GitHub - keflavich/brick-jwst-2221: Data reduction & analysis for the Brick JWST 2221 data

This repository contains notes & scripts for reduction and analysis of the JWST Cycle 1 project 2221 GO that observed the Brick and Cloud C in NIRCAM narrow-band filters and the MIRI F2550W filter.

While some aspects of this repository are well-organized, the notebooks are not: they often include asides and experiments that lead to dead ends or just got dropped. There are a lot of 'notes in the margins'. Most of them should work if you have the directory structure set up right and the right files in place, but I haven't put in the time needed to make clear what that structure should be.

Package split

As of 2026, the generic JWST photometry pipeline that was originally developed in this repository has been extracted into a separate package, jwst-gc-pipeline, so that other Galactic Center JWST programs can share the same reduction and crowded-field photometry code.

brick2221 (this repository) keeps the Brick / Cloud C science: ice analyses (make_icecolumn_fig9*, icecolumn_*, make_ccd_with_icemodels, co_*, etc.), foreground extinction modeling, paper-figure generators, selections, distance analysis, and analysis_setup / paths configuration.
jwst-gc-pipeline holds the field-agnostic pipeline: PipelineRerun*, destreak, align_to_catalogs, saturated_star_finding, filtering, make_merged_psf, merge_a_plus_b, crowdsource_catalogs_*, merge_catalogs, make_reftable, generic plot_tools, and isochrones.

The old import paths inside brick2221.reduction.* and brick2221.analysis.* continue to work as backward-compatibility shims — they re-export the implementation from jwst_gc_pipeline. New code should import directly from jwst_gc_pipeline.

Iter1 / Iter2 / Iter3 / Iter4 cataloging cycle [OBSOLETE]

After the JWST pipeline (PipelineRerunNIRCAM-LONG.py etc.) produces per-frame *_destreak_o<NNN>_crf.fits images, photometry is built up across four iterative passes. Every pass runs the same script (jwst_gc_pipeline.photometry.crowdsource_catalogs_long / brick2221.analysis.crowdsource_catalogs_long) on a per-frame basis, but with different inputs and goals.

New default photometry pipeline (2026-06-09): the manual-iteration pipeline (jwst_gc_pipeline.photometry.cataloging.run_manual_pipeline) replaces IterativePSFPhotometry as the default PSF-photometry path. It runs the m12 → m3 → m4 → m5 → m6 → m7 phases in-process as a single non-array job (phases are sequential, per-frame fits within a phase parallelize via --parallel-workers). Output tokens are _m1.._m7, _dao_basic — disjoint from the legacy _iter2/_iter3/ _iter4 / _daoiterative products. See jwst-gc-pipeline/PHOTOMETRY_PIPELINE.md for the algorithm.

Use brick2221/shellscripts/submit_manual_pipeline.sh to launch the new path (see "Manual-iteration pipeline" section below).

Legacy scripts retired (2026-06-16): the legacy iter1→iter2→merge→ iter3→iter4 shell submitters (submit_full_chain.sh, run_iter4resbgrefit.sh, run_merge_each*.sh, run_residbg_cataloging.sh, the run_all_cataloging*/sbatch_cataloging* drivers, the sgrb2 recovery/continuation scripts, run_mosaicing_only_sickle.sh, …) now live in brick2221/shellscripts/legacy pipeline/ and exit 1 with a deprecation banner if invoked. run_iter3_cataloging.sh is temporarily retained in place (unmoved) only because 65 in-flight V9 iter3-launch jobs exec it by absolute path; retire it once those drain. See brick2221/shellscripts/README.md for the full active-vs-legacy list.

The legacy iter1/iter2/iter3/iter4 description below is retained for reference and remains accurate for the (now-retired) legacy path.

