IPP Progress Report for the week 2012.04.09 - 2012.04.13

(Up to IPP Progress Reports)

Eugene Magnier

  • helped Bill get started on investigations in the psphot kron radius values.
  • lots of iterations with Chris on the false positives & noise model
  • working on understanding why relphot is not getting the right zero point for some of the non-ubercal images. I discovered that those zero points can be reset for images in some of the RA 'orange slices' due to the 0,360 overlap. I ended up re-running relphot on a couple of times on the database. even parallel, this still takes lots of time.

Serge Chastel

Heather Flewelling

  • alaska travel plans
  • sdss stripe 82 to bill
  • psps
    • new sas dvodb is done (sas.20120403)
    • psps training session with roy and thomas
    • ingested sas with roy and thomas
  • installed ipp on new mac
    • commited changes to make it work, otherwise it was fairly clean (and not hacky)
  • ssp (ingested through end of March)
  • dvodb
    • lap/threepi are still building (3/5ths through on cam, 1/3 on ssky for lap)
  • megacam
    • md06 has bad astrometry
      • we get better on average astrometry if we do not use higher orders
      • there are still some that fail (why?)

Roy Henderson

  • finished UML diagram for ippToPsps code
  • shifted processing over to IPP user. Have been running as myself up until now
  • added unique index on objID on Object table to ensure no duplicates from DVO
  • various fixes to the Object batches after attempts to load to PSPS:
    • 'dec' is reserved word in MySQL, causing a problem for the OB dec coord field. Implemented a workaround
    • the crazy -999 PSPS default does not fit in a byte, especially not an unsigned bytes
    • units issues for mags - converting from millimags (provided by IPP) to mags
    • huge mag errors found in some DVO objects: setting to NULL using a cut set by Gene
  • problem with dvograbber program when running as IPP user: had been building against old libraries: fixed
  • changes to incorporate Bill's new skycells table, which is now part of gpc1
  • found duplicates in DVO object tables. Because DVO tables relate using row numbers, I keep the duplicates, populate mags and colors, then discard the duplicates
  • code in Dvo class to check for any ingested tables that need to be cleaned up
  • ippToPsps training for Heather and Thomas, helped loading new SA3
  • lots of documentation

Mark Huber

  • out sick Tuesday+Wednesday
  • DRAVG minutes, stack and staticsky samples for Dave Thilker, tracked down corrupted MD.GR0 stack Ryan Chornock found
  • ppStack: continued looking at the rejected (or should be) input images as well as the rejection logic.

Bill Sweeney

  • More psphotStack debugging. Decided to defer dealing with the memory explosion problem for now. Now fault the run if > 40000 peaks are found. We can revert those and rerun later once the code is ready.
  • spent quite a bit of time looking at the extended source results for galaxies. Noticed that we were occasionally getting extremely large negative kron flux values. Tracked this down to a problem with the kron window calculation. Gene suggested re-implementing the function using the psLib ellipse functions which are prepared for misbehaved values. With some analysis, reverse engineering and debugging got the new version to work, but discovered that many faint sources with poorly measured kron radii are putting garbage in the window. That is something to debug next week.
  • Worked a bit on modes for sctool for listing skycell coordinates.
  • Spent quite a bit of time trying to build ipp at MPIA. There is something very strange about the library configuration on their hosts. Finally found a symlink command that worked aroudnd the problem. They now have a recent IPP buidld
  • tried to investigate why the postage stamp server has recently been having intermittent communication problems with the MOPS data store. Didn't find anything specific. Need to change the code to try again occasionally
  • Examined the first STS operations. Found and fixed a problem with the recipe that caused it to use the synthetic reference catalog
  • packaged up some data for the outreach program. Cleaned up the scripts so that will be a 5 minute job next time.

Chris Waters

  • Dark/detrend/false positive work: http://svn.pan-starrs.ifa.hawaii.edu/trac/ipp/wiki/Background_Dark_Model#April92012
  • Finished creating new array of darks to cover the time period from 2010-01-23 to the present. Finished creating preliminary noisemap detrends for the full time range of the survey up until now. Ran set of tests to confirm that the OTA67 slopes are improved by these new dark models.
  • Worked at developing a method to scale the noisemap values to do a better job at rejecting false positives. The difficulty in this task is arriving at a conclusion that is minimally arbitrary.