Closed zaneselvans closed 2 years ago
I ended up adding a local copy of the data to the shared
folder so that the catalog can be pointed there, rather than having folks potentially cache multiple different copies of the data in different user spaces. It's also faster locally, and catalog access works just the same.
Updated the notebook and jupyterhub configuration. Still waiting on conda-forge distribution of the catalog though.
Once we have an initial release of the data catalog on PyPI /
conda-forge
(see #10):catalystcoop.pudl_catalog
to the environment specified in the pudl-examples repo which builds the Docker container for our JupyterHub.pd.read_parquet()
and through the Intake catalog, both for the monolithic and partitioned datasets, with and without caching turned on to see what the user experience is like.