Implement GBU conditions from Steve's pipeline

[mgalloy]

Steve's implementation is in UCoMP_GBU and the current conditions are:

 1   spar guider control loop is not locked
 2   spar guider intensity below threshhold
 4   background > gbu_max_background
 8   background anamolously low, defined as < gbu_min_background
16   spurious Stokes V signal is above threshhold
32   chi-squared of occulter fit is above threshhold for either camera 0 or 1
64   deviation of the difference of the image with the median is above threshhold

Need:

• [ ] GBU thresholds for wave regions 670, 761, 802, and 991.
• [x] What happened before 20210727 with the SGS values? Is using RAZR as LOOP and DIMS as DIMV sufficient, or are there other changes I should make?
• [x] Should I move ucomp_gbu_check_nominal_temps and ucomp_gbu_check_identical_temps to quality?
• [x] How to implement the ucomp_gbu_median_diff check? Does it really need the daily mean/median for each pixel?

Produce:

• [x] GBU text file with columns: Filename, Wave, Ntune, GBU, Back, Ctlk, Reason, Sigmas

[mgalloy]

The GBU cutoff values are not available for 670, 761, 802, and 991.

[mgalloy]

There are epoch cutoffs for the SGSLOOP value:

if (jd lt 2459423.d) then begin
  sgs_loop = sxpar(header, 'SGSRAZR')
  sgs_volt = sxpar(header, 'SGSDIMS')
endif

Julian date 2459423.0 is July 27, 2021.

[mgalloy]

There are now two GBU checks that I used to do, that are no longer done:

• ucomp_gbu_check_nominal_temps
• ucomp_gbu_check_identical_temps

Maybe they should be moved to quality?

[mgalloy]

The last GBU test must be done after all the level 1 processing for all the files and all the previous conditions have been tested for all the other files, where "all" above means at least all the files of the same wave region and number of unique wavelengths.

[mgalloy]

Method:

• compute median, mean, and sigma for all files not rejected by other GBU conditions (Steve uses all the files of the same wave region and same number of tunings)
• then, for each file, for each polarization state and wavelength, computes a number of standard deviations from the median
• if the maximum number of standard deviations for any wavelength/polarization state combination is greater than the threshold, the file is marked bad

Need to:

• [ ] all sigma to averaging step
• [ ] do a GBU step after all the files are done processing for a wave region
• [ ] calculate temporary mean/median/sigma stage (don't write the files)
• [ ] do final GBU check

[mgalloy]

See #150 for the fix of SGS keywords before 2021-07-26.