More FIBERSTATUS bits

[julienguy]

This PR addresses issue #2304

• Add desispec.maskbits.fibermask bits NEARCHARGETRAP ("Fiber trace near charge trap in one of the CCDs") and VARIABLETHRU ("Fiber has throughput variations we cannot model well")
• Add NEARCHARGETRAPFIBERS and VARIABLETHRUFIBERS to the list of keywords of calibfinder.badfibers
• Add code in desispec.io.fibermap.assemble_fibermap to set NEARCHARGETRAP and VARIABLETHRU fibermask bits by parsing the DESI_SPECTRO_CALIB configs of all cameras using the calibfinder.badfibers function and the raw fits file headers.

Example:

from desispec.io.fibermap import assemble_fibermap
from desispec.maskbits import fibermask

hdulist = assemble_fibermap(night=20220102, expid=116442)
# INFO:fibermap.py:1147:assemble_fibermap: Bad fibers from DESI_SPECTRO_CALIB: NEARCHARGETRAPFIBERS:[551 552 553]

fmap = hdulist["FIBERMAP"].data
print(fmap["FIBER"][(fmap["FIBERSTATUS"] & fibermask.NEARCHARGETRAP) > 0])
# [551 552 553]

To run this, you need an updated version of DESI_SPECTRO_CALIB svn and you need the environment variable FIBER_ASSIGN_DIR=/global/cfs/cdirs/desi/target/fiberassign/tiles/trunk. It is set by the 'main' version of the DESI environment source /global/cfs/cdirs/desi/software/desi_environment.sh main but not the default one.

So far, DESI_SPECTRO_CALIB has been updated only for the problematic fibers 551,552,553 of SM10 r1. I will work on flagging more bad fibers in the yaml files (but we don't need to wait for those changes to the calibrations to merge this PR).

[akremin]

Thank you, Julien, for both the code and for updating the changes.rst. As usual, the changes look good. I'll merge and do further testing as you update more calibration files.