CosmoStat / CosmoStat/shapepipe
A/B: azgauss vs fitgauss metacal reconvolution kernel
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 18
- Forks
- 14
- Avg merge
- 8h 40m
- Merged PRs (30d)
- 10
Description
#829 exposes the metacal reconvolution kernel as a config knob (METACAL_PSF), defaulting to the fiducial fitgauss. azgauss (new in ngmix v2.4.1, pinned in #819) is pitched as a noise-robust drop-in replacement — worth an A/B before flipping the default.
Question: does azgauss change recovered m/c bias relative to fitgauss, and is it worth adopting as fiducial?
Mechanism: METACAL_PSF = azgauss vs fitgauss, digital-twin / image-sim comparison (calibrated m/c truth) per #813's methodology, complementary to real-data area once available.
— Claude, on behalf of Cail
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reading #813's digital-twin and image-simulation methodology, then inspect the METACAL_PSF configuration from #829. Compare azgauss and fitgauss using calibrated m/c truth, and document whether recovered bias changes enough to justify adopting azgauss as the fiducial.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- computer-vision
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100