wiki:DC3APostMortemMeetingMOPS
Last modified 11 years ago Last modified on 05/20/2009 04:15:53 PM

Integrating MOPS into the LSST Framework

This is Where We Are

  • Transition Plan (see attached document).
  • Current Status.
  • Experience in integrating with the LSST pipeline framework.
  • Experience in integrating with the LSST persistence framework.
    • How much throw-away code do we need to write?
  • Experience in integrating with the LSST policy framework.

Where Do We Want to Go?

  • Sky tessellation
  • ORM Layer
    • See integrating with the LSST persistence framework.
  • Standard units for policy files
  • Use of positional astronomy library
  • Re-implementation of Auton Kd-Tree code?
  • Switch to statistical orbit determination (i.e. OpenOrb)?
  • Requirements for PEX_HARNESS (e.g. updating clipboard for slices in preprocess)?
  • Will we run DayMOPS every day?
  • What about the run mode for data release?
  • Can we run on a one-month worth of data? One night? One image?
  • Do we have to compute values for all columns in the DB tables (e.g. MovingObject)?
  • Who tracks the ground truth? How?
  • When will we know which input data we'll use?
  • Can we leverage MPC work on re-implementing the JPL code?
  • Cadence?

The Future?

  • Timeline for integration
  • Requirements for DayMOPS?
    • Input data?
    • Output data?
      • Which parameters for
        • MovingObjects
        • Tracklets
      • MOID?
    • User-level tools?
    • Efficiency/accuracy levels?
    • How many nights to determine an orbit?
    • 5-sigma vs 3-sigma DiaSources?
    • Multiple orbits per MovingObject?
    • Provenance?
    • Are we going to simulate the telescope acquiring data during the night?
      • Are we going to be using !DiaSourceIDForTonight?
    • What does SDQA mean for MOPS?

Science Users

  • Which database table do you want to see?

Results/Decisions?

Attachments