IPP Progress Report for the week 2010.01.25 - 2010.01.29

(Up to IPP Progress Reports)

Eugene Magnier

Heather Flewelling

• culled raw files, freed up 1% of space.
• liason duties: emailed Matt Holman
• fixed various cells on the magic mask.
• created magic verification test set: found from various emails between Gene, Ken, and Will, and added extra md and 3pi. The 'best' set:
  • magicverify.3pi.%20100127 3 pi fields
  • magicverify.md.%20100127 md04 with lots of satellite streaks
  • magicverify.dec07.%20100127 dec=-7.0
  • magicverify.ss.%20100127 sweetspot
  • magicverify.lgl.%20100127 low galactic lat
  • magicverify.m31.%20100127 m31
  • magicverify.shortmd.%20100127 mdfields g,r,i,z 30s
  • magicverify.md20.%20100127 mdfields (sets of 20, g,r,i,z)
  • magicverify.dec35.%20100127 dec=-3.5
• stacked and diffed and diffed and magicked and reverted and reverted as necessary.
• created VerifyStreaks for the above labels, and investigated causes of streaks.

Bill Giebink

Roy Henderson

• workaround for issue of blank rows in DVO tables/jumps in smf data
• now batching-up data as requested by PSPS: includes job and batch dirs with XML manifest files (md5sum etc)
• created 'product' on Maui datastore (PSPS_test)
• throughput test perl script now publishes data to datastore for Conrad to pick-up
• lots of back and forth with Conrad regarding specifics of FITS format
• working on SQL queries to pull out locations of diffs/batch files

Paul Price

• Added new camera format configuration (format_20100122.config) to handle change to temperature keywords on camera side.
• Added function, pmConfigRecipeValueByView to get recipe values dependent upon the camera component of interest. Applied to pattern subtraction within ppImage.
• Parameter tuning for stacks; checked works on diff
• Magic recertification
  • Merged pattern subtraction enhancements into branch
  • Diagnosed memory growth problems as due to change in temperature keywords
  • Merged new gpc1 configuration file into branch to get temperatures compatible with existing dark frames
  • Fix to PSF-matching code to measure normalisation independently from other kernel parameters: normalisation is measured in an aperture of carefully-chosen size
  • Fix to PSF-matching code to only worry about lit pixels to prevent covariance matrix from being evaluated at non-constrained position
  • Fix to variance renormalisation code to only worry about lit pixels to prevent insane renormalisation values
• Compared diffs with dual/single and convolved/unconvolved stacks: now obsolete following above fixes, to be re-done with new code

Bill Sweeney

• Visiting MPG.

Chris Waters