Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-7446

Write tech note describing DCR algorithm and results

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: Science Pipelines
    • Labels:
      None
    • Story Points:
      10
    • Sprint:
      Alert Production S17 - 2, Alert Production S17 - 3, Alert Production S17 - 4, Alert Production S17 - 5, Alert Production S17 - 6, Alert Production F17 - 7, Alert Production F17 - 8, Alert Production F17 - 9, Alert Production F17 - 10
    • Team:
      Alert Production

      Description

      Once the DCR algorithm bakeoff is complete, a tech note should be written describing the approaches that were tried and the results. Highlights should include:

      • Edge cases found in DM-5699
      • In particular, an example using data with varying seeing
      • Plots of the DCR severity metric (DM-5920) to highlight which technique to use under different conditions.
      • Example spectra recovered from the images, compared with the input (simulated) spectra. An estimate on the accuracy of recovering spectra or at least differentiating between different types of objects (e.g. between stars and quasars) would be great.
      • If possible, provide an example using real data.

        Attachments

          Activity

          Hide
          swinbank John Swinbank added a comment -

          Hey Ian Sullivan — I'm a little concerned that this issue is infinitely extensible, as you can continue developing/refining/testing/documenting the DCR algorithm indefinitely. I'm wondering if we could focus this a little more tightly — say, document achievements as of a particular date — and then take it as read that this technote will continue to be expanded to describe improvements as they happen. What do you think?

          Show
          swinbank John Swinbank added a comment - Hey Ian Sullivan — I'm a little concerned that this issue is infinitely extensible, as you can continue developing/refining/testing/documenting the DCR algorithm indefinitely. I'm wondering if we could focus this a little more tightly — say, document achievements as of a particular date — and then take it as read that this technote will continue to be expanded to describe improvements as they happen. What do you think?
          Hide
          sullivan Ian Sullivan added a comment -

          I'm hoping to wrap this up very soon. I have only been waiting on the visualizations tools ticket, which I finally finished last month. I'll spend a little while editing the text and adding a few more figures, and then I'll send it out for review.

          Show
          sullivan Ian Sullivan added a comment - I'm hoping to wrap this up very soon. I have only been waiting on the visualizations tools ticket, which I finally finished last month. I'll spend a little while editing the text and adding a few more figures, and then I'll send it out for review.
          Hide
          sullivan Ian Sullivan added a comment -

          The current draft of the technote is at:
          https://dmtn-037.lsst.io/v/DM-7446/index.html

          Show
          sullivan Ian Sullivan added a comment - The current draft of the technote is at: https://dmtn-037.lsst.io/v/DM-7446/index.html
          Hide
          ebellm Eric Bellm added a comment - - edited

          I put a couple of questions in the review on github.

          Show
          ebellm Eric Bellm added a comment - - edited I put a couple of questions in the review on github.
          Hide
          ebellm Eric Bellm added a comment -

          Ian Sullivan I think the tech note is in good enough shape for now. Your section on the finite bandwidth effects is interesting. I think the choice of the number of sub-bands bears more attention, and I wonder if there's a way of casting this problem where we don't have to introduce that discretization. Fodder for thought.

          Show
          ebellm Eric Bellm added a comment - Ian Sullivan I think the tech note is in good enough shape for now. Your section on the finite bandwidth effects is interesting. I think the choice of the number of sub-bands bears more attention, and I wonder if there's a way of casting this problem where we don't have to introduce that discretization. Fodder for thought.
          Hide
          sullivan Ian Sullivan added a comment -

          There is still plenty more that can be added to this tech note, but it should now be ready for an initial distribution. Additional changes may be made later if there is significant feedback.

          Show
          sullivan Ian Sullivan added a comment - There is still plenty more that can be added to this tech note, but it should now be ready for an initial distribution. Additional changes may be made later if there is significant feedback.

            People

            • Assignee:
              sullivan Ian Sullivan
              Reporter:
              sullivan Ian Sullivan
              Reviewers:
              Eric Bellm
              Watchers:
              Eric Bellm, Ian Sullivan, John Swinbank
            • Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: