Uploaded image for project: 'Request For Comments'
  1. Request For Comments
  2. RFC-882

Remove ap_verify_hits2015 dataset

    XMLWordPrintable

    Details

    • Type: RFC
    • Status: Adopted
    • Resolution: Unresolved
    • Component/s: DM
    • Labels:
      None

      Description

      The ap_verify framework was originally designed to handle both nightly CI runs and occasional larger runs. We created one dataset, ap_verify_hits2015, to support the latter case. However, developments since then (in particular, the introduction of BPS) have shifted ap_verify's purpose to CI alone. The AP team has not performed an ap_verify run with ap_verify_hits2015 in over two years, but we still try to keep it in sync with its more frequently used CI subset.

      We would like to retire the ap_verify_hits2015 package from active support. Since there may be third parties still using the dataset, we propose doing this through the standard deprecation procedure (i.e., waiting for one major version before removal), and moving the package into lsst-dm/legacy- once removed. Since ap_verify_hits2015 is not a code package and is never imported, we'll have to find some other way of getting the deprecation warning out (possibly injecting the warning into the dataset-specific configs?).

      This proposal does not affect the two CI packages, ap_verify_ci_hits2015 and ap_verify_ci_cosmos_pdr2.

        Attachments

          Issue Links

            Activity

            Hide
            krzys Krzysztof Findeisen added a comment -

            Escalating as a deprecation request (albeit an unusual one).

            Show
            krzys Krzysztof Findeisen added a comment - Escalating as a deprecation request (albeit an unusual one).
            Hide
            Parejkoj John Parejko added a comment -

            As possibly the only person who has done an ap_verify run on this dataset in the past two years, I support this RFC. Pretty much each time I've tried to run it, I've run into one issue or another that needed debugging (because it's not run regularly), and the marginal benefit it gives is not worth the bother.

            Show
            Parejkoj John Parejko added a comment - As possibly the only person who has done an ap_verify run on this dataset in the past two years, I support this RFC. Pretty much each time I've tried to run it, I've run into one issue or another that needed debugging (because it's not run regularly), and the marginal benefit it gives is not worth the bother.
            Hide
            ebellm Eric Bellm added a comment -

            Very happy to retire this, as suggested.

            Show
            ebellm Eric Bellm added a comment - Very happy to retire this, as suggested.
            Hide
            mrawls Meredith Rawls added a comment -

            Happy to support this RFC.

            Show
            mrawls Meredith Rawls added a comment - Happy to support this RFC.

              People

              Assignee:
              krzys Krzysztof Findeisen
              Reporter:
              krzys Krzysztof Findeisen
              Watchers:
              Eric Bellm, Ian Sullivan, Jim Bosch, John Parejko, Krzysztof Findeisen, Meredith Rawls, Tim Jenness, Yusra AlSayyad
              Votes:
              0 Vote for this issue
              Watchers:
              8 Start watching this issue

                Dates

                Created:
                Updated:
                Planned End:

                  Jenkins

                  No builds found.