BlueJeans: https://bluejeans.com/426716450
attending: Eric Bellm Meredith Rawls Unknown User (cmorrison) Ian Sullivan John SwinbankKrzysztof FindeisenUnknown User (gkovacs)
Topics for discussion:
- HiTS processing (Meredith Rawls ):
- still working on collecting functions, talked to Ian about useful pieces, digging through some of Yusra's notebook
- Meredith hasn't yet run the monthly HITS processing but will bump it up the priority list–Leanne is encouraging us to put this into SQUASH and be regular about it.
- DES processing (Unknown User (emorganson)):
- not here
ap_association (Unknown User (cmorrison)):- finishing up plugin system ticket, working; about 30% faster than current system. Marginally lower for just a handful of visits (e.g. CI) because the overhead is proportionally higher
- SDM has been blocked, but pyarrow now into conda which should break things loose
- DCR (Ian Sullivan ):
- Has been working on defining a metric for observation's contribution to DCR model.
- Eric: does this go in the stack, or to the sims team?
- mainly for the sims, although Ian points out that could be relevant after we do various selection
- we think it would be worth having it in the stack for using during the modeling process, quantifying performance, etc.
- seems like perhaps running with 2015 HITS as templates may be more DCR performant
- Image differencing (Unknown User (gkovacs) ):
- talked about calibration with John & John; started pair-coding to put coadds into nanoJansky (not recalibrating, just rescaling); plans to do same with calexp prior to diffim, so we don't have an arbitrary scale to solve in diffim
- will discuss decorrelation at the Wednesday brownbag
- (another potential brownbag topic, for a couple of weeks from now: brainstorming ip_diffim restructuring–maybe Nov. 13th)
- Review CI (https://chronograf-demo.lsst.codes/):
- nothing notable
- AOB
- no meeting next week (John & Eric at DMLT)