For the coming engineering run, I put a script version of the fluxCalQA at /work/tanaka/fluxcalqa/fluxCalQaScript.py

You can launch it this way:

$ python fluxCalQaScript.py  rerun  visit  outputDir

For example,

$ python fluxCalQaScript.py rerun/kiyoyabe/erun/run15/20240328 107831 /work/tanaka/fluxcalqa/

This will generate two plots: one is i-mag difference between pfsSingle and PS1 photometry for FLUXSTDs, and the other is g-y color. I have shown these plots at telecons and slack, and you should be able to read them (if not, let me know).

It also dumps summary stats in the log:

fluxCalQAStats:  visit  median(imagDiff)  sigma(imagDiff)  median(colorDiff)  sigma(colorDiff)

Please grep.

It looks that /home/tanaka/processing/synthmag is not accessible so I couldn't import the module. Could you copy them to somewhere accessible (e.g. /work/tanaka/fluxcalqa/)?

I found that there was `/work/tanaka/processing/synthmag/`.  Is it the same as `/home/tanaka/processing/synthmag/`?

I forgot to change the permission of my home directory, but all of my work is in /work/tanaka/.  /home/tanaka/processing/synthmag/ and /work/tanaka/processing/synthmag/ are the same directory as you guessed.

I assume this will end up in the drp_qa repo, so I can help set that up if needed. 

• There is still hardcoded path in synthmag.py
  • self.basepath = '/home/tanaka/processing/synthmag/'
  • so I need to specify the path explicitly like `synth = synthmag.synthmag(path=...`
• The catId of FLUXSTD was specified as 4 but now the catId has been changed to 3006 (from the last run, I think)
• It looks that the calculated magDiff for ps1_i band is all nan (other bands look OK), so `makeMagDiffPlot` fails to make a plot
• Anyway, it would be very helpful if somebody can check whether this is working in the recent rerun and whatever visits

I will fix the code when I get a chance. Please leave the rerun and visit numbers you looked at in this ticket, so I can reproduce the problem.

I'm not sure which one is used for the report above but probably the same in both. This script file includes my hacks.

python /work/kiyoyabe/erun/run17/calibs/generateCalibs/20240619/scripts/fluxCalQaScript.py rerun/kiyoyabe/erun/run17/20240528 111347 /work/kiyoyabe/erun/run17/calibs/generateCalibs/20240619/qa/old /work/drp/CALIB
python /work/kiyoyabe/erun/run17/calibs/generateCalibs/20240619/scripts/fluxCalQaScript.py rerun/kiyoyabe/calibs/run17/20240619/science 111347 /work/kiyoyabe/erun/run17/calibs/generateCalibs/20240619/qa/new /work/kiyoyabe/erun/run17/calibs/generateCalibs/20240619/CALIB-kiyoyabe-20240619

I've started working on moving this into drp_qa and should have a PR within a day (or so). I've copied the synthmag and fulxCalQaScript.py, so please let me know if the originals change for any reason.

Comparison between plots produced by Masayuki Tanaka original code vs the fluxCalQa.py.  The command line produces all filters in a single plot, so the comparison is just between the i. The median values are similar (both bad in 110477 and both good in 110478). The scatter/rms values differ slightly and by-eye there appear to be a few different outliers.

What's the significance of circles, diamonds, + and x?

The markers correspond to the spectrograph. On the newer plots (with all the filters) there is a legend in the upper right. The markers are not consistent between the old plots and new.

This has been merged, making the fluxCalQa.py script available via the command line. Further improvements to the QA should be in new tickets.