IPP Progress Report for the week 2011.06.27 - 2011.07.01

(Up to IPP Progress Reports)

Eugene Magnier

Serge Chastel

  • IPP czar on Friday
  • Mainly busy with the gpc1 database replication issues
  • Talked with Conrad about PSPS Test Plan
  • Created database for Haydn/Paul Sydney to move stare data and free 30TB on Maui cluster
  • Finished? lapMonitor for Chris
  • Updated ippMonitor for Heather and Bill

Heather Flewelling

  • skyprobe
    • queued 3 months of skyprobe. stopped processing it about halfway because we were out of disk space
    • investigated how to sort out the bad observations, but couldn't find a good cut (yet). Bad observations are ones where the dome gets in the way
    • ingested comments/alt/az for the last 5 or so months for isp (from their headers).
  • addstar management
    • managing LAP addstar (cam and staticsky) and ThreePi? merging
    • debugged addtool/addTasks a bit

Roy Henderson

  • Czar Monday and Tuesday, but credit to Bill for doing most of the work
  • ippToPsps: loading of new LAP detections began, after the following work:
    • issues with likelihoods: no erfc() function in MySQL, so using a Java method and updating tables with a cursor, which is the fastest way.
    • newly populated detection fields, thanks to pointers from Chris:
      • analysis version in detection FrameMeta
      • bias and biasScat in detection ImageMeta
      • kronFlux and kronFluxErr in Detection table
      • psf, extended and cosmic ray likelihoods in Detection (see above)
    • added more values to config file for better control or program, including a 'force' option and 'test' mode
    • bug in FITS header reading code: if one header failed to read, then all subsequent headers also failed: fixed
    • new column for ippToPsps database: stage_id, which is either a cam_id (for detections) or a stack_id (for stacks)
    • worked on new query, with help from Heather, to improve my extraction of stack cmf path names from GPC1
    • updated more documentation
  • Quite a lot of user support:
    • some emails back and forth with Dave Clark regarding the usage of the postage stamp server
    • advice to Kent Wood and his team on how they can store data locally prior to our deletion of 3PI PSPS database. This led to...
    • ...a fair bit of work on queryDRLClient script to enable long queries to be run and the results retrieved and saved locally
  • Other stuff
    • planning and discussion for the ingest of new data earlier in the week
    • some more back-and-forth regarding the wiki migration: no have access to new machine, Raja is looking into upgrading old wiki to newest version

Bill Sweeney

Chris Waters

  • Stacking speed: Improved speed of KMM rejection code by tweaking algorithm. Sped up iterations for convergence from ~100+ in bad cases to ~5.
  • Diskspace: Added and tested cleanup code for stacks, and queued old quickstacks for cleanup. Delayed merging changes into tag because the disk space freed by cleaning stacks is somewhat small.
  • LAP: stopped g filter processing and launched y filter processing.
  • Dark: Confirmed that reprocessing with the current master dark is producing higher background noise than the original processing did with the old dark, as reported by Peter Draper. Detrend verify runs using a wide range of dates show that the current master dark does not correct any dark frame well. No good solution yet, other than attempting to construct a new dark.
  • WISE: Looked into importing WISE data into DVO. A new reading function needs to be defined, as the WISE catalogs do not easily match the existing functions.