Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-5109

Offset in gaussian-psf in ci_hsc

    Details

    • Type: Bug
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: ci_hsc
    • Labels:
      None

      Description

      I'm seeing what looks like an aperture correction problem in psf-gaussian on ci_hsc coadds. This gets in the way of our ability to do star/galaxy classification, and suggests potentially more serious problems elsewhere.

        Attachments

          Issue Links

            Activity

            Hide
            lauren Lauren MacArthur added a comment -

            Jim, are you sure there isn't a problem? I was actually thinking you might need to add the multi-step I implemented in processImage.py (see https://github.com/lsst/pipe_tasks/commit/d904e3d188698b4f57bf3dad1516b0bf201078f5) to the measurement run in multiband.py at:
            https://github.com/lsst/pipe_tasks/blob/master/python/lsst/pipe/tasks/multiBand.py#L539 (i.e. as is, any algorithm with execution order > APCORR_ORDER in the algorithm list, will be executed prior to aperture correction).

            Show
            lauren Lauren MacArthur added a comment - Jim, are you sure there isn't a problem? I was actually thinking you might need to add the multi-step I implemented in processImage.py (see https://github.com/lsst/pipe_tasks/commit/d904e3d188698b4f57bf3dad1516b0bf201078f5 ) to the measurement run in multiband.py at: https://github.com/lsst/pipe_tasks/blob/master/python/lsst/pipe/tasks/multiBand.py#L539 (i.e. as is, any algorithm with execution order > APCORR_ORDER in the algorithm list, will be executed prior to aperture correction).
            Hide
            jbosch Jim Bosch added a comment -

            Hmm, you may be right. Thanks for being persistent. I hadn't realized the workaround for the bad logic involved changes to ProcessImageTask. Taking this out of review while I investigate.

            Show
            jbosch Jim Bosch added a comment - Hmm, you may be right. Thanks for being persistent. I hadn't realized the workaround for the bad logic involved changes to ProcessImageTask . Taking this out of review while I investigate.
            Hide
            jbosch Jim Bosch added a comment -

            Nate Lust, back to in review for you. There are now branches for meas_base and pipe_tasks, in addition to the ci_hsc change noted above.

            Show
            jbosch Jim Bosch added a comment - Nate Lust , back to in review for you. There are now branches for meas_base and pipe_tasks, in addition to the ci_hsc change noted above.
            Hide
            nlust Nate Lust added a comment -

            Provided the things like the workaround for aperture correction are just a temporary workaround while a better solution is researched I am fine with this. The true solution to that is beyond the scope of this ticket, but I lament the need for this in general, and especially seeing the code duplicated all over the place. This looks fine to merge

            Show
            nlust Nate Lust added a comment - Provided the things like the workaround for aperture correction are just a temporary workaround while a better solution is researched I am fine with this. The true solution to that is beyond the scope of this ticket, but I lament the need for this in general, and especially seeing the code duplicated all over the place. This looks fine to merge
            Hide
            jbosch Jim Bosch added a comment -

            Merged to master.

            These changes are indeed workarounds, and I hope to eventually improve matters with the approach described in RFC-143.

            Show
            jbosch Jim Bosch added a comment - Merged to master. These changes are indeed workarounds, and I hope to eventually improve matters with the approach described in RFC-143 .

              People

              • Assignee:
                jbosch Jim Bosch
                Reporter:
                jbosch Jim Bosch
                Reviewers:
                Nate Lust
                Watchers:
                Jim Bosch, Lauren MacArthur, Nate Lust
              • Votes:
                0 Vote for this issue
                Watchers:
                3 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved:

                  Summary Panel