iter1 — basic per-frame DAO seed. crowdsource_catalogs_long.py --each-exposure --daophot --skip-crowdsource (no --iteration-label). Finds peaks above noise on each per-frame destreak/crf image and runs one DAOPHOT PSF-fit pass. Output is the seed catalog used downstream.
iter2 — residual-aware refit, per filter. Same script with --iteration-label=iter2 --postprocess-residuals. Uses the iter1 catalog as a seed, fits the residual for missed sources, and writes a cleaned per-frame catalog + residual. merge_catalogs.py --merge-singlefields --indiv-merge-methods=daoiterative --iteration-label=iter2 then merges per-frame catalogs within each filter (cross-exposure merge) — the per-filter daoiterative output is what feeds the iter3 cross-band union seed.
iter3 — cross-band union-seeded fit. build_union_seed_catalog.py unions the iter2 per-filter merged catalogs into a single seed_union_iter3_<target>.fits, then crowdsource_catalogs_long.py --iteration-label=iter3 --seed-catalog=<union> --postprocess-residuals is run per frame. This is the canonical "all filters, all frames" simultaneous fit and produces the final per-frame iter3 photometry + residual mosaics. merge_catalogs.py --iteration-label=iter3 produces the merged iter3 multiwavelength catalog. Per-frame "forced photometry on satstar residuals" (forced_photometry_residuals.py) is gated on iter3 to recover sources undetected in individual filters.
iter4 — residual-background refit (optional refinement). run_iter4resbgrefit.sh builds median-smoothed background images from the iter3 residuals, then refits the iter3 catalog as seeds with tight xy_bounds on the residual-bg-subtracted data. Output --iteration-label=iter4resbgrefit is purely additive — iter1/2/3 are not overwritten. A parallel iter2residbg / iter3residbg cascade exists in run_residbg_cataloging.sh for early-stage residbg-driven refitting.

The cross-iteration dependency tree is

iter1 (per-filter, per-frame DAO)
  ↓ (per-filter merge)
iter2 (per-filter, per-frame, with residual postprocess)
  ↓ (per-filter merge → daoiterative)
seed_union_iter3 (cross-band union)
  ↓
iter3 (per-frame, all sources fit jointly, with residuals)
  ↓ (per-target merge)
iter4resbgrefit (residual-bg refinement, optional)

Job runner coverage matrix

Active entry points in brick2221/shellscripts/ (post-2026-06-16 cleanup). Cataloging is the manual-iteration pipeline (default). Reduction scripts are upstream (they produce the *_crf inputs) and remain required.

Script	Stage	Pipeline	Refcat	Notes
`submit_manual_pipeline.sh`	cataloging (m12→m7)			THE cataloging submitter. Single in-process job, `--manual-iterations` default; `_m1.._m7`/`_dao_basic` output
`sickle_miri_f770w_cataloging_*.sh`	cataloging (MIRI)			Hand-written manual-path MIRI F770W jobs (single in-process, not array)
`run_full_pipeline_common.sh`	reduction + iter1	✓	✓	First-pass + refcat + second-pass; its step-4 per-exposure cataloging now runs the manual path
`run_full_pipeline_<target>.sh`	reduction	✓	✓	Thin wrappers over `_common.sh` (arches, cloudef, gc2211, quintuplet, sgra, sgrb2, sgrc, w51, wd1, wd2)
`run_pipeline_long_sickle.sh` / `run_pipeline_miri_sickle.sh`	reduction	✓		Sickle NIRCam-LONG / MIRI calibration
MIRI f2550w `*_image3/edgetrim/homogenize/colprofile`, `cloudc_miri_f2550w_pipelinemiri.sh`	reduction	✓		MIRI f2550w mosaic post-processing
`sickle_build_miri_large_psf.sh`	utility			Builds large MIRI PSF grids

Retired to legacy pipeline/ (deprecated, exit 1): the legacy iter1→iter2→merge→iter3→iter4 submitters — submit_full_chain.sh, run_iter4resbgrefit.sh, run_merge_each{,_brick,_cloudc}.sh, run_residbg_cataloging.sh, run_all_cataloging{,_sgrb2,_sickle}.sh, run_f115w_cataloging.sh, run_cataloging_eachexposure{,_f115w}.sh, run_merge_f115w.sh, all_cataloging_merge{,_blur}.sh, sbatch_cataloging.sh, sbatch_commands_nobgsub.sh, run_{continuation,continuation2,recovery,recovery2}_sgrb2.sh, resubmit_failed_brick_sickle_2026-04-20.sh, run_mosaicing_only_sickle.sh. run_iter3_cataloging.sh stays in place until its 65 in-flight V9 iter3-launch jobs drain (it is exec'd by absolute path), then retire it too.

