Closed sbailey closed 3 years ago
@sbailey: The for nskey in ... blocks
work perfectly.
For the record, these are used to generate HEALPixel-split files via the bundlefiles
option as follows, e.g.:
LS=/global/cfs/cdirs/cosmo/data/legacysurvey/dr9
select_targets $LS/south/sweep/9.0 $CSCRATCH -s2 $LS/north/sweep/9.0 --bundlefiles 1 --nside 8 --numproc 4 --gaiasub --scnddir $CSCRATCH/ADMsecondary --dark-subpriorities subpriorities-dark.fits --bright-subpriorities subpriorities-bright.fits
If you were to try this, you'd find that the --dark_subpriorities subpriorities-dark.fits --bright_subpriorities subpriorities-bright.fits
flags are added for each HEALPixel in the slurm script. Kudos!
I tested that this works for each of select_targets
, select_skies
and supplement_skies
. It does.
Let's check in later on on Slack and discuss next steps (which will start by merging this PR).
Thanks for checking. I updated docs/changes.rst and added desitarget.subpriority to api.rst . I think this is ready to merge.
Feel free to merge, once tests pass.
This PR adds header keyword provenance for what subpriority overrides were used, if any. Example outputs are in /global/cscratch1/sd/sjbailey/desi/targets/provenance, with headers e.g.
I went with the DEPNAMnn/DEPVERnn to avoid cryptic 8-char limited keywords like SUBPBROV.
Belt and suspenders, I also added and "CMDLINE" keyword that records the full command line call directly from sys.argv, e.g.
this applies to select_targets, select_skies, and supplement_skies, which are the 3 scripts that accept subpriority overrides.
@geordie666 please double check my implementation of passing the options into the
for nskey in ...
blocks. I think I followed your instructions correctly, but that's also the piece I didn't know how to test / verify.