Add FIBERSTATUS bits for VARIABLETHUR and NEARCHARGETRAP

[sbailey]

Several tasks for adding new non-fatal but problematic bits to FIBERSTATUS:

• Add new FIBERSTATUS bits in desispec.maskbits:

  • VARIABLETHRU - fibers 1098, 3994, 4720, 4891 (re)identified in N(z) K-S test to identify bad fibers in jura #2283
  • NEARCHARGETRAP - edge effects of CTE corrections, fibers 551-553, 3969, see Jura bad fibers 551-553 due to z1 CTE edge effects #2286
    (Alternate bit name suggestions are welcome.)

• Add bookkeeping in desispec.calibfinder to identify which fibers should be masked based upon new optional entries in desi_spectro_calib yaml files.

• Add those new entries to desi_spectro_calib yaml files.

• Add code to desispec.io.fibermap.assemble_fibermap to set those bits based upon calibfinder.

• Review functions in desispec.fiberbitmasking for how these should be treated. We should not set ivar=0 for these bits, but we should probably not use them for sky and stdstar/fluxcalib calculations. It's not clear to me whether the current set of functions supports that case.

In the future there may be additional FIBERSTATUS bits for cases like this; problematic but not fatal.

[julienguy]

• See PR More FIBERSTATUS bits #2313 for the changes to the code.
• I have also updated the yaml files for fibers 551,552,553. See issue Jura bad fibers 551-553 due to z1 CTE edge effects #2286

[julienguy]

Added NEARCHARGETRAPFIBERS: 3969 to some configs of $DESI_SPECTRO_CALIB/spec/sm8/sm8-z.yaml . Tested with assemble_fibermap(night=20240213, expid=225329).

[julienguy]

Added VARIABLETHRUFIBERS: 4891 to $DESI_SPECTRO_CALIB/spec/sm3/sm3-b.yaml (equivalent to editing any other camera), which is a fiber we will probably never fully recover (see issue #2305). I am waiting for the end of the study on improvements to the TPCORR parameters before flagging other fibers.

[sbailey]

Bits have been added to both the code and desi_spectro_calib yaml files. In the end we have been fairly conservative for when to set the bit so that we don't over-mask. Even if NEARCHARGETRAP is strictly informative, many users instinctively require FIBERSTATUS=0 and would discard mostly ok spectra, so we only set it for the worse cases.