Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-9610

Refactor Coadd Tasks: Add Pipe Task that makes all Coadds for DRP

    XMLWordPrintable

Details

    • 6
    • DRP S17-5, DRP S17-6, DRP F17-1
    • Data Release Production

    Description

      Artifact removal during co-addition requires working with both directWarps and psfMatchedWarps within a single task. Furthermore, makeCoaddTempExp.py creates both psfMatchedWarps and directWarps with one call to the command-line task to save computing time by avoiding warping twice, and assembleCoadd should have the same behavior and command-line configs.

      Currently, AssembleCoaddTask holds the "warpType" through the config parameter "doPsfMatch" in the config, which remains immutable for the lifetime of the task instance. This config controls whether directWarps or psfMatchedWarps are read from disk, anytime warps are needed. AssembleCoadd was only meant to make one warpType Coadd at a time. This is fine, but we need a supertask that can work with psfMatched, direct and likelihood Warps simultaneously.

      This ticket will implement a higher level task that uses the lower-level assembleCoadd as a subtask. Change and name will be RFC'ed.

      Must provide appropriate skeleton to enable both Safe-Clipped artifact removal and New artifact removal.

      Attachments

        Issue Links

          Activity

            Per discussion with yusra & jbosch of 2017-06-05, Yusra has done most of the thinking behind this ticket. However, she wants to have an RFD session to collect input from at least Paul & Russell before finishing up. We expect that to happen in parallel to work on DM-10005 (likely the discussion will bubble away in the background for a few days while she makes progress on that ticket).

            swinbank John Swinbank added a comment - Per discussion with yusra & jbosch of 2017-06-05, Yusra has done most of the thinking behind this ticket. However, she wants to have an RFD session to collect input from at least Paul & Russell before finishing up. We expect that to happen in parallel to work on DM-10005 (likely the discussion will bubble away in the background for a few days while she makes progress on that ticket).

            Per discussion of 2017-07-20, yusra will schedule a discussion, including at least price and jbosch, for the week starting 31 July to finalize the design of this work.

            swinbank John Swinbank added a comment - Per discussion of 2017-07-20, yusra will schedule a discussion, including at least price and jbosch , for the week starting 31 July to finalize the design of this work.

            The intent of this ticket diverged with the title, and new ticket DM-11446 is jumping ahead of it in the queue.

            yusra Yusra AlSayyad added a comment - The intent of this ticket diverged with the title, and new ticket DM-11446 is jumping ahead of it in the queue.

            We'll think about this during supertask conversion

            yusra Yusra AlSayyad added a comment - We'll think about this during supertask conversion
            yusra Yusra AlSayyad added a comment - - edited

            Found a workaround for CompareWarpAssembleCoadd (DM-8491) and with Gen3 it's now trivial to make many types of coadds in a single DRP.

            yusra Yusra AlSayyad added a comment - - edited Found a workaround for CompareWarpAssembleCoadd ( DM-8491 ) and with Gen3 it's now trivial to make many types of coadds in a single DRP.

            People

              yusra Yusra AlSayyad
              yusra Yusra AlSayyad
              Jim Bosch, Yusra AlSayyad
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Jenkins

                  No builds found.