DM-29344: Investigate the CI differences between Gen 2 and 3 in COSMOS field (continued) #13
Conversation
| @@ -1,7 +1,6 @@ | |||
| # Config override for lsst.ip.isr.IsrTask | |||
|
|
|||
| # Can't ingest the necessary datasets | |||
| config.doBrighterFatter = False | |||
| config.doAttachTransmissionCurve = False | |||
| config.doUseOpticsTransmission = False | |||
| config.doUseFilterTransmission = False | |||
There was a problem hiding this comment.
Given the comment above (and the assumption that removing the line doesn't break things), I wonder how many of the other lines are also unneccessary. It looks like we weren't correctly handling curated calibs when this file was written (i.e., before lsst/ap_verify#91).
Edit: the underlying problem is that the way obs_subaru handled these datasets in Gen 2 was very non-standard (because there was no standard), and ap_verify can't handle the necessary datasets without special-casing.
There was a problem hiding this comment.
EDIT: the Gen 2/3 discrepancy is actually a bug in ap_verify; it hard-codes the files it can load in Gen 3, and doesn't load isr.py. It should be resolved by DM-26140.
|
After discussion with Krzysztof, we don't actually want to turn this back on here, because it won't work. Going to close this PR and open one for ap_verify instead. |
This PR removes the line in
config/isr.pywhich explicitly turned off the brighter-fatter correction (in gen2) for this dataset. The goal is to match the gen3 results, which currently has brighter-fatter on by default.