Present: Lauren MacArthur, Unknown User (tmorton), Sophie Reed, Jim Bosch, Robert Lupton, Paul Price, John Swinbank.

  • Over the next two months, we (or rather, all of us that available on any given week) will convene weekly at 3pm Eastern on Tuesdays to discuss progress on QA for the November LSST milestone & next HSC data release.
  • We agreed to produce a list of QA metrics with brief descriptions which we will aim to measure on the next data release.
    • These metrics will be binned over the survey area.
    • This list will be a "living document": we will seed it with existing LSST KPMs (that's the last time I'll use that term in this context), the stellar locus metric, and any other metrics that Lauren is planning; after that, curation and prioritization will be up to Jim and Robert.
    • Let's regard (the initial version of) that list as the deliverable of DM-11312.
    • John will create a Confluence page for this document.
  • For next meeting (Tuesday 3 October), Lauren will focus on:
    • Completing her work with DESC (DM-11452);
    • Developing a new HSC RC dataset (DM-11345).
    • That will free her up to focus on developing QA metrics over the following ~2 months.
  • For the next meeting, Tim will focus on:
    • Loading the results of the previous HSC PDR reprocessing into his QA system and checking that there are no gotchas. Let's regard this as completing DM-10859.
    • Extending his QA system to cover spatially binned metrics. I don't think we have a ticket for this: Tim, if I'm wrong please correct me; if I'm right, please create one.
  • For the next meeting, Tim & Lauren will work together to ensure that the pipe_analysis scripts efficiently & effectively dump all the information that Tim's plotting scripts need to Parquet files without wasting time generating redundant plots or stripping out useful data. I have ticketed this as DM-12038 and assigned it to Tim. Please check and fix that ticket if necessary.
  • For future meetings, we will focus on developing metrics as described on the Confluence page, with rapid feedback from Robert (& Jim) on what works and what doesn't.
  • I'll (John Swinbank) be out for the meeting next week and (probably) the week after.  Please go ahead meet without me. Bonus points if anybody takes notes.

  • No labels