Fix Version/s: None
Reprocess the HSC RC2 dataset, as defined in
DM-11345, using Stack w_2019_14. Steps include makeSkyMap.py, singleFrameDriver.py, mosaic.py, skyCorrection.py, coaddDriver.py, and multiBandDriver.py.
DM-19150 Update /datasets/refcats/htm/* using meas_algorithms/bin/convert_refcat_to_nJy.py
processCcd/singleFrameDriver for all visits/ccds finished running.
There were two non-fatal errors seen in w_2019_10 (
DM-17940). The LengthError from --id visit=38950 ccd=64 is still seen ( DM-15121). DM-18400 has fixed the other LengthError with Number of matches is smaller than request. There are many more warnings in the logs.
Lots of WARN of Found version 0 reference catalog with old style units in schema. appear in the log, expected by the implementation of
RFC-575 as the reference catalogs on /datasets/refcats/ have not been updated.
Attempting to use meas_mosaic results to make coadd products (the HSC default) failed, due to a bug likely in meas_mosaic or afw from
DM-10156; the bug is reported at DM-19015.
Meanwhile, continue the coaddDriver runs by using the jointcal outputs instead. The output butler repo will be /datasets/hsc/repo/rerun/RC/w_2019_14/
DM-18300-jointcal/ (different from the usual path pattern!)
Had to break down the patches to finish multiband run for the COSMOS tract.
coadd/multiband: the number of data products and empty patches are the same as w_2018_15/
DM-14123 (this comment)
The logs have been copied to /datasets/hsc/repo/rerun/RC/w_2019_14/
The coadd and multiband products are based on the jointcal outputs (not meas_mosaic).
The new BrightObjectMasks should be in place (
DM-18147). Look out for "Unable to read brightObjectMask for" DataId warn messages in assembleCoadd.