Target / observation coverage

Each target's full per-frame set must be reduced across all observations that belong to it. Missing an observation here means a permanent hole in the merged catalog. Coverage as of 2026-06-07:

Target / "name"	Proposal	Obs (fields)	Filters
brick (2221 narrowband)	2221	001	F182M F187N F212N F405N F410M F466N
brick-1182 (broadband)	1182	004	F115W F200W F356W F444W
cloudc	2221	002	F182M F187N F212N F405N F410M F466N
sickle	3958	007 (NIRCam); MIRI 001, 002	F187N F210M F335M F470N F480M (NIRCam); F770W F1130W F1500W (MIRI obs 001/002)
brick (3958 MIRI)	3958	003	F770W F1130W F1500W — MIRI obs 003 is the BRICK, not the sickle (routes to `brick/`)
sgrb2	5365	001	F150W F182M F187N F210M F212N F300M F360M F405N F410M F466N F480M
sgra	1939	001	F115W F212N F405N
cloudef	2092	002 + 005	F162M F210M F360M F480M (Cloud E = 002, Cloud F = 005, merged)
arches	2045	001	F212N F323N
quintuplet	2045	003	F212N F323N
sgrc	4147	012	F115W F162M F182M F212N F360M F405N F470N F480M
gc2211	2211	023, 028, 046, 049, 050	obs 028: F150W,F277W; others: F200W,F277W

Multi-obs targets (cloudef, gc2211) require the runner to loop over every obs id; missing one silently drops part of the field. Brick is covered as two separate launcher targets (brick for proposal 2221 + brick-1182 for the broadband proposal 1182) because the per-frame destreak_o<NNN>_crf suffix is obs-specific.

Program 3958 spans two targets. NIRCam obs 007 and MIRI obs 001/002 are the sickle; MIRI obs 003 is the brick (it shares the 3958 program id but is a separate pointing). Both PipelineMIRI.py and the cataloging field_to_reg_mapping route 3958/003 to the brick/ tree, so its images and catalogs land under /orange/.../jwst/brick/ (== the /blue/... brick symlink) and never collide with sickle/ products. Catalog it with --target=brick --proposal_id=3958 --field=003.

Reduction process

Pipeline files are in reduction/.

PipelineRerunNIRCAM-LONG.py run the JWST pipeline with modifications to the tweakwcs stage
destreak.py runs "Massimo's Destriper", which is a simple percentile-subtraction across the X-axis of each horizontal quadrant of each detector (this is run by the pipeline)
align_to_catalogs.py includes some post-facto re-alignment tools. One function runs only on the final processed data & matches it to VVV. (this is run by the pipeline)
saturated_star_finding.py performs PSF fitting on saturated stars and removes them. (this is run by the pipeline)
crowdsource_catalogs_long.py runs the crowsource extraction algorithm. This must be run on the long-wavelength channels before running the short-wavelength pipeline to provide the reference catalog we use for the shortwave data.
merge_catalogs.py merges the multiwavelength catalogs.
make_reftable.py makes the reference table from the long-wavelength (F410M) data to be used on the short-wavelength data (this is run by merge_catalogs) - note that this is in analysis/
PipelineRerunNIRCAM-SHORT.py run the JWST pipeline with modifications to the tweakwcs stage including a reference catalog generated from F410M
crowdsource_catalogs_long.py on the short data (it has _long in the name, but I merged both into this one; the _short version is deprecated) - note that this is in analysis/
merge_catalogs.py again to finally merge wavelengths - note that this is in analysis/

Notebooks. There are a lot of these.

BrA_separation_nrc{a,b}.ipynb should be run after the pipeline. They subtract BrA from F410M, then subtract the rescaled continuum from the BrA image.
F466_separation_nrc{a,b}.ipynb does the same thing, but for the F466 filter. It doesn't work as well because of the wavelength difference.
StarKILLER_nrc{a,b}.ipynb manually remove stars from the images, only for display purposes!
Stitch_A_to_B.ipynb pulls the two modules together. It makes the "star-free" three-color merged image
nrca_contsub_colorimages.ipynb and BrickImages.ipynb make 3-color images.

Most of the rest are test / WIP things.

