• Install dependencies: uv sync
• Run tests: uv run pytest
• Run tests with coverage (matches CI): uv run pytest tests --cov=./ --cov-report=term-missing
• Type checking: uv run pyright

This project uses ruff for both linting and formatting.

After editing Python files, run:

uv run ruff format .
uv run ruff check --fix .

• ruff format — black-compatible formatting
• ruff check --fix — linting with auto-fix (import sorting, safe fixes)

Benchmarks use pytest-benchmark and live in benchmarks/. They are not part of the main test suite — run them explicitly.

# Run all benchmarks
uv run pytest benchmarks/ -v --override-ini="addopts="

# Exclude slow (1M feature) cases
uv run pytest benchmarks/ -v --override-ini="addopts=" -m "not slow"

# Only GLM benchmarks
uv run pytest benchmarks/ -v --override-ini="addopts=" -k "glm"

# Save a named baseline
uv run pytest benchmarks/ -v --override-ini="addopts=" --benchmark-save=baseline

# Compare against a saved baseline
uv run pytest benchmarks/ -v --override-ini="addopts=" --benchmark-compare=baseline

Saved baselines go to .benchmarks/.

• Source code: src/bayesianbandits/
• Tests: tests/
• Benchmarks: benchmarks/