting appendix, which I like) that includes a few plots showing completeness of the Fakes at a few different flux or significance thresholds.