Analysis

The analysis has been done haphazardly in notebooks, but then I tried to reconcile different notebooks and merge them.

analysis_setup.py - loads the latest versions of 3-color images and catalogs. Meant to be called before running plotting things.
co_fundamental_modeling.py - extracts from the CO2_PhoenixModel.ipynb, CO2_Phoenix_Example.ipynb, and COFundamentalModeling_and_IceModeling.ipynb notebooks. This script produces a wider range of models than those notebooks
plot_tools.py - contains many plot templates for color-color, color-magnitude, etc.
selections.py - downselects from the full merged catalog of NIRCAM data of the brick, in all 6 filters, to the subset(s) that are reliable
ColorVsCOIceAnalysis.ipynb - makes several color-color plots

Manual-iteration pipeline (new default, 2026-06-09)

The new default PSF-photometry pipeline (run_manual_pipeline in jwst_gc_pipeline.photometry.cataloging) is launched via brick2221/shellscripts/submit_manual_pipeline.sh. Phases m12 → m3 → m4 → m5 → m6 (→ m7 if multi-filter) run sequentially in one process; per-frame fits inside each phase parallelize via --parallel-workers=<N> (defaults to --cpus-per-task).

Defaults

env var	default	meaning
`MANUAL_CPUS`	32	`--cpus-per-task` AND `--parallel-workers`
`MANUAL_MEM`	256gb	`--mem` per node
`MANUAL_TIME`	96:00:00	walltime

Override at launch, e.g.:

MANUAL_CPUS=64 MANUAL_MEM=512gb \
  bash submit_manual_pipeline.sh sgrb2 F212N nrcb V9

Usage

submit_manual_pipeline.sh <target> <filter[,filter,...]> <module> [tag] [extra_dep]

Module token rules (see PHOTOMETRY_PIPELINE.md):

A single token nrcb (or nrca) auto-expands to nrcb1..nrcb4 for SW filters and nrcblong for LW filters within one run.
Use merged to process all detectors via the module-merging codepath.

The <filter> argument is passed verbatim to --filternames. A single filter token = single-filter run (no m7 cross-band seed). COMMA-SEPARATED filters engage the m7 cross-band seed across all given filters. The full-coverage call for each target is multi-filter; single-filter calls are only useful for smoke tests on one band.

Multi-filter calls must share the same each_suffix family — sgrb2 LW/SW mixes are rejected by the script (split into two calls; see "Special-case targets" below).

Per-target copy-pastable examples

The first line for each target is the canonical full-coverage call (multi-filter, engages m7). The second line is a single-filter smoke-test.

# brick (proposal 2221 narrowband, 6 filters)
bash submit_manual_pipeline.sh brick F182M,F187N,F212N,F405N,F410M,F466N merged
bash submit_manual_pipeline.sh brick F405N merged

# brick-1182 (proposal 1182 broadband, 4 filters; outputs under /jwst/brick/)
bash submit_manual_pipeline.sh brick-1182 F115W,F200W,F356W,F444W merged
bash submit_manual_pipeline.sh brick-1182 F115W merged

# cloudc (proposal 2221 obs 002, 6 filters)
bash submit_manual_pipeline.sh cloudc F182M,F187N,F212N,F405N,F410M,F466N merged
bash submit_manual_pipeline.sh cloudc F410M merged

# sickle (proposal 3958 obs 007, 5 filters)
bash submit_manual_pipeline.sh sickle F187N,F210M,F335M,F470N,F480M nrcb
bash submit_manual_pipeline.sh sickle F470N nrcb

# sgra (proposal 1939, 3 filters)
bash submit_manual_pipeline.sh sgra F115W,F212N,F405N merged
bash submit_manual_pipeline.sh sgra F212N merged

# arches (proposal 2045 obs 001, 2 filters)
bash submit_manual_pipeline.sh arches F212N,F323N merged
bash submit_manual_pipeline.sh arches F212N merged

# quintuplet (proposal 2045 obs 003, 2 filters)
bash submit_manual_pipeline.sh quintuplet F212N,F323N merged
bash submit_manual_pipeline.sh quintuplet F212N merged

# sgrc (proposal 4147 obs 012, 8 filters)
bash submit_manual_pipeline.sh sgrc F115W,F162M,F182M,F212N,F360M,F405N,F470N,F480M merged
bash submit_manual_pipeline.sh sgrc F212N merged

