Details
-
Type:
Bug
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: ap_verify
-
Labels:None
-
Story Points:1
-
Epic Link:
-
Sprint:AP S23-6 (May)
-
Team:Alert Production
-
Urgent?:No
Description
The dc2 dataset failed with:
lsst.daf.butler.registry._exceptions.MissingDatasetTypeError: "Registry is missing an input dataset type DatasetType('fakes_goodSeeingRealBogusSources', {band, instrument, detector, physical_filter, visit}, Catalog)"
|
Attachments
Issue Links
- relates to
-
DM-36743 Add the real-bogus PipelineTask to ap_verify for the DC2 CI dataset
- Done
Activity
Pushed a fix.
Looks good. GitHub seems to be stuck on the linting checks, which might be due to GitHub problems today, but could also need to be unstuck with a new push. You can try to unstick the check by amending the previous commit message, and pushing again.
New push didn't help the situation 
Ian Sullivan, "new push" was needed back in the Travis days. With GitHub Actions, you can always force a check to rerun from its Details page.
Krzysztof Findeisen
added a comment - - edited Ian Sullivan , "new push" was needed back in the Travis days. With GitHub Actions, you can always force a check to rerun from its Details page. I take it back – some of these checks are so corrupted, they can't be rerun.
Krzysztof Findeisen
added a comment - I take it back – some of these checks are so corrupted, they can't be rerun.
ap_verify always uses --register-dataset-types, so it's a bit strange that it would get this error. Possibly the pipeline graph isn't connected properly (i.e., it's an input but not an output)?