.gitignore

# =============================================================================
# Personal data — NEVER commit. These contain or derive from individual genomes.
# =============================================================================
# Raw provider files (the original 23andMe / AncestryDNA / etc. downloads)
raw-source-genomes/

# Standardized chip parquet + any per-subject files at this level
standardized-genomes/*.parquet
standardized-genomes/imputed/

# Per-subject haplogroup outputs (MT VCFs, yhaplo genos + per-sample results)
standardized-genomes/haplogroups/

# Per-subject metadata (names, sex, self-reported phenotypes, etc.)
profiles/*.json

# Append-only ledger — contains subject-specific findings and inferences
ledger/*.jsonl
ledger/*.jsonl.bak.*

# Generated reports
reports/

# Per-subject session notes (see local/README.md)
local/

# Per-subject empirical PRS calibrations (full-panel <PGS>.json files are
# public-content-safe; subject-observed <PGS>.<subject>.json ones are not).
# Using a two-dot glob that matches <PGS>.<subject>.json but not <PGS>.json.
reference/population_cache/prs_empirical/*.*.json

# Session-level user memory (lives under ~/.claude/ but guard anyway)
.claude/memory/
.claude/scheduled_tasks.lock

# =============================================================================
# Large re-fetchable infrastructure — instructions in README.md § "One-time setup"
# =============================================================================
# 1000 Genomes imputation reference panels (~16 GB; scripts/imputation_download.py)
reference/imputation/1kg_ref_b37/

# 1000 Genomes EBI release VCFs (~25 GB; scripts/download_1kg_canonical.py)
reference/imputation/1kg_ebi_release/

# Portable JDK (~200 MB; scripts/install_portable_jdk.py)
reference/imputation/jdk/

# Genetic maps (~25 MB; scripts/imputation_download.py --maps)
reference/imputation/genetic_maps/

# Beagle + conform-gt JARs (~500 KB vendor binaries; scripts/imputation_download.py)
reference/imputation/beagle/*.jar

# HaploGrep3 install (~50 MB vendor binary; scripts/haplogrep_download.py)
reference/haplogroups/mtdna/haplogrep3/

# Canonical 1000G EUR AF table (~440 MB; scripts/extract_eur_afs.py, ~20 min)
reference/population_cache/1kg_eur_afs.parquet
reference/population_cache/1kg_eur_afs.chr*.parquet
reference/population_cache/1kg_eur_afs.json

# PGS Catalog weight files (each ~20 MB; scripts/prs_download.py <PGS_id>)
reference/prs_weights/*.txt.gz

# =============================================================================
# Build / runtime artifacts
# =============================================================================
__pycache__/
*.pyc
*.pyo
.venv/
venv/
env/
.env
.ipynb_checkpoints/

# =============================================================================
# OS / IDE junk
# =============================================================================
.DS_Store
Thumbs.db
desktop.ini
.vscode/
.idea/
*.swp

# =============================================================================
# Intentionally tracked (keep commented as documentation)
# =============================================================================
# reference/curated_snps.tsv                          — curated SNP table, shareable
# reference/carrier_panels/*.tsv                      — curated ACMG panel variants
# reference/haplogroups/mtdna/rCRS.fasta              — public mtDNA reference sequence
# reference/population_cache/prs_empirical/<PGS>.json — full-panel 1000G EUR calibrations,
#                                                       content-safe (per-sample scores are
#                                                       public 1000G data, not a subject's)
# reference/population_cache/prs_empirical/README.md  — documents subject-observed pattern
reference/haplogroups/mtdna/rCRS.fasta