1# Raw & processed data — never committed (see data/raw/README.md for provenance)2data/raw/*3!data/raw/README.md4data/processed/*5!data/processed/.gitkeep67# Python8__pycache__/9*.pyc10.pytest_cache/11.venv/12*.egg-info/1314# LaTeX build artifacts (paper/main.pdf IS committed)15paper/*.aux16paper/*.log17paper/*.out18paper/*.toc19paper/*.lof20paper/*.lot21paper/*.fls22paper/*.fdb_latexmk23paper/*.bbl24paper/*.blg25paper/*.spl26paper/*.synctex.gz27paper/*.run.xml28paper/*.bcf2930# Logs31logs/32*.log3334# macOS35.DS_Store3637# WP45 specific: copyrighted derived full texts and raw model outputs never committed38data/derived/parsed/39data/derived/text/40data/derived/embeddings/*.npy41experiments/outputs/raw/42.env43*.duckdb4445# evaluation fixtures contain copyrighted excerpts46experiments/fixtures/heldout_v1/*/47experiments/outputs/*/outputs/48