Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-23946

Change localWcs plugins to use a localGnomonicWcs transform.

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: meas_base, pipe_tasks
    • Labels:
      None
    • Story Points:
      8
    • Sprint:
      AP S20-4 (March), AP S20-5 (April)
    • Team:
      Alert Production
    • Urgent?:
      No

      Description

      Jim Bosch found that the current functor for localWcs would not return a valid transform when far from the Wcs center as it uses getCdMatrix/getTANWcs which behave poorly when far from the Wcs origin/do not utilize larger distortions. This ticket will change the plugins and functors to use linearizePixelToSky instead.

        Attachments

          Activity

          Hide
          cmorrison Chris Morrison added a comment -

          Hey Jim, implemented everything we talked about previously. The localGnomonic Wcs seems to work great. As we spoke about, there's a quirk in ast or something that causes the accuracy of the localWcs to drop when not using arcSeconds as the internal units. I've thus hard coded the function to use those units internally but return radians for convenience in later functor calculations.

          Considering the back and forth on this ticket, I'm wondering how to squash the commits. Should there be just one per ticket branch or can you see some the deserve keeping?

          Jenkins: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/31547/pipeline

          Show
          cmorrison Chris Morrison added a comment - Hey Jim, implemented everything we talked about previously. The localGnomonic Wcs seems to work great. As we spoke about, there's a quirk in ast or something that causes the accuracy of the localWcs to drop when not using arcSeconds as the internal units. I've thus hard coded the function to use those units internally but return radians for convenience in later functor calculations. Considering the back and forth on this ticket, I'm wondering how to squash the commits. Should there be just one per ticket branch or can you see some the deserve keeping? Jenkins: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/31547/pipeline
          Hide
          jbosch Jim Bosch added a comment -

          I think the only reason not to squash the commits would be to preserve the history of what approaches didn't work here, and I think the Jira comments do that better. Maybe squash them all together, and add a note in the meas_base commit message with a link to the Jira page and comment that alternatives were discussed there?

          Show
          jbosch Jim Bosch added a comment - I think the only reason not to squash the commits would be to preserve the history of what approaches didn't work here, and I think the Jira comments do that better. Maybe squash them all together, and add a note in the meas_base commit message with a link to the Jira page and comment that alternatives were discussed there?
          Hide
          cmorrison Chris Morrison added a comment -

          Sounds good to me. Doesn't look like things have changed do much. Did you take a look at the code/unittests? Are you satisfied there?

          Show
          cmorrison Chris Morrison added a comment - Sounds good to me. Doesn't look like things have changed do much. Did you take a look at the code/unittests? Are you satisfied there?
          Hide
          jbosch Jim Bosch added a comment -

          Yup, all looks good (just couldn't hit "Approve" on the PRs because apparently I was the one who created them).

          Show
          jbosch Jim Bosch added a comment - Yup, all looks good (just couldn't hit "Approve" on the PRs because apparently I was the one who created them).
          Hide
          cmorrison Chris Morrison added a comment -

          Added link to Jira ticket in commits.

          Show
          cmorrison Chris Morrison added a comment - Added link to Jira ticket in commits.

            People

            • Assignee:
              cmorrison Chris Morrison
              Reporter:
              cmorrison Chris Morrison
              Reviewers:
              Jim Bosch
              Watchers:
              Chris Morrison, Eric Bellm, Jim Bosch, John Swinbank
            • Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: