Scope work on RAIL for n(z) modelling and propagation into TXPipe.

[tms-epcc]

From FY24 Q2 QU

Start-up of WP3: scope work on RAIL for n(z) modelling and propagation into TXPipe.

[tms-epcc]

07/MAY/24

• Jaime writing up the proposal following feedback
• once feedback dealt with, will then start coding
• expect two outputs from this
• this will be a means for adding analytical methods (likely a paper from this in the long term)

[tms-epcc]

29/MAY/24

• this will be the RAIL projector
• discussions with various DESC teams on reqs and scope
• creating a library to be hosted in RAIL
• started implementing basic functionality using the old RAIL
• now integrating the package with the RAIL 1.0, needs updating to QP files
• this will then be the MVP of the package for RAIL
• will then extend the models

[tms-epcc]

26/JUN/24

• met with Joe Zuntz and Tilman Tröster to discuss this
• working on interfacing QP to SACC
• software is now an independent library https://github.com/LSSTDESC/rail_prior
• now have MVP with all functionality
• can go from QP output to photometric uncertainty
• so have a end-to-end working example
• now looking to create a test dataset
• discussions on this will take place at the RAIL workshop

[tms-epcc]

24/JUL/24

• at Zurich workshop, presented at Algorithms and pipeline session
• met JoeZ and Tilman
• people need a RAIL example or dataset to try different analyses so now creating a photometric catalog of galaxies based on simulations. This is being shared with rest of collaboration via NERSC.
• also met Alex Knowles and Benjamin Stölzner at UCL workshop and agreed what is needed to connect RAIL interface with SACC - PR in place - https://github.com/LSSTDESC/sacc/pull/107 - expect to finish soon
• met Alex Vicente from Firecrown and have a PR for the integration of the uncertainty models - https://github.com/LSSTDESC/firecrown/pull/441
• also have a PR for the priors - the package from RAIL output to prior photometric uncertainty. Package is orphan so trying to identify a place for this.
• People interested in uncertainty and correlation between uncertainties. This had been merged in the code. Once code settled can hook to SACC.

[tms-epcc]

18/SEP/24

• still working on PRs in Firecrown and SACC : see #107 above
• visited Bochum and implemented the comm models for the priors (https://github.com/LSSTDESC/nz_prior/pull/10)
  • this is now a full DESC workpackage so no longer an orphan ie. https://github.com/LSSTDESC/nz_prior