Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-9159

Catch-and-ignore should not apply to any exception

    XMLWordPrintable

    Details

      Description

      The code at https://github.com/lsst/pipe_tasks/blob/master/python/lsst/pipe/tasks/assembleCoadd.py#L584 tries to force coadd creation even in the presence of problems that are not related to the input data (e.g., AttributeErrors or TypeErrors from code bugs). In particular, it does not guarantee that the resulting coadd contains any valid data, leading to hard-to-trace failures elsewhere in the code.

      I request that this code be rewritten, either to only catch exceptions that are specifically known to be ignorable, or to deal with missing subregions in some more robust way than simply ignoring them. I assume the former approach is more appropriate...?

        Attachments

          Issue Links

            Activity

            Hide
            Parejkoj John Parejko added a comment -

            pipe_tasks in general seems to be littered with these, and they don't even obey doRaise, in general.

            Show
            Parejkoj John Parejko added a comment - pipe_tasks in general seems to be littered with these, and they don't even obey doRaise , in general.
            Hide
            jbosch Jim Bosch added a comment -

            Processing this in a triage sweep - it's worth keeping, as there are still except Exception cases in assembleCoadd.py, and those are definitely worth fixing (note that the original line numbers are no longer valid). But I do not want this ticket to become an umbrella for all such cases in pipe_tasks, because that'll make it too big to ever close (and possibly too big for anyone to try starting).

            Show
            jbosch Jim Bosch added a comment - Processing this in a triage sweep - it's worth keeping, as there are still except Exception cases in assembleCoadd.py, and those are definitely worth fixing (note that the original line numbers are no longer valid). But I do not want this ticket to become an umbrella for all such cases in pipe_tasks , because that'll make it too big to ever close (and possibly too big for anyone to try starting).

              People

              Assignee:
              Unassigned Unassigned
              Reporter:
              krzys Krzysztof Findeisen
              Watchers:
              Jim Bosch, John Parejko, Yusra AlSayyad
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Dates

                Created:
                Updated:

                  Jenkins

                  No builds found.