Special-case targets

sgrb2 (proposal 5365) — LW filters use align_o001_crf, SW filters use destreak_o001_crf. Multi-filter calls that mix LW + SW are rejected by the script; split into two calls per family:

# SW family (canonical full coverage, 5 filters)
bash submit_manual_pipeline.sh sgrb2 F150W,F182M,F187N,F210M,F212N nrcb
# LW family (canonical full coverage, 6 filters)
bash submit_manual_pipeline.sh sgrb2 F300M,F360M,F405N,F410M,F466N,F480M nrcb
# Single-filter smoke tests
bash submit_manual_pipeline.sh sgrb2 F212N nrcb   # SW
bash submit_manual_pipeline.sh sgrb2 F360M nrcb   # LW

cloudef (proposal 2092) — two obs (002 Cloud E, 005 Cloud F) that must be reduced together. submit_manual_pipeline.sh cloudef auto-loops the fields array and submits ONE sbatch per obs. Each obs's run_manual_pipeline produces per-obs outputs under /orange/adamginsburg/jwst/cloudef/. Cross-obs catalog union is a separate manual step (call merge_catalogs.py over the two output trees once both jobs complete; cross-obs auto-union in the manual pipeline is a TODO).

bash submit_manual_pipeline.sh cloudef F210M nrcb
bash submit_manual_pipeline.sh cloudef F162M,F210M,F360M,F480M nrcb

gc2211 (proposal 2211, asteroid survey, 5 GC pointings) — each obs is its own launcher target (gc2211-023, gc2211-028, gc2211-046, gc2211-049, gc2211-050). Filter sets vary per obs. Submit one call per obs:

bash submit_manual_pipeline.sh gc2211-023 F200W,F277W merged
bash submit_manual_pipeline.sh gc2211-028 F150W,F277W merged   # obs 028 is F150W not F200W
bash submit_manual_pipeline.sh gc2211-046 F200W,F277W merged
bash submit_manual_pipeline.sh gc2211-049 F200W,F277W merged
bash submit_manual_pipeline.sh gc2211-050 F200W,F277W merged

Output locations

Per-frame and merged outputs land under <basepath>/<filter>/pipeline/ and <basepath>/catalogs/ with tokens _m1.._m7, _dao_basic, plus _vetted and _allcols catalog variants (see PHOTOMETRY_PIPELINE.md §Outputs). These are disjoint from the legacy _iter2/_iter3/_iter4 tokens — both paths coexist in one tree without collision.

Setup for Reduction

There are a few basic setup requirements before reduction of a new target:

The target data root must exist at /orange/adamginsburg/jwst/{target}/.
A crds/ directory must exist under that target root.
A reduction/fwhm_table.ecsv file must exist under that target root.
A catalogs/ directory should exist under that target root (it will be created automatically by helper scripts if missing).
If you want VVV-based alignment for that target, provide a field-of-view region file under regions_/ and wire it into fov_regname in brick2221/reduction/PipelineRerunNIRCAM-LONG.py.

Validated mapping locations in the codebase (these must be coherent for any new target):

brick2221/reduction/PipelineRerunNIRCAM-LONG.py
- proposal/field -> target mapping in field_to_reg_mapping
- proposal/field -> reference catalog mapping in REFERENCE_ASTROMETRIC_CATALOG_BY_FIELD
- optional target -> fov region mapping in fov_regname
brick2221/analysis/crowdsource_catalogs_long.py
- proposal/field/target mapping in field_to_reg_mapping
- visit-count policy in nvisits
brick2221/analysis/merge_catalogs.py
- filter inventory per target/proposal in obs_filters
- target/proposal -> field mapping in project_obsnum

Current first-pass/second-pass behavior:

If the configured reference catalog does not exist yet, PipelineRerunNIRCAM-LONG.py now runs a first pass and skips reference-catalog realignment steps.
After building a reference catalog, rerun with --skip_step1and2 to reuse existing _cal products and perform the aligned second pass.

Full Pipeline runners

The "full pipeline" for one target is the end-to-end dependency tree:

First-pass pipeline jobs for all configured filters (no refcat).
Reference-catalog build via make_reference_from_pipeline_catalogs.py.
Second-pass pipeline jobs with --skip_step1and2.
Per-exposure iter1 DAO cataloging arrays via crowdsource_catalogs_long.py.
Cross-filter merge via merge_catalogs.py --merge-singlefields.

