All
Faro paper submitted to SPIE. Draft: dmtn-211.lsst.io . Acceptance notification from SPIE is 06 Apr 2022, so nothing more to do until then.
- Along with other changes in Rubin leadership, there will be a reorganization of SIT-Com. Plan by the AMCR.
CI status, bugs, issues of the week
- CI job has not been running for a week. There is an issue with Jenkins. Jobs so slow that they don't finish. KT is on it.
- The new weekly should have the fix for non-deterministic wiggles in metric values. Since dashboard hasn't been running, Jeff is running his own tests.
- When source tables are combined into a single parquet table, they are combined in a random order. Eli implemented change to enforce ordering by detector so that repeatable. Before fix, FGCM was getting slightly different calibrations each run. The algorithms have an order dependence.
- This gives us a sense of the systematic floor of the algorithms
- Separate issues. Require 50 repeated measurements to compute photometric repeatability. Jeff will change configuration for the CI jobs to be 20. yaml configs.
- Fall 2021 epic
- Backlog epic:
- Progress on conversion to parquet
- Last week merged ticket so that refcat loader gives a single merged dataframe that includes refcat columns + proper motion corrected astrometry and color-term-corrected photometry
- Dan Taranu had developed a matcher. To use, we would do matching before the measurement task
- Jeff and Peter working on selectors. There is PR. Once this is merged we can finish converting metrics to use parquet format
- ATLAS REFCAT2 RFC. General support to move forward. Closes tomorrow. Already shared up in the desired format.
Data processing campaigns
- RC2/DC2 reprocessing epic :
- w_2021_40 RC2:
- w_2021_44 DC2 :
- w_2021_40 RC2:
- AuxTel
- AP/DiffIm
Faro for DP0.2
Potential co-working session ideas here
Next meeting Homework task for everyone to take a look at tickets from the previous development epic and we'll do a big review of status + planning for new development cycle next week.
