Note that we are going to have to skip filter NB0921, because it does not have any metrics defined in validate_drp.

Per discussion of 2017-12-04, we agreed that the success criterion for this ticket is to generate the Jointcal run which will actually be used for comparison with meas_mosaic. That will mean (has meant) running Jointcal on the data multiple times, figuring out what breaks, and filing blockers.

Latest news here: Pierre Astier's work on outlier rejection have helped photometry fits complete, but I still have NaNs sneaking in, apparently from NaN source catalog centroid_sigma values that aren't getting rejected by the astrometrySourceSelector (e.g., Centroid_flag is still False).

With the work done on DM-14786, I think this ticket may be finally ready to close? I have uploaded summary statistics to DM-11787.

The code to run the jointcal and validate_drp jobs lives in jointcal_compare. However, validate_drp has moved on enough since the meas_mosaic data was originally processed that I can no longer run validate_drp on the meas_mosaic output that lives in /datasets/hsc/repo/rerun/private/lauren/DM-11786, nor can I work with the old .json files produced from my old validate_drp runs on that data. I do still have the .rst summary files (as uploaded to DM-11787), which will probably be enough for these purposes.

Per discussion of 2018-08-06, we believe that this is done — there's now a complete Jointcal run for comparing with meas_mosaic. Further refinements to that will be handled on separate tickets.