.venv/ __pycache__/ *.py[cod] .pytest_cache/ # Runtime data (connections, pending jobs, logs) must never be baked in data/ # IDE / VCS noise .idea/ .vscode/ .git/ # Plan + dev-only docs (keep CLAUDE.md is fine; drop the rest) docs/ # Note: spark/ and hadoop-conf/ are NOT ignored — the Dockerfile COPYs them # as runtime configs. If your environment doesn't have them yet, add a # placeholder file in each (e.g. .gitkeep) so COPY doesn't fail with # "not a directory" or "no such file or directory".