Verification Datasets Meeting 2015-12-16 Minutes

Data Management
#1

Verification Datasets Meeting 2015-12-16 Minutes
Attendees: David, Angelo, Hsin-Fang, Domonique, Simon, Gruendl
Regrets: Ciardi, Francisco, Colin
Notetaker: David

Angelo

  • working on visualizing the stack processing with firefly
  • having some issues selecting the images
  • no progress on bulge processing since focusing on QA visualization

David

  • rerunning procesCcdDecam on COSMOS data, still running now
  • will look for failures due to initial PSF FWHM being set too high (e.g. too few PSF stars).
  • still need a program to tell me the unique patches and visits to run makeCoaddTempExp on,
    having difficulty getting Lauren’s code running, butler.get failure
  • Hsin-Fang has Lauren’s showVisitSkymap script working will send and figure out how to get it into stack

Dominique

  • good progress on CFHT processing

  • u-band coadd problem, it’s in the tan-sip fitter, lower S/N in u-band and higher background
    maybe limit tan-sip polynomial order to 3, will open ticket on this

  • running simultaneous astrometry fitting and updating WCS before coaddition, it’s working
    improvement, stars in coadds are narrower, see plots

  • astrometric scatter was 25 mas for <19 mag, after simultaneous fitting it went down to 9.5 mas

  • sip convention requires inverse polynomials, still failing sometimes

  • code has been tested on HSC and is working

  • Simon said that it’s being modified to fit better into the stack, will try to get it done by the end of this cycle

  • David: what forced photometry code to use forcedPhotCcd.py or forcedPhotCoadd.py?

  • Simon uses forcedPhotCcd.py, runs forced photometry on calexps using coadd measurements
    measure on coadds, merge the catalogs, forced photometry on single-epoch images
    Simon’s “tutorial” on running LSST simulated data through DRP

  • David: how best to add usage information into a command line task, will ask Russell

Hsin-Fang

  • running raw Stripe82 DECam data through DRP, one patch single-band, no obvious failures but some warnings, something wrong with WCS, doesn’t quite trust results yet
  • modifications to obs_decam config policy file, will try to merge soon

Simon

  • Frank Massey has differencing program that does an order of magnitude better than hotpants, Simon will talk with him about it

  • Joachim has compiled day MOPS now, can now distribute it with the stack

  • Russell breaking processCcd.py into three command line tasks: ISR (e.g. ccdproc), psf determination (e.g. PSFEx), detection/measurement (e.g. SExtractor)
    will be easier to document these, there will still be a processCcd command line task that will run these three

  • will think about command line task documentation priorities

  • H-F: some DES folks would like to contribute their code to the LSST stack, satellite trail masking code

  • Simon, process not well fleshed out, working with DESC group, try to use the stack to implement their algorithms but keep it in separate repositories
    eventually if theirs is much better than ours then we will adopt it, maybe not the best process

  • it’s in the plan to develop this kind of masking code, but not developed yet

Colin, not on
Francisco/Yusra, not on
Michael Wood-Vasey, not on

Will try to have call next week

#2

Excellent. Is this the version that uses PAL?

If the DESC code is a standalone package that’s obviously easier to manage regarding license and copyright. If they are wanting to modify a lot of our code and use pull requests that could also work but we’d really have to be certain about our license and copyright policy before we wanted to turn that on.