Iter2/iter3/iter4 are NOT included — run submit_full_chain.sh / run_iter3_cataloging.sh / run_iter4resbgrefit.sh afterwards.

Entry points (one per target):

run_full_pipeline_common.sh <target> — generic per-target runner. Supports cloudef sgrc sgrb2 arches quintuplet sgra gc2211.
run_full_pipeline_<target>.sh — thin per-target wrappers (arches, cloudef, quintuplet, sgra, sgrb2, sgrc).
run_full_pipeline_gc2211.sh — wraps _common.sh and loops over the 5 gc2211 obs IDs (one full pipeline per obs). Required because each gc2211 obs is a separate field with its own filter set.
run_full_pipeline_cloudef_sgrc.sh — multi-target launcher. Despite the name, runs the full pipeline for whatever target list you pass on the command line, OR a default list of cloudef sgrc sgrb2 arches quintuplet sgra when called with no args. Does NOT include brick or gc2211 in its default list — brick is reduced separately (see below), and gc2211 needs its own per-obs wrapper.

Brick is handled separately. Proposal 2221 (narrowband) and proposal 1182 (broadband) have historical per-frame outputs already on disk under /orange/adamginsburg/jwst/brick/. There is no run_full_pipeline_brick.sh; bring brick through submit_full_chain.sh (handles iter1/iter2/merge/iter3) once per (target, filter, module), where target is brick (2221) or brick-1182 (1182). See the iter1+ cataloging example below.

Copy-pastable run examples

From brick2221/shellscripts/:

# --- Per-target full pipeline (steps 1-5: pipeline + refcat + iter1 merge) ---

bash run_full_pipeline_cloudef.sh        # cloudef (auto-loops obs 002 + 005)
bash run_full_pipeline_sgrc.sh           # sgrc
bash run_full_pipeline_sgrb2.sh          # sgrb2
bash run_full_pipeline_arches.sh         # arches
bash run_full_pipeline_quintuplet.sh     # quintuplet
bash run_full_pipeline_sgra.sh           # sgra
bash run_full_pipeline_gc2211.sh         # gc2211 (auto-loops obs 023 028 046 049 050)

# Single gc2211 obs only:
GC2211_OBS=028 bash run_full_pipeline_gc2211.sh

# Multi-target convenience launcher (default list: cloudef sgrc sgrb2 arches quintuplet sgra):
bash run_full_pipeline_cloudef_sgrc.sh
# Or a custom subset:
bash run_full_pipeline_cloudef_sgrc.sh cloudef sgrc

# --- iter1 + iter2 + iter2-merge + iter3-launch chain (per filter+module) ---
# submit_full_chain.sh <target> <filter> <module> [tag] [extra_dep]

bash submit_full_chain.sh brick        F405N merged V8   # proposal 2221 narrowband
bash submit_full_chain.sh brick-1182   F115W merged V8   # proposal 1182 broadband
bash submit_full_chain.sh cloudc       F410M merged V8
bash submit_full_chain.sh cloudef      F480M merged V8   # auto-loops obs 002 + 005
bash submit_full_chain.sh sickle       F470N nrcb  V8
bash submit_full_chain.sh sgrb2        F212N merged V8
bash submit_full_chain.sh sgra         F212N merged V8
bash submit_full_chain.sh arches       F212N merged V8
bash submit_full_chain.sh quintuplet   F212N merged V8
bash submit_full_chain.sh sgrc         F212N merged V8
bash submit_full_chain.sh gc2211-028   F277W merged V8   # per-obs (one of 023/028/046/049/050)

# --- iter3 launch only (assumes iter2 done) ---
bash run_iter3_cataloging.sh --target brick
bash run_iter3_cataloging.sh --target brick-1182
bash run_iter3_cataloging.sh --target cloudef           # loops obs 002 + 005
bash run_iter3_cataloging.sh --target gc2211-046

# --- iter4 refinement (assumes iter3 done) ---
bash run_iter4resbgrefit.sh --target sickle
bash run_iter4resbgrefit.sh --target brick
bash run_iter4resbgrefit.sh --target sgrb2

Defaults (full pipeline runners + `submit_full_chain.sh`)

Proposal id / field(s) / filters / reference filter used when no env-var override is given. Multi-obs targets are marked bold.

