Summary of extraction QA for edr2-20231203.

extractionQA-20231203-b.pdf

extractionQA-20231203-r.pdf

extractionQA-20231203-nm.pdf

Points:

• For each arm, the rms of chi is plotted for all fibers and visits in a plot (the colors of points show the chi’s rms values). The problems of extraction are listed for all arms in the slides (and in this post).
• As known, it can be seen from the comparison between the chi rms maps and the mean pfsArm flux maps that most large chi value spots be related to the bright fibers. A few fibers around very bright fibers also show large chi values. The effects of the bright fibers on the next fibers should be investigated more quantitatively (see P.3 of the b-arm slide for plots).
• All arm data of the quartz visits (97036-97047, 97050-97060, 97063. 97064) have large chi because of the scattered light problem. The scattered light affects the measurement of the profile widths in extractionQA and makes dσ/σ negative (meaning the profiles of calExp image is broader than those of model image).
• Two fibers, ID=1373 and 1415, are found to have a problem in their profiles in b3 data (P.12-16 of b-arm slide).
• The extraction quality of the r3 data of 82595-83249 visits (November run) is found to be poor, because the fiberProfiles of April run, in which r3 arm data was out of focus, are used as the calibration data. (P.10-12 of r-arm slide)
• As known, the m-arm data of visits (92536, 92537, 92733, 92734, 92892, 92893 93301, 93302) have large residuals because of the poor fiberProfile quality of m-arm data.
• The results of the detectorMap QA and extraction QA should be compared numerically as commented for weeks. I tried to show some comparison in the drp telecon, but there is not enough time to prepare plots.
   

extractionQA script is put at /work/shamano/fiberProfile/extractionQa.py on the pfsa server.

This script requires following arguments.

• The rerun name (positional argument)
• The output directory name (positional argument. If the directory does not exist, the script makes it.)
• Visit number(s) (optional argument with -v or --visit)

The default paths of the data repository and calib are set as follows.

• data repository: /work/drp (the path can be changed with --repository option)
• calib: /work/drp/CALIB (the path can be changed with --calib option)

Example command (extraction QA for all arms for the visits 98042 and 98043):

python extractionQa.py pfs/internal/edr2-20231203 extractionQa-edr2-20231203/ -v 98042 98043 

The arms and spectrographs can be limited with the optional arguments a (-arm) and -s (–specrograph).

Example command (extraction QA for r1 and r3 data for visit=98042):

python extractionQa.py pfs/internal/edr2-20231203 extractionQa-edr2-20231203/ -v 98042 -a r -s 1 3 

Output files are as follows:

• {visit}_{arm}{spectrograph}_chistats.pdf: the plots of some statistics from extraction QA (e.g., the standard deviation of chi for each fiber, the maps of X offsets and width differences).
  * {visit}

  _{arm}

  {spectrograph}_chifiber.pdf: the detailed plots of the fiber regions for the fibers with relatively bad extraction quality.
  * {visit}_{arm}{spectrograph}

  _chifiber.pickle: the numerical output from extraction QA is stored. dictionary object.

Plotter using the pickle files is in the notebook at /work/shamano/fiberProfile/extractionQA_summary.ipynb.

The extractionQA results have been delivered, so can we close this ticket? Masayuki Tanaka

Yes, we can close this ticket.