# AILANG
.ailang/

# Python
__pycache__/
*.pyc
.venv/
*.egg-info/

# Environment
.env
.env.local
.env.*.local

# mcp-publisher local auth tokens
.mcpregistry_*

# Specs (too large for git)
specs/

# Stress test files (large, downloaded on demand)
data/test_files/stress/*.docx
data/test_files/stress/*.xlsx
data/test_files/stress/*.pptx
!data/test_files/stress/poi_many_merges.xlsx

# arxivbench corpus: commit only .tex/.bbl/.arxiv_id; PDFs, figures, and
# style/class files are re-fetchable via benchmarks/arxivbench/download_corpus.sh.
data/test_files/arxiv/**/*.pdf
data/test_files/arxiv/**/*.png
data/test_files/arxiv/**/*.jpg
data/test_files/arxiv/**/*.jpeg
data/test_files/arxiv/**/*.eps
data/test_files/arxiv/**/*.ps
data/test_files/arxiv/**/*.sty
data/test_files/arxiv/**/*.cls
data/test_files/arxiv/**/*.bst
data/test_files/arxiv/**/*.tar.gz
data/test_files/arxiv/**/*.svg
data/test_files/arxiv/**/*.synctex
data/test_files/arxiv/**/*.synctex.gz
data/test_files/arxiv/**/.svn/
data/test_files/arxiv/**/*.aux
data/test_files/arxiv/**/*.log
data/test_files/arxiv/**/*.out
data/test_files/arxiv/**/*.toc
# Preskill lecture-note bundle is 15MB and exceeds our default --max-bytes;
# exclude wholesale (download on demand).
data/test_files/arxiv/preskill_lectures/

# Tesseract OCR model (5MB binary, not used by the parser)
eng.traineddata

# IDE
.DS_Store
.vscode/
.idea/

# Build artifacts
dist/
build/
*.whl
*.tar.gz

# OS
Thumbs.db

# Test outputs
docparse/data/*.json
docparse/data/*.md
docparse/data/*_summary.txt

# UV
uv.lock
.claude/skills/benchmark/.previous-summary.json
.claude/scheduled_tasks.lock
