Child pages
  • 22nd April 2015
Skip to end of metadata
Go to start of metadata

Attending: Scott Daniel, Chris, Kem, Rahul, Michael, Peter, Cathy, Lynne, ajc, Bryce, Darko, 

 

Quick items before Lynne's presentation:

  • DDS rate issues - the current limitation is due to the SAL (goal is to expose improved functionality of DDS by 4th quarter this year). This is a concern as we are basing our design on a given performance. Need to know DDS can perform with a 1-5ms latency or better. Should we strip away SAL?
  • Stash - Michael will move repos to github and leave problematic PRs till later. All PR must be merged by Friday.

Lynne's notebook demo

  • Note - you need to start the notebook in a terminal with MAF etc set up
  • Comments
    • Should we start with a notebook on access to opsim DB? Lynne might mean people won't use MAF and just resort to opsim queries
    • We need to have the opsim DBs available for the workshop (Tier 1)
    • Should add a demo of how to extract the data that is plotted (e.g. the airmass as a function of healpix pixel)
    • Can we zoom in on healpix regions. Yes - should put that example in a notebook
    • Include an example of how to slice in time, create a movie?
    • How do I find the simple metrics - metric has function metrics.list() that will provide a list (there are about 30 of them) or metrics.list(docs=True)
    • Stackers arent in current example
  • For the workshop the notebooks are design documents that we want to keep (almost like pseudo code)
    • What notebooks will we need:
      • how to split data into years
      • how to interact with the metric data (not just plot or summarize)
      • healpix functionality and changing the plots
      • using contributed metrics
      • writing your own metric
      • how to use stackers
      • notebook for each of the science plots
        • show depth and then per year then as a function of time
        • number of colors as a fn of year
        • metric values in a separate notebook: notebook on these projections etc

 

  • No labels