Keep record of patching process log files

[araichoor]

We did so far two rounds of patching of the fiberassign files:

• one in Oct. 2021:
  • script here: PR Add old script for patching fiberassign files (from Oct 5/2021) #434
  • log files (~0 GB): I see some here /global/cscratch1/sd/dstn/patched-fa-v2/patch-??.log, but those date from Sept. 8; @dstndstn could you confirm / or point to the correct log files?
• one in Oct. 2022:
  • scripts here: PRs Patch 202210 #436 and Patch 202210 v2 #438
  • log files here (3.4GB): $DESI_ROOT/users/raichoor/fiberassign-patch/20221019-patch/patch-??.log and $DESI_ROOT/users/raichoor/fiberassign-patch/20221019-patch/???/fiberassign-??????-patching_202210.ecsv.

We want to keep record of those files.
One suggestion was to create some svn folder in $DESI_TARGET/fiberassign/tiles/, provided that those are not too large for that.
@sbailey : let s iterate in this ticket to make the final decision.

[moustakas]

If I may interject: let's try to write this "record" in the form of documentation that we can use as-is in the EDR and DR1. It's not quite a "known issue" but something like the list that @geordie666 has been curating for the imaging data releases?
https://www.legacysurvey.org/dr9/issues/

[araichoor]

thanks for the suggestion.
I m happy to take the lead on this (with help on technical details on the first round of patching from @dstndstn).
I just need to know where that should be done.

one question, though: what should be mentioned?
would something like the following be in the good direction:

• edr: the fuji redux has been launched on Jan. 24, 2022 ([desi-data 5825]), so after the 1st patching (but way before the 2nd one)
  • so the "record" could be: "there are some inconsistencies in xxx columns due to some fiberassign bug, propagated downstream" (and add some details)
• dr1: that iron redux will use the twice-patched files, so hopefully the correct thing
  • so the "record" could be: "some differences in xxx columns could be seen w.r.t. edr; the dr1 values are the correct ones"

also, kind of related: we ll likely create a svn tag for each release (at least fuji and iron), "freezing" the fiberassign files version that went into the production.