Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-10504

Make more stringent cuts to input visits for HSC

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: pipe_tasks
    • Labels:
    • Team:
      Data Release Production

      Description

      In HSC we filter out visits that go into the coadd based on their PSF quality. However, we found that the current cuts do not reject enough bad CCDs so that we can reproduce the same PSF quality on the coadd that we had in the S16A data. This ticket is to adjust the cuts so that we can reproduce the same quality as before.

        Attachments

          Issue Links

            Activity

            Hide
            rearmstr Bob Armstrong added a comment -

            Here are the results for the correlation of the ellipticity residuals as a function of separation for the tract 9455 i-band:

            I found a new set of cuts using:

            • maxEllipResidual = 0.006
            • maxSizeScatter = 0.015

            That produces cuts equivalent to the S16A data:

            These new cuts remove double the amount of CCDs from the S16A i-band data ( from 4% to 8%).

            Show
            rearmstr Bob Armstrong added a comment - Here are the results for the correlation of the ellipticity residuals as a function of separation for the tract 9455 i-band: I found a new set of cuts using: maxEllipResidual = 0.006 maxSizeScatter = 0.015 That produces cuts equivalent to the S16A data: These new cuts remove double the amount of CCDs from the S16A i-band data ( from 4% to 8%).
            Hide
            rearmstr Bob Armstrong added a comment -

            Paul, can you give this a quick look?

            Show
            rearmstr Bob Armstrong added a comment - Paul, can you give this a quick look?
            Hide
            price Paul Price added a comment -

            Looks fine. Could you perhaps explain in the commit log the need for the change and how you chose the new parameters?

            Show
            price Paul Price added a comment - Looks fine. Could you perhaps explain in the commit log the need for the change and how you chose the new parameters?
            Hide
            rearmstr Bob Armstrong added a comment -

            Merged.

            Show
            rearmstr Bob Armstrong added a comment - Merged.
            Hide
            swinbank John Swinbank added a comment -

            It looks as though the changes here are motivated by HSC-specific requirements, but they've gone right into the default configs in pipe_tasks rather than an override in obs_subaru. Given that this code was only added to the LSST stack based on testing on HSC, I don't think that's really a problem, but I do note it here for the benefit of those doing JIRA archaeology to understand the changes in the future.

            Also, we should probably make issues like this clearer in our commit messages: while I think the changes here are appropriate, it seems a bit weird for LSST's commit log to be so clearly driven by HSC requirements, so we should take care to explain that.

            Show
            swinbank John Swinbank added a comment - It looks as though the changes here are motivated by HSC-specific requirements, but they've gone right into the default configs in pipe_tasks rather than an override in obs_subaru. Given that this code was only added to the LSST stack based on testing on HSC, I don't think that's really a problem, but I do note it here for the benefit of those doing JIRA archaeology to understand the changes in the future. Also, we should probably make issues like this clearer in our commit messages: while I think the changes here are appropriate, it seems a bit weird for LSST's commit log to be so clearly driven by HSC requirements, so we should take care to explain that.

              People

              • Assignee:
                rearmstr Bob Armstrong
                Reporter:
                rearmstr Bob Armstrong
                Reviewers:
                Paul Price
                Watchers:
                Bob Armstrong, John Swinbank, Paul Price
              • Votes:
                0 Vote for this issue
                Watchers:
                3 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: