Skip to content

Installation & setup

CubeDynamics (cubedynamics) runs anywhere xarray does—laptops, clusters, or hosted notebooks. Use this guide to install the package, configure environments, and find ready-to-run notebooks.

Choose an installation source

Install from GitHub today

Grab the latest commits straight from main. Installing inside a virtual environment (venv, Conda, or uv) is recommended but optional.

pip install "git+https://github.com/CU-ESIIL/climate_cube_math.git@main"

Install from PyPI once released

The first PyPI release will ship as soon as the streaming loaders stabilize. At that point you can simply run:

pip install cubedynamics

Until then, the GitHub install above is the canonical way to pick up fixes and examples.

Environment notes

  • Python version – target Python 3.10+ to match the test matrix.
  • xarray + dask – both dependencies ship automatically; if you already manage these packages with Conda, install CubeDynamics inside that environment to avoid duplication.
  • Optional extras – notebooks rely on jupyterlab/notebook, Lexcube visualizations require a live frontend (VS Code, JupyterLab, or Binder).

First steps after install

  1. Launch a notebook (JupyterLab, VS Code, Colab, Binder, etc.).
  2. Import the helpers and stream a cube:

```python import cubedynamics as cd from cubedynamics import pipe, verbs as v

cube = cd.load_prism_cube( lat=40.0, lon=-105.25, start="2000-01-01", end="2020-12-31", variable="ppt", )

pipe(cube) \ | v.anomaly(dim="time") \ | v.month_filter([6, 7, 8]) \ | v.variance(dim="time") ```

  1. Continue into the First PRISM cube guide for more context, AOI patterns, and Lexcube screenshots.

Documentation + notebooks

These notebooks match the code snippets in the documentation, so you can copy/paste cells or launch them on Binder for a fully hosted workflow.