update qa-fiberassign for stuck/broken/unassigned

[sbailey]

This PR fixes qa-fiberassign for counting of broken, stuck, and unassigned fibers. In particular:

• corrects identification of broken and stuck fibers (from FIBERMASK, not TARGETID)
• unassigned broken/stuck are not considered errors, but unassigned good fibers are
• only print non-error per-tile stats if using new --verbose option

Note: unassigned good fibers and insufficient standards/sky could be due to insufficient target densities in the input catalog; it is not necessarily an error for fiberassign itself, but we do want to flag it.

The mask values for broken and stuck are still hardcoded; updating that is related to the broader question of where these masks should be defined and I'm not trying to fix that here. See desihub/desiutil#113.

@weaverba137 please check this with your minitest notebook output, e.g.

qa-fiberassign --verbose $SCRATCH/desi/dev/end2end/fiberassign/tile*.fits

[weaverba137]

Here's the output of qa-fiberassign:

bweaver@cori08[1008]: qa-fiberassign --verbose --targets $SCRATCH/desi/dev/end2end/targets/targets.fits $SCRATCH/desi/dev/end2end/fiberassign/tile*.fits
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_01165.fits - ERROR
    Petal 9 has 5/10 stdstars
    3 stuck, 2 broken, 400 sky, 95 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_06927.fits - ERROR
    Petal 0 has 5/10 stdstars
    Petal 3 has 9/10 stdstars
    3 stuck, 2 broken, 400 sky, 94 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_11108.fits - OK
    3 stuck, 2 broken, 400 sky, 100 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_16870.fits - OK
    3 stuck, 2 broken, 400 sky, 100 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_18465.fits - ERROR
    1 unassigned good fibers
    Petal 1 has 9/10 stdstars
    Petal 8 has 8/10 stdstars
    3 stuck, 2 broken, 400 sky, 97 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_24227.fits - OK
    3 stuck, 2 broken, 400 sky, 100 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_28408.fits - OK
    3 stuck, 2 broken, 400 sky, 100 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_34170.fits - ERROR
    8 unassigned good fibers
    Petal 0 has 8/10 stdstars
    Petal 1 has 8/10 stdstars
    Petal 7 has 8/10 stdstars
    Petal 8 has 8/10 stdstars
    Petal 9 has 8/10 stdstars
    3 stuck, 2 broken, 400 sky, 90 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_39942.fits - ERROR
    Petal 1 has 8/10 stdstars
    Petal 2 has 5/10 stdstars
    Petal 6 has 8/10 stdstars
    Petal 7 has 9/10 stdstars
    Petal 9 has 9/10 stdstars
    3 stuck, 2 broken, 400 sky, 89 stdstars
/global/cscratch1/sd/bweaver/desi/dev/end2end/fiberassign/tile_45704.fits - ERROR
    1 unassigned good fibers
    Petal 0 has 8/10 stdstars
    Petal 5 has 9/10 stdstars
    Petal 6 has 8/10 stdstars
    Petal 7 has 7/10 stdstars
    3 stuck, 2 broken, 400 sky, 94 stdstars
6/10 tiles had errors
----------
56.1% targets (129340/230516) covered by 10 tiles
  Dropping 101176 targets not covered by any tile:fiber
Fiber assignment efficiency
  ELG       35.8%  (17348/48429)
  LRG       72.2%  (5798/8025)
  QSO       94.2%  (3449/3661)
  MWS_ANY    3.8%  (1031/27165)
  BGS_ANY   31.8%  (13391/42061)

---------------------------------------------------------------
WORK IN PROGRESS
The stats below mix dark/bright/gray without considering that a
dark target can't be observed on a bright tile
---------------------------------------------------------------

Percent of targets covered exactly N times
            1     2     3     4     5     6     7     8     9
          33.7  26.9  17.9  11.4   4.9   3.8   1.1   0.3   0.0

Percent of targets covered at least N times
            1     2     3     4     5     6     7     8     9
         100.0  66.3  39.4  21.5  10.1   5.2   1.4   0.3   0.0

Assignment probability if covered exactly N times
            1     2     3     4     5     6     7     8     9
ELG       17.6  33.0  46.5  59.3  70.4  79.1   ---   ---   ---
LRG       59.5  79.7  86.1  87.5  98.6   ---   ---   ---   ---
QSO       87.3  99.1 100.0 100.0 100.0   ---   ---   ---   ---
MWS_ANY    1.1   2.2   2.8   4.4   7.2  10.2   9.7  11.2   ---
BGS_ANY   31.4  36.1  26.7  18.1  32.5  45.0  48.1  48.3   ---

[weaverba137]

The code looks fine. Please double-check the qa-fiberassign output I pasted into the PR.

[sbailey]

Looks good; thanks for the cross check. Note that the efficiency numbers aren't particularly meaningful here because of large edge effects from this tiny field.