|News, announcements and task review|
RHL: Setting up a meeting || to the CAP to deal with problems directly related to data taken on the mountain
ComCAM/AuxTel calibrations, RAT spectroscopy. Role of this meeting is to keep pushing data through the pipelines and available for analysis. Scientific interpretation of the data would happen at a different meeting (probably this one).
Thinking globally about the organization of efforts around collecting data, processing data, and analyzing data. Thinking about the whole lifecycle of the data.
RHL: Move DRP focus from test and deploy to deploy and test. Once something works on a small scale - good enough then through it to the wider QA group. Early large-scale testing on real data. Example: PSFEx → PIFF.
LPG: Need to define the criteria for pushing a component out.
CTS: makes sense from RHL perspective of processing LATISS data, but for someone working on HSC data he considerations are not the same. How do we partition the different cadences? Not everyone will work at the bleeding edge.
CI status, bugs, issues of the week
Why do we have daily apparently randomly fluctuating metric values? NCSA maintenance this morning. Jeff's plan to test two successive runs.
Non-deterministic issue with photometric / astrometric calibrations?
Perhaps due to a config override for FGCM.
One check would be to directly check catalog values output from the Science Pipelines.
Suspicion: non-determinism in the order of input data products in Science Pipelines. Any algorithms that depend on order of input data products could be affected. See DM-32451.
DC2 astrometric repeatability is larger than expected according to Yusra. But maybe DC2 and HSC not directly comparable without a bit of thought about exposure times, atmosphere corrections, etc.TE
|Faro Development status|
TEx is under review.
Peter working on astrometric.
Jeff working on source selector task.
Dan T working on a matching algorithm to take parquet tables as input. Peter implemented a KDTree to start development with. Matching to a refcat - 1-1 matcher not a n-way matcher. Should look at Dan's matcher. Peter and Dan should talk.
KB got a ZP estimte for one of the filters (felh0600 filter cuts on at 600nm). ED picked an example exposure for the g-band. KB will do the same ZP estimate analysis on this. Exposure is from a dense field so do not expect processing issues.
Some config changes coming to deal with getting an astrometircs solution in sparse fields. KB should not need this for the g-band image procesing.
January plans are to put gri in the filter wheel.
ED: Next steps? SFP is working. Can we build a skymap?
Check with Merlin!
Faro for DP0.2
|Trying to wrap up coaad analysis - no outstanding issues found so far.|
Potential co-working session ideas here
Last meeting of 2021 on 14 December. Meet again in 18 January 2022