AP Pipeline QA Meeting, 2024-01-29

2pm Pacific Time

Zoom: https://washington.zoom.us/j/98846655020

during-meeting chat, please use #dm-alert-prod

attending:

• Eric Bellm
• Nima Sedaghat 
• Meredith Rawls 
• Hsin-Fang Chiang 
• Ian Sullivan 
• Brianna Smart 
• Erin Howard 
• Markus Rabus 
• Keith Bechtol 
• Krzysztof Suberlak 
• Ari Heinze 
• Krzysztof Findeisen 
• Jeffrey Carlin 
• Bob Armstrong 
• Kenneth Herner 
• Bryce Kalmbach 
• Rick Kessler
• Colin Slater 
• Zeljko Ivezic 
• Harshit Rai 
• Bruno Sanchez 

regrets:

• Michael Wood-Vasey 

Topics for discussion:

• Announcements
  • Mirror cell off the telescope, met requirements for forces at 100% speed/acceleration/jerk.  Glass to coating chamber....
• AuxTel Processing (all):
  • Prompt Processing with AuxTel status (Hsin-Fang Chiang , Krzysztof Findeisen , Eric Bellm , Neven Caplar ):
    • returning to on-sky tonight
  • 10am processing status (Colin Slater , Meredith Rawls ):
    • will run another round at the end of the week
  • updates to survey (Keith Bechtol ,Colin Slater ?):
    • no updates...  no update to scheduler logic.  will discuss further in Tuesday meeting tomorrow
    • some discussion of what fields are being observed, what templates are available
      • Hsin-Fang Chiang will add updated 10am collection from December so we can perhaps get some subtractions.  We'd prefer a more formal process in the future!
• Precursor processing (AP team):
  • Summary of diffim pipeline changes (Ian Sullivan ):
    • https://docs.google.com/presentation/d/16nidoAkCqejku47gfU-rYCMflztBgxgQoj2Orp_eDp4/edit?usp=sharing 
    • various bugfixes
    • new flags to reject early sources with bad 3x3 central pixels
    • Rick K asks if we plan (or should consider) using fixed kernel selection candidates–not until after ~DR2
    • Rick asking about how to interpret/characterize use of flags for science use–interpretation was challenging in DPs
  • Streak Masking status (Meredith Rawls ):
    • new streak masking running in single frame–looks for linear features in the DETECTED mask plane
    • currently it's over-zealous, adding some metrics (in DRP processing)–counts DIASources overlapping streak footprints or centers.  
    • new plotting utility in analysis_tools to enable overplotting different (DIA)source types spatially
    • showing some LATISS "worst case" examples
    • Zeljko: should we be doing the streak finding on the difference images instead?  MR: thinks using the science
• SITCOM team updates (Andrew Connolly, Mario Juric , Bryce Kalmbach , Krzysztof Suberlak , Neven Caplar  ):
  • Bryce: starting to think about commissioning tests & analysis notebooks to be ready for Comcam
• DIA Transients & Variables Science Unit update (Eric Bellm ; #rubinobs-sciunit-transient-variable):
  • Reorganization into technical teams for the next six months:
    • Real/Bogus (Bellm, Sedaghat, Rai, Bianco, Acero Cuellar) 
      • Prepare a model to go on-sky with ComCam
    • Source Injection (Smith, Sánchez, Ryczanowski, Narayan(?), Knop, Kelvin)
      • Complete injection framework and prepare appropriate input catalogs
        
    • DIA Metrics (Wood-Vasey, Caplar, Sánchez)
      • Implement metrics for completeness & purity, astrometric & photometric precision
    • DIA Algorithms (Sullivan, Sánchez, Liu, Wood-Vasey, …)
      • Prepare algorithmic configuration to go on-sky with ComCam
• DIA Solar System Science Unit update (Mario Juric ; #sitcom-solsys):
  • out today
• PreOps V&V (Colin Slater ): 
  • nothing AP related–working on metrics and upcoming ops rehearsal
  • Keith summarizing ops rehearsal: 21-night window has been simulated, including the 3-day ops rehearsal.  Planning to make templates with the earlier images.  Image simulations are higher-fidelity than DC2: includes better treatment atmosphere, optics, use of test data.  wider range of stellar density.  Observations are a series of repeated observations with transnational and rotational dithers, range of temporal sampling.  Stretch goal to include solar system objects.  So far on pace, first images being generated now.
    • Hsin-Fang: can we make sure next_visit messages are correct?  Colin: Tiago working on this.  needs a simulated Comcam scheduler
    • YA: prediction for image availability?  Keith: a few this week, rest in a week or two
• Review CI (https://usdf-rsp-dev.slac.stanford.edu/chronograf/):

• Next meetings (all at 2pm Pacific):
  
  • March 4
  • April 1 (pending ops rehearsal)
• AOB
  • none

Action Items