picarlo 0.4.3

Calculate Pi via the Monte Carlo method

Calculating Pi with Modern Tooling

We need a few things:

1. a CLI tool to specify the number of iterations and the number of cores it runs on: picarlo --cores 4 --iterations 10000
2. a library that can be imported in a notebook

[!CAUTION] The CLI requires a few more dependencies like typer

Tooling setup (runs on each commit)

• uv self update & uv sync --group dev
• linting/format checking: uv run ruff check
• auto-formatting: uv run ruff format
• type checking: uv run pyrefly check
• testing: uv run pytest
• integrate into pre-commit pre-commit run --all-files
• use maturin framework to run MC via Rust
• build a GitHub actions pipeline to run prior to merge into main (and post to publish to PyPi)
• run tests in parallel
• split between dev and prod dependencies: uv add --group dev
• add a build system, hatchling in pyproject.toml
• run a build uv build

Checks

1. Check that the package can be installed: uv run --with picarlo --no-project -- python -c "import picarlo"

Goal

1. run from command-line (uvx)
2. import lib into notebook (either via PyPi or from local dist)
3. published module

Useful stuff

1. Create docstrings via LLM
2. Create docs from docstrings
3. Calculate test coverage
4. Tracing
5. server-sent event via starlette

More interesting stuff

• parallelism, e.g. concurrent.futures (see prime number example)
• bulk python version testing through tox

Release preparation

1. Generate Changelog: make changelog (requires Conventional Commits)
2. Bump Version: Update pyproject.toml and Cargo.toml
3. Build: make build
4. Publish: make publish

Commit Convention

This project uses Conventional Commits to automate changelog generation.

• feat: ... for new features
• fix: ... for bug fixes
• docs: ... for documentation
• perf: ... for performance improvements

Example: git commit -m "feat: add parallel processing support"

vscode extensions

1. Ruff
2. TOML syntax highlighting