Serve sumission schema files from backend

[pkalita-lbl]

Fixes #1166

Summary

This is partially to support the Field Notes app, but I think it also has benefits for the main submission portal as well. The gist is to allow API clients (whether that's the submission portal or the app) to get submission schema artifacts from the API instead of bundling them in the client application. For the submission portal this hasn't been a major problem (other than the weirdness of having a PyPI artifact as a frontend dependency) since the frontend and backend are part of the same codebase and released at the same time. That won't necessarily be true for the Field Notes app so we want to be able to dynamically fetch the schema.

Details

• A new module nmdc_server/static_files.py defines a path to hold static files and a function which knows how to introspect files provided by the nmdc_submission_schema package and write JSON files into the static file directory. In particular it produces a JSON file representing the whole submission schema (used by the submission portal), files representing individual classes of the schema (likely to be used by the Field Notes app), and the GOLD ecosystem tree (used by the submission portal and likely to be used by the Field Notes app).
• When the backend app starts up, the function to generate the static files is called and the static file directory is mounted under the /static path.
• The frontend app fetches the necessary JSON files from the /static routes instead of importing packaged files.
• Include nmdc-submission-schema as a backend dependency, remove it as a frontend dependency.
• Increment the nmdc-schema dependencies to the latest version.

Comments

I decided on the static file approach because FastAPI will serve the right headers (I think mayble Etag) in the responses to allow browsers to aggressively cache the responses.

In the future we could consider whether we can piggyback on this approach to also remove the nmdc-server frontend dependency.

[naglepuff]

I think this is a pretty elegant solution. I had