Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: obs_decam
-
Labels:None
-
Story Points:2
-
Epic Link:
-
Sprint:AP S20-4 (March)
-
Team:Alert Production
-
Urgent?:No
Description
The decam convert tests in DM-22655 rely on a gen2 repo that does not exist in our CI system. To make those tests able to be run as part of CI, we need to shrink the calibration data that lives in ap_verify_ci_hits2015, and make a gen2 repo with it. If the data is small enough, that data could even live in obs_decam.
Tim Jenness's suggestion is to not zero all the data, but to set each data array to the HDU number, so that we can still test that gen3 and gen2 read the same pixel data.
Attachments
Issue Links
- is triggered by
-
- Done
Activity
The new test repo looks good. I think you need to fix the repositoryCfg.yaml files.
After some struggles with compressed HDUs, I've now got obs_decam using this data, so it should run as part of regular CI.
Jenkins run: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/31355/pipeline
Tim Jenness: do you mind doing this small review? The +1000 lines in testdata_decam is because of all the new data files there. I did update that readme, so at least look at that.
I think it's good enough that we have files in testdata_decam (instead of in obs_decam itself): the bias/flat files are float32, so they compress pretty well when set to all one value, but int data would do better. I'd rather leave the data in the same format as the original files to keep things consistent (even if the values are now nonsense).
Looks great. Thanks.
It's possible that
DM-21849might not merge until Jim Bosch comes back online after the workshop so it's probably best not to wait on that. Master should be fine and it would be good to get Jenkins testing this – I only had to make a couple of tweaks on obs_decam to get it to work withDM-21849.