Misc things for eScience accelerator 2025
https://escience.washington.edu/global-high-resolution-modeling-a-new-lens-on-the-southern-ocean/
Default 'NPL' Conda environments have a ton of great packages and are updated twice a year. Currently, NPL2025a is the latest environment and has very recent packages (e.g. xarray=2025.1.1). But you can't update existing packages or add new ones to this environment.
For a shared custom Python environment on the NCAR JupyterHub we can use pixi (https://pixi.sh/). It creates "lock" files to guarantee two collaborators use the same packages and versions.
curl -fsSL https://pixi.sh/install.sh | bash
Environments are defined by pixi.toml and pixi.lock files which live in a "project" folder/repository. The actual packages are downloaded and installed in a hiddent .pixi subfolder.
git clone https://github.com/scottyhq/accelerator2025.git
cd accelerator2025
pixi shell # when done type "exit" instead of "conda deactivate"
You might have to logout of the JupyterHub before using the new kernel:
pixi shell
python -m ipykernel install --user --name=accelerator2025