brick: proposal 2221, field 001, filters F182M,F187N,F212N,F405N,F410M,F466N, reference filter F405N
brick-1182: proposal 1182, field 004, filters F115W,F200W,F356W,F444W, reference filter F405N (cross-proposal — brick narrowband seed)
cloudc: proposal 2221, field 002, filters F182M,F187N,F212N,F405N,F410M,F466N, reference filter F405N
cloudef: proposal 2092, fields 002,005 (Cloud E + Cloud F), filters F162M,F210M,F360M,F480M, reference filter F210M (merge_ref_filter=f360m)
sgrc: proposal 4147, field 012, filters F115W,F162M,F182M,F212N,F360M,F405N,F470N,F480M, reference filter F212N
sgrb2: proposal 5365, field 001, filters F150W,F182M,F187N,F210M,F212N,F300M,F360M,F405N,F410M,F466N,F480M, reference filter F210M
sgra: proposal 1939, field 001, filters F115W,F212N,F405N, reference filter F212N
arches: proposal 2045, field 001, filters F212N,F323N, reference filter F212N
quintuplet: proposal 2045, field 003, filters F212N,F323N, reference filter F212N
sickle: proposal 3958, field 007, filters F187N,F210M,F335M,F470N,F480M, reference filter F470N
gc2211: proposal 2211, obs 023, 028, 046, 049, 050 (each is its own field; launched per-obs by run_full_pipeline_gc2211.sh).
- obs 028: filters F150W,F277W, reference filter F150W
- obs 023/046/049/050: filters F200W,F277W, reference filter F200W
- merge_ref_filter=F277W (only filter common to all five obs)

Override defaults with env vars before launching, e.g.:

CLOUDEF_FIELD, CLOUDEF_FIELDS (CSV for multi-obs), CLOUDEF_FILTERS, CLOUDEF_REF_FILTER
SGRC_FIELD, SGRC_FILTERS, SGRC_REF_FILTER
SGRB2_FIELD, SGRB2_FILTERS, SGRB2_REF_FILTER
ARCHES_FIELD, ARCHES_FILTERS, ARCHES_REF_FILTER
QUINTUPLET_FIELD, QUINTUPLET_FILTERS, QUINTUPLET_REF_FILTER
SGRA_FIELD, SGRA_FILTERS, SGRA_REF_FILTER
GC2211_OBS (CSV subset of 023,028,046,049,050)
MODULES (default merged)
ARRAY_RANGE (default 0-23; auto-sized from destreak file count once those exist)

Name		Name	Last commit message	Last commit date
.github		.github
alma		alma
brick2221		brick2221
catalogs		catalogs
docs		docs
licenses		licenses
offsets		offsets
regions_		regions_
synthetic_source_recovery		synthetic_source_recovery
MANIFEST.in		MANIFEST.in
README.md		README.md
README.rst		README.rst
pyproject.toml		pyproject.toml
sbatch_commands.md		sbatch_commands.md
setup.py		setup.py
tox.ini		tox.ini

Navigation Menu

Repository files navigation

Package split

Iter1 / Iter2 / Iter3 / Iter4 cataloging cycle [OBSOLETE]

Job runner coverage matrix

Target / observation coverage

Reduction process

Analysis

Manual-iteration pipeline (new default, 2026-06-09)

Defaults

Usage

Per-target copy-pastable examples

Special-case targets

Output locations

Setup for Reduction

Full Pipeline runners

Copy-pastable run examples

Defaults (full pipeline runners + `submit_full_chain.sh`)

About

Releases

Packages

Used by

Contributors

Languages

Latest commit

History

Folders and files

Repository files navigation

Package split

Iter1 / Iter2 / Iter3 / Iter4 cataloging cycle [OBSOLETE]

Job runner coverage matrix

Target / observation coverage

Reduction process

Analysis

Manual-iteration pipeline (new default, 2026-06-09)

Defaults

Usage

Per-target copy-pastable examples

Special-case targets

Output locations

Setup for Reduction

Full Pipeline runners

Copy-pastable run examples

Defaults (full pipeline runners + submit_full_chain.sh)

About

Resources

Stars

Watchers

Forks

Releases

Packages

Used by

Contributors

Languages

Defaults (full pipeline runners + `submit_full_chain.sh`)