Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

MCCD on a small DR6 area after the PSFEx production run (proposal)

Aperta
  • Qualcosa non va?
#972 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 1 giorno

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
5/5
Tempo stimato
Più di una settimana
Idoneità per principianti
20/100
Tipo di issue
Funzionalità
Chiarezza
Da chiarire
Stato di attività
Attiva
Stack tecnologico
python
Ambito
data, machine-learning

Direzione di ricerca

Read #674 and inspect workflow/config/cfis/config_exp_mccd.ini after resolving its removed-rule reference. The proposal depends on #601, #630 and #402, and still leaves the patch size, location, runtime plan and comparison owner undecided. Done would mean running MCCD through the v2 workflow on the selected DR6 tiles and comparing the listed PSF residual, leakage, rho/tau and shear catalogue measures against PSFEx on identical objects.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

Proposal, not a decision (UNIONS tomography call, 20 Oct): once the current PSFEx production run on DR6 is done, run MCCD on a small DR6 area through the v2 workflow and compare it with PSFEx on the same tiles. A larger MCCD run would be considered only after that.

Why: on the older catalogues, MCCD gives similar or slightly better PSF systematics, and the 11–12 arcmin B-mode bump improves after footprint matching. A provisional MCCD−PSFEx jackknife on a common footprint shows nothing above ~1.5σ. That is suggestive but not enough to switch the production PSF model. A matched test on DR6 data separates the PSF model from the footprint and selection differences that confound the old comparison.

What the test would be:

  • Area: a contiguous patch already run with PSFEx in v2 production, so the PSFEx side costs nothing extra. Size to be chosen so that rho/tau statistics and leakage are measurable, likely tens of tiles.
  • Run: the same v2 workflow and code commit, with only the PSF model changed (config_exp_mccd.ini path).
  • Compare: per-exposure PSF residuals, star-galaxy leakage, rho/tau statistics, and the shear catalogue on identical objects.

Prerequisite: the MCCD exposure chain has to run in the v2 workflow first. That is #674. Its known blockers are #601 (ModOpt), #630 (interp) and #402 (install), and workflow/config/cfis/config_exp_mccd.ini still references a removed rule.

Open questions: patch size and location; whether MCCD's cost per exposure changes the plan; who runs the comparison side (the cosmo_val / B-mode stack already used for the 2D comparison is the natural place).

Claude (Opus 5.5) on behalf of Cail

Lingua principale
Python
Stelle
18
Fork
14
Merge medio
4g 14h
PR unite (30g)
44

Preparare l'ambiente

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di CosmoStat/shapepipe

Tutte le issue di CosmoStat/shapepipe

Issue simili

Altre issue su Python

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.