Notebooks

This directory holds the showcase notebooks rendered into the MyST docs site.

Layout¶

Notebooks are organized by the source repo or topic they showcase:

notebooks/
├── <source-or-topic>/
│   ├── <name>.ipynb        # executed notebook, outputs embedded
│   └── <name>.md           # optional prose write-up / landing page
└── README.md               # this file

Authoring workflow¶

1. Open JupyterLab: pixi run -e jupyterlab lab.
2. Author the notebook under notebooks/<source-or-topic>/<name>.ipynb.
3. Execute all cells end-to-end so outputs (figures, prints, tables) embed into the .ipynb — the committed file is both the source and the rendered output.
4. Optionally add notebooks/<source-or-topic>/<name>.md alongside it for longer prose commentary that links back to the notebook.
5. Commit. MyST picks up both .ipynb and .md via the Notebooks section pattern in myst.yml — no TOC edit needed.

Figures render inline via the notebook’s own outputs; do not save separate PNGs or reference a docs/images/ directory.

Current showcases¶

• pyrox/ — equinox + NumPyro + JAX probabilistic programming (GP regression/classification, regression masterclass, ensemble methods).
• coordax/ — coordinate-aware arrays for JAX (foundations, finite-difference and finite-volume derivatives, ODE/PDE integration and parameter estimation).

Larger sub-projects with their own source tree, tests, and pixi feature live under projects/ rather than here:

• projects/methane_pod/ — thinned marked temporal point processes + POD models for satellite methane retrieval. Self-contained methane_pod package with 4 library modules, 80+ tests, and 8 notebooks (theory .md + executed .ipynb galleries + a NumPyro NUTS fit on synthetic data).
• projects/plume_simulation/ — atmospheric plume dispersion forward models. Currently ships gauss_plume, a steady-state Gaussian plume (JAX forward model + NumPyro Bayesian inference) with 36 tests and three notebooks covering forward simulation, emission-rate parameter estimation, and time-varying-state estimation.