Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-38948

Fix dataset query constraint bugs introduced on DM-38780

    XMLWordPrintable

    Details

      Description

      While starting DM-38498, I noticed that some logic involving the "dataset query constraint" option to QG generation should have been updated on DM-38780, but hadn't been (which is largely my fault, as the reviewer of that ticket just earlier the same day).  I had originally planned to put a fix on DM-38498 itself, but the original goal of that ticket now looks harder than it once did so I'm putting the fix here.

        Attachments

          Issue Links

            Activity

            Hide
            jbosch Jim Bosch added a comment -

            Possible fix is on the branch.  Given the difficulty of adding unit tests on short notice (needs moderately complex data and tasks), I'm going to run some one-off tests this weekend against the USDF repos before putting it out for review.  I also want to fix the test coverage but don't want to get involved in that while the bug is on main.

             

            Show
            jbosch Jim Bosch added a comment - Possible fix is on the branch.  Given the difficulty of adding unit tests on short notice (needs moderately complex data and tasks), I'm going to run some one-off tests this weekend against the USDF repos before putting it out for review.  I also want to fix the test coverage but don't want to get involved in that while the bug is on main.  
            Hide
            jbosch Jim Bosch added a comment -

            One-off test was successful; the command I used to reproduce the problem is:

            $ pipetask qgraph -b /repo/main -i HSC/runs/RC2/w_2023_11/DM-38360 -p /sdf/group/rubin/sw/conda/envs/lsst-scipipe-6.0.0/share/eups/Linux64/drp_pipe/g10915be422+84f8699254/pipelines/HSC/DRP-RC2.yaml#step2cde -o u/jbosch/DM-38948 

            Before this ticket this results in an exception.  After, I get:

            lsst.ctrl.mpexec.cmdLineFwk INFO: QuantumGraph contains 869 quanta for 7 tasks, graph ID: '1682731670.6923134-3810612'
            Quanta           Tasks           
            ------ --------------------------
               432   finalizeCharacterization
                 1 fgcmBuildFromIsolatedStars
                 1               fgcmFitCycle
                 1         fgcmOutputProducts
               432         updateVisitSummary
                 1             makeVisitTable
                 1          makeCcdVisitTable 

            which is the same as I get with w_2023_17, before the bug was introduced.

            Show
            jbosch Jim Bosch added a comment - One-off test was successful; the command I used to reproduce the problem is: $ pipetask qgraph -b /repo/main -i HSC/runs/RC2/w_2023_11/DM- 38360 -p /sdf/group/rubin/sw/conda/envs/lsst-scipipe- 6.0 . 0 /share/eups/Linux64/drp_pipe/g10915be422+84f8699254/pipelines/HSC/DRP-RC2.yaml#step2cde -o u/jbosch/DM- 38948 Before this ticket this results in an exception.  After, I get: lsst.ctrl.mpexec.cmdLineFwk INFO: QuantumGraph contains 869 quanta for 7 tasks, graph ID: '1682731670.6923134-3810612' Quanta Tasks ------ -------------------------- 432 finalizeCharacterization 1 fgcmBuildFromIsolatedStars 1 fgcmFitCycle 1 fgcmOutputProducts 432 updateVisitSummary 1 makeVisitTable 1 makeCcdVisitTable which is the same as I get with w_2023_17, before the bug was introduced.
            Hide
            jbosch Jim Bosch added a comment -

            One-off tests were a bit easier than I thought and Jenkins is green, so it's time to put this in review.

            Only PR is https://github.com/lsst/pipe_base/pull/324

            I have some thoughts about how to improve test coverage that I'll put on another ticket.

            Show
            jbosch Jim Bosch added a comment - One-off tests were a bit easier than I thought and Jenkins is green, so it's time to put this in review. Only PR is https://github.com/lsst/pipe_base/pull/324 I have some thoughts about how to improve test coverage that I'll put on another ticket.
            Hide
            salnikov Andy Salnikov added a comment -

            Looks good, thanks for fixing my mess.

            Show
            salnikov Andy Salnikov added a comment - Looks good, thanks for fixing my mess.

              People

              Assignee:
              jbosch Jim Bosch
              Reporter:
              jbosch Jim Bosch
              Reviewers:
              Andy Salnikov
              Watchers:
              Andy Salnikov, Jim Bosch
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Dates

                Created:
                Updated:
                Resolved:

                  Jenkins

                  No builds found.