IPP Progress Report for the week 2012.04.16 - 2012.04.21

(Up to IPP Progress Reports)

Eugene Magnier

  • more work on the relphot-ed zero points. In order to speed up the testing, I have threaded the most expensive steps of the analysis process in relphot. This gains almost a factor of 2 in processing time (though some of the gain actually comes from reducing the number of times we calculate and report advisory statistics on the whole database).
  • some images are duplicated in the database (due to an old addstar queuing problem which has since been addressed). these image duplicates can cause errors with the zero point calculation, so I need to be a bit careful.
  • WISE allsky is out, so i downloaded it and updated the dvo ingest functions. this will be merged with the full dvo database before general release
  • worked on dvomerge to allow merging into a parallel database. code is in place, but testing is still needed.

Serge Chastel

  • I created a database containing all files in the nebulous directory on the mhpcc storage nodes. I'm currently using it to check if I didn't miss anything for the ippb02 repair.
  • Then I will use it to classify the files in the various types to automatically build the nebulous usage information.
  • I also wrote pieces of software to update the entries in that database.
  • I updated the condor framework with the new ipp tag

Heather Flewelling

  • relphot on SAS.20120403
    • we need to think about how to do this more periodically on the masters.
  • sextractor
    • helped sidik with tuning sextractor for tdi
  • megacam
    • made a super-stack (stack of all MD01 u band) for various configurations of trunk and recipes. Discoveries:
      • eam_branches/20110505 looks lovely
      • r33783 looks bad
      • r31601 with r33783 ippconfig looks worse than eam_branches/20110505
      • r31601 with eam_branches/20110505 config looks good
      • from this, decided to do a binary search of trunk (using eam_branchse/20110505 ippconfig directory)
        • r31668 through r33783 all produce bad stacks
        • still doing binary search
  • czar
    • mostly uneventful, including when Cindy et al worked on the computers at mhpcc.
  • abstract aas

Roy Henderson

  • ippToPsps:
    • fixed queuing bug - an old UNIQUE constraint was still in place on 'pending' table. Removed
    • improvements to exception handling
    • installed and tested 64-bit Java JVM
    • loaded objects to production for first time: for new SAS
    • added more complete data to the PSPS news page detailing each survey including counts of stacks, frames etc
    • issues with Kron fluxes highlighted by Nigel: now scaling by exposure time
  • Czartool:
    • purged magic and magicDS
    • purged burntool
    • added staticsky
    • added chipbackground
  • Ken's/Chris' masks for false detections:
    • updated PSPS FAQ with masks for poor and bad, as well as a union of both
    • updated PSVO plugins accordingly
  • PSVO:
    • fixed minor bug where selecting a mask and using the mask generator popup would not work for mask > 32 bits
    • fixed bug where angles outwith constraints would not be 'remembered' by coords popup
    • released new version
  • Other:
    • meetings and demo-prep for ORR
    • investigated strange case of nDetections = 1 but no corresponding P2 detection in new objects loaded to PSPS. Answer: relphot not run on DVO master

Mark Huber

  • ppStack continued in tracing out input warp problems.
  • ppSub revived test sets, re-visited the few recent samples from QUB page.
  • MD08 full reprocessing and MD08 refstack creation.

Bill Sweeney

  • Continued with the kron measurements investigation. Fixed the cause of the huge negative kron fluxes by re-implementing the window function. This took several iterations.
  • Spent almost all of Thursday recovering from crash of ippc17, the host that runs the postage stamp server and data store.
  • out of the office a couple of days.

Chris Waters

  • Calculations on boosting the noisemap. Using a set of detections measured on the bias images, I was able to construct a model of the amount the noisemap under measured the detector noise. As this was measured on teh bias frames, any issues with the gain should not influence this result. Calculated boost factors, and applied the result to the noisemaps and ran a footprint study in g. This did not show the expected dramatic improvement in the number of false positives.
  • AAS Anchorage meeting abstract.
  • Determined with Ken and Gene that the corner anti-glows visible on many exposures are a result of the video cells on that OTA. This suggests that we should be able to model this effect and remove it with the dark. This will remove one more detector issue that sometimes prints through into the stacks.