4pm Pacific Time
BlueJeans: https://bluejeans.com/426716450
attending:
- Eric Bellm
- Ian Sullivan
- Yusra AlSayyad
- Jeffrey Carlin
- Kenneth Herner
- Colin Slater
- Zeljko Ivezic
- Meredith Rawls
- Krzysztof Suberlak
- Unknown User (cmorrison)
- Leanne Guy
- Andrew Connolly
- John Parejko
regrets:
Topics for discussion:
- Precursor processing (AP team):
- Eric Bellm completeness and purity tests of HSC-COSMOS:
- followup: purity (Zooniverse):
-
DM-31341Getting issue details...
STATUS
- edge flags vs. human edge labels: latter are at edges, as expected; former are all over in CCD coordinates, which is weird
- followup Slack discussion identified the issue: edge flags are set both on EDGE and NO_DATA pixels, and pixels which are saturated in all input images of the coadd are excluded with NO_DATA set.
- nondetections: PSF Flux measurements show +/- 5 sigma detections, but aperture flux measurement doesn't. Science looks like noise so this probably points to a narrow variance plane issue being caught up in PSF model
- both have obvious next steps to dig in and investigate
- edge flags vs. human edge labels: latter are at edges, as expected; former are all over in CCD coordinates, which is weird
- followup: purity (Zooniverse):
-
DM-31341Getting issue details...
STATUS
- Kenneth Herner / Ian Sullivan : diffim sprint configurations
- PREOPS-630. made best and middle-third seeing templates.
- best seeing templates
- preconvolve best seeing
- default middle seeing
- convolve science with the middle-third seeing template
- least successful right now–large numbers of excess negative sources
- Ian Sullivan : next step is to restrict analysis to the science images we are interested (e.g., science seeing better than template)
- PREOPS-630. made best and middle-third seeing templates.
- Meredith Rawls : Work-in-progress for getting HiTS working in gen3 with fakes and multi-tract processing
- showing off the pipeline stages for this
- single-CCD completeness plot, for now
- Eric Bellm completeness and purity tests of HSC-COSMOS:
- SITCOM team updates (Andrew Connolly et al.):
- Krzysztof Suberlak showing donut detection & Zernike estimation on AuxTel real data
- also thinking about how to use in-focus images to seed fits of out-of-focus images.
- some AuxTel pointing issues have been resolved
- Zeljko: do you have the piston from the telescope? Andy: yes. We're trying to work out the mirror collimation errors though
- PreOps V&V (Colin Slater ):
- schematic plan for DP0.2
- w40 weekly frozen. Processing team now running a set of precursors at IDF
- V&V then will start validating single-frame, so processing team can start single-frame processing while V&V does coadds/diffim
- schematic plan for DP0.2
- Review CI (https://chronograf-demo.lsst.codes/):
- bump in the object count metrics due to DM-30465
- very modest improvement in gen2 vs gen3 difference?
- memory usage jumped on subsequent config tweak (not currently annotated)
- Ian Sullivan investigate metrics change due to DM-30465
- Next meetings (all at 4pm Pacific):
- November 1
- AOB
- John Parejko merged a command-line Zooniverse cutouts code
Action Items
Description | Due date | Assignee | Task appears on |
---|---|---|---|
| 17 Jul 2023 | Eric Bellm | AP Pipeline Meeting, 2023-06-26 |
| 11 Sep 2023 | Eric Bellm | AP Pipeline Meeting, 2023-08-21 |
| 25 Sep 2023 | Eric Bellm | AP Pipeline Meeting, 2023-08-14 |