Details

    • Type: Story
    • Status: Done
    • Priority: Major
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None
    • Templates:
    • Story Points:
      3
    • Sprint:
      DRP S17-1, DRP S17-2
    • Team:
      Data Release Production

      Description

      In preparation for a new multi-color deblender the API for lsst.meas.deblender.baseline.deblend needs to be modified to (optionally) accept lists for the footprint, maskedImage, psf, and psffwhm parameters.

      The PerParent and PerPeak classes will also be modified to organize parent and peak objects in different colors, and the underlying code will be modified as needed to use the new objects.

      When this ticket is completed there should be no difference in the behavior of lsst.meas.deblender.baseline.deblend from using the current deblender API (other than the new/modified classes), but this will be the last ticket before a long-lived branch is created to store work on the new deblender.

        Activity

        Hide
        fred3m Fred Moolekamp added a comment -

        I think there are two different changes we are discussing. The first is modifying pipe_tasks so that instead of running meas.deblender.baseline.deblend on each filter in a loop, it calls meas.deblender.baseline.deblend once with the parameters from all of the filters once. That is what I planned to add to this ticket, although I am fine with delaying it until DM-8621 if you prefer.

        The separate issue is refactoring the pipe.tasks that will run differently when the multi-color deblender is working. Until the multi-color deblender is working this can't really be done, because until we have a functioning multi-color deblender we still need to merge the footprints and perform forced photometry on them.

        Show
        fred3m Fred Moolekamp added a comment - I think there are two different changes we are discussing. The first is modifying pipe_tasks so that instead of running meas.deblender.baseline.deblend on each filter in a loop, it calls meas.deblender.baseline.deblend once with the parameters from all of the filters once. That is what I planned to add to this ticket, although I am fine with delaying it until DM-8621 if you prefer. The separate issue is refactoring the pipe.tasks that will run differently when the multi-color deblender is working. Until the multi-color deblender is working this can't really be done, because until we have a functioning multi-color deblender we still need to merge the footprints and perform forced photometry on them.
        Hide
        rearmstr Bob Armstrong added a comment -

        > I think there are two different changes we are discussing. The first is modifying pipe_tasks so that instead of running meas.deblender.baseline.deblend on each filter in a loop, it calls meas.deblender.baseline.deblend once with the parameters from all of the filters once. That is what I planned to add to this ticket, although I am fine with delaying it until DM-8621 if you prefer.

        I think I'm a little confused. The looping over the different filters happens implicitly in pipe_drivers. The only command line tasks in multiband.py that work on multiple filters is the footprint merging and the merge measurements. So I think you can't only modify pipe_tasks to do this. I think it makes sense to put this into DM-8621.

        Show
        rearmstr Bob Armstrong added a comment - > I think there are two different changes we are discussing. The first is modifying pipe_tasks so that instead of running meas.deblender.baseline.deblend on each filter in a loop, it calls meas.deblender.baseline.deblend once with the parameters from all of the filters once. That is what I planned to add to this ticket, although I am fine with delaying it until DM-8621 if you prefer. I think I'm a little confused. The looping over the different filters happens implicitly in pipe_drivers. The only command line tasks in multiband.py that work on multiple filters is the footprint merging and the merge measurements. So I think you can't only modify pipe_tasks to do this. I think it makes sense to put this into DM-8621 .
        Hide
        fred3m Fred Moolekamp added a comment -

        Ok Bob Armstrong, in that case I guess this ticket is finished after all.

        Show
        fred3m Fred Moolekamp added a comment - Ok Bob Armstrong , in that case I guess this ticket is finished after all.
        Hide
        fred3m Fred Moolekamp added a comment -

        Changes made based on the review and py 2,3 tests pass in Jenkins.

        Show
        fred3m Fred Moolekamp added a comment - Changes made based on the review and py 2,3 tests pass in Jenkins.
        Hide
        fred3m Fred Moolekamp added a comment -

        Removed oldBaseline.py, which was included as a way to verify that the old and new API generated deblender results that were bitwise the same and re-ran on Jenkins py2,3.

        Show
        fred3m Fred Moolekamp added a comment - Removed oldBaseline.py , which was included as a way to verify that the old and new API generated deblender results that were bitwise the same and re-ran on Jenkins py2,3.

          People

          • Assignee:
            fred3m Fred Moolekamp
            Reporter:
            fred3m Fred Moolekamp
            Reviewers:
            Bob Armstrong
            Watchers:
            Bob Armstrong, Fred Moolekamp, Jim Bosch
          • Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development

                Agile