Details
-
Type:
Story
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: obs_subaru, pipe_drivers, pipe_tasks
-
Labels:
-
Story Points:1
-
Team:External
Description
I think we agreed at the 2018-12-11 Monday Meeting that it's better to included interpolated CR pixels in the coadd (since the interpolation is pretty good) rather than leave them out and get lots more pixels masked with INEXACT_PSF.
Paul Price, if you're in a position to do this quickly before the next HSC release, please steal it. Otherwise I'll try to get to it later.
Attachments
Issue Links
- is blocked by
-
DM-12985 Bad masking and interpolation around bleeds
- Done
Activity
Looks good.
1) No other obs_packages have overrides, but double check that the default badMaskPlanes for all cameras is desirable:
self.badMaskPlanes = ["NO_DATA", "BAD", "CR", ]
https://github.com/lsst/pipe_tasks/blob/master/python/lsst/pipe/tasks/assembleCoadd.py#L139
2) A search for INTRP in github won't turn up your commit message spelled INTERP. The commit message is clear regardless.
Here's an N_Image (visits contributing to each pixel for a few configs left to right:
1) PRE DM-12968 WarpCompare HSC /datasets/hsc/repo/rerun/private/yusra/psfMatching/DM-12697/deepCoadd/HSC-G/9813/5,8_nImage.fits
2) POST DM-12968 WarpCompare HSC /datasets/hsc/repo/rerun/private/yusra/psfMatching/w_2017_48/warpCompare_DM-12968/deepCoadd/HSC-G/9813/5,8_nImage.fits
3) POST DM-12968 SafeClip HSC /datasets/hsc/repo/rerun/private/yusra/psfMatching/w_2017_48/safeClipOverride/deepCoadd/HSC-G/9813/5,8_nImage.fits
4) POST DM-12968 SafeClip ALL OTHER CAMERAS (see my comment #1 above): /datasets/hsc/repo/rerun/private/yusra/psfMatching/w_2017_48/safeClip/deepCoadd/HSC-G/9813/5,8_nImage.fits
I hadn't realized the defaults for other cameras were in such bad shape - thanks for bringing that to my attention.
In response, I've added some commits to this ticket that moves the (new) HSC defaults into AssembleCoaddTask and removes the overrides in pipe_drivers and obs_subaru. I also moved the badMaskPlanes config option itself from CoaddBaseTask to AssembleCoaddTask, since we weren't actually using it in MakeCoaddTempExpTask.
Yusra AlSayyad, do you mind taking another look, since the changes themselves are now fairly different from what they were before?
OK to merge Jim Bosch
Yusra AlSayyad, since you've already looked at most of this config-only change, could you do the formal review?
I think the most important question is whether there is anywhere else I need to modify these configuration options.
Note that final merge of this is blocked on
DM-12985, but that shouldn't actually require any changes to this ticket.