IPP Progress Report for the week 2010.07.12 - 2010.07.16

(Up to IPP Progress Reports)

Eugene Magnier

The Sersic model fitting is now converging well for unconvolved model fits. I have started to modify the convovled model fit process to match. That codes needed a bit of rework in order to accept the same kinds of modifications. I have made some progress on adjusting the psf-convolved fitting, but this effort is not yet finished.

I have also made some other needed fixes: I have identified the reason psphotForced was failing to return negative fluxes even in empty test images. There was an artificial limit on the fit parameters to prevent negative sources. I also fixed the calculation of the difference stats related to mask vs unmasked pixel fractions: it turns out the analysis was counting pixels outside of the standard aperture as 'masked' instead of simply ignoring them. Finally, I have modified dvomerge to address a problem with the merge speed: it was looping over all catalogs in the output database (and reading them all into memory), rather than only merging the catalogs in the input (presumably smaller) database.

I got sucked away to Jury Duty on Thursday, and will be on the jury again next week Wednesday and Thursday.

This was Paul's last week with the IPP Team : We will all miss him!

Serge Chastel

  • IPP / IppMonitor
    • Warp Processed Exposures scripts (both in raw directory)
    • Simplification of stack summary script
    • JavaScript? is now used in IppMonitor, precisely in ipp.php (to update dynamically the cells of a table)
  • PSPS
    • Meetings with Conrad / Roy
    • Writing a report/description of the PSPS subsystem for Conrad (possibly being reused to complete/update the wiki pages).

Heather Flewelling

  • published mops smf files
  • detrend work
    • queueing up of various dark detrends
    • plotting/fitting darks to understand why the dark residuals are not like what they were on previous runs. problems include:
      • first or second exposure of a night are bad
      • a whole set of darks for a night are bad
    • I've randomly select a set of good exposures for a dark run
  • vacation day july 15th

Roy Henderson

  • Short week due to vacation on Tuesday and Wednesday
  • PSPS
    • Batches finally loading to sandbox with new DXLayer, new batch structure and new schema.
    • developed new test batch to provide Jim with necessary data to look at 'orphans' problem
    • bug in new DXLayer: objIDs were all screwed up. Tracked down to fitsConverter: code was built with wrong cfitsio lib
    • finally loaded 'bad' astrometry frame for Sue to test her new code
    • tried to explain the inner workings of the PSPS interface to Serge, so he can begin designing his test code
  • IPP
    • Czar on Friday:
      • various hassles with reference stacks
      • more work on roboczar

Bill Sweeney

  • modified the skycells program to produce skycells that overlap.
  • queued MD08 reference stack using new tessellation with overlapping skycells.
  • since we have recent MD08, 9, and 10 data that has been warped to the older non-overlapping continued building of reference stacks with old non-overlapping tessellation
  • spent an unbelievable amount of time shepherding these runs through. Under heavy load stacking has high rate of failures due to system problems. All but one of the skycells is finally done.
  • Two days of this was as official processing czar.
  • We ran into two classes of problems with cfitsio triggered by the deep stacks (over 100 inputs). Spent many hours debugging these problems. Created a patched cfitsio that fixes one the of the problems. Still debugging the other errors.

Chris Waters

  • Burntool: Reprocessed all data back to 2009-05-01 via burntool pantasks, with the exception of 4 dates with unknown errors. Will debug in the future when we do science reprocessing.
  • Detector linearity: investigated detector linearity for cells and edge-of-cell rows and columns. Will make wikipage with results when conclusions can be reasonably made. Compared IPP linearity test results with Camera group linearity test results provided by Sidik.
  • Determined that the functions necessary to put WCS values into summary fits images does not exist.