This is a package to enable running notebook-based diagnostic workflows. Based on my-cesm-experiment by matt-long: https://github.com/matt-long/my-cesm-experiment.
- Integration with data catalogs
- Inject custom parameters into code and Markdown cells
- Run many notebooks at once on a parameter set, or different parameters per notebook
- Run different notebooks in different environments
- Cache intermediate data products
- Quickly build a Jupyter book with results
-
Clone this repo
-
Activate the environment you want to install
nbscuid
in. Within the clonednbscuid
directory, run:pip install .
Alternatively, to install the commands
nbscuid-run
andnbscuid-build
without installing all ofnbscuid
's dependencies, first installpipx
withpip install pipx
, then run:pipx install .
Create a new folder that contains a config.yml
file. (A guide to what goes in a config.yml
file is coming soon!) This will be the run directory for your collection of notebooks, where all the computed notebooks will appear.
To run all the notebooks you've specified, execute:
nbscuid-run path/to/config.yml
To build the jupyter book if desired, execute:
nbscuid-build path/to/config.yml