Model static data file dependencies

[hellkite500]

To support a deployable sandbox with a selection of available models, we need to identify, document, and curate the required static resources for each model.

This issue is intended to track known existing data dependencies required, where they currently exist, and how we intend to support deploying these into a given sandbox instance.

In order to support a standard, reproducible sandbox with well defined applications, these static data should be pinned to particual model code versions that support those, and I suspect we will need to copy these files into an s3 sandbox archive to guarantee reproducibility? Or we will want to checksum each static resource and keep those available to verify each deployment (assuming the source of the data will be available as long as this project is.)

[patrick-tripp]

We have been saving versioned static data for the NOSOFS model applications since early in this project. We have also saved this data for each of the model applications we have tested. Each institution's model application will have different dependencies and might do things a lot differently, e.g. LiveOcean.

See:
https://github.com/ioos/Cloud-Sandbox/blob/main/cloudflow/workflows/scripts/getNOSOFSfix.sh
and
https://github.com/ioos/Cloud-Sandbox/blob/main/cloudflow/workflows/scripts/get_fixfiles_s3.sh

Saving them to an S3 bucket is the best bet to ensure long-term availability.