IPP Progress Report for the week 2010.03.08 - 2010.03.12

(Up to IPP Progress Reports)

Eugene Magnier

Heather Flewelling

Bill Giebink

Roy Henderson

• PSPS JHU sandbox testing:
  • Test of big merge - 36 full sized exposures. Successful
  • DXLayer crash when loading 'init' batch: debugged (was requiring min/max object ID)
  • Database table name issue in ODM (photoZRecipe): debugged with Suzanne at JHU
  • Found bug when loading 'init' batch. Issue with VARBINARY types in FITS->csv. Fixed.
  • Attempted to load some diffs. Ready on IPP side, but changes needed at JHU

• Maui cluster loading:
  • Working with Conrad to load Jim's simulation data on Maui cluster
  • Converted Jim's stack and Object sim data, but problem with detections. Fixed.
  • Problem submitting batches to Maui ODM. Samba mount issue?

• IPP-side development:
  • Finished off XML configuration work. Now, table descriptions, 'init' data, IPP->PSPS mappings etc all stored as XML for full flexibility

Paul Price

• Running STS background-preserved processing
• Cluster dramas
• Copied AF-approved opposition data for Ocarina.
• Discovered camera masks were being by psastro without error checking, leading to failures in warp because mask had zero size. Fixed psastro and added check for non-zero size in PS::IPP::Config->file_exists().
• Attempting to process Microtest suite. Discovered problems with automatic burntooling; reverted nightly_stacks.pro
• Debugged sudden background changes causing magic masking: bad cells (all pixels zero) weren't being recognised as such. Changed CELL.BAD from -100 to 0 and set MASK.BUILD in ppImage recipe to ensure this 'high' value doesn't damage residuals.
• Discovered and fixed bug in script that deletes diff convolved images
• Debugged assumptions on data volume
• Marked destreaked ThreePi.nightlyscience and all STS.nightlyscience warps for cleanup
• Debugged failing magic runs: many had missing convolved diff files, which have been recreated.
• Reworked PS1_DataRequestManagement page
• Using own astrometry with MD04 DVO database was fairly straightforward once I discovered the correct command lines and changed the psastro recipe. This work is not continuing because the excess clipping in the stack is caused by a different issue.
• Discovered and fixed half-pixel bug in pswarp producing bad positions in warp CMF files. Fixing this issue also results in improved stacks because the warp positions are used to produce the fake image that is the target for PSF matching.

Bill Sweeney

• debugged and fixed problem with magic masking.
• started working on fixing import of sdss data into dvo
• monitored dvo database creation. Discovered a problem causing dvo database corruption.
• fixed some problems with the postage stamp server caused by incorrect user input.
• hacked the distribution system to rebundle some data that was cleaned before ROE was able to download it
• various searches for data to clean up.

Chris Waters

• Observed at UKIRT.
• Added microtest handler into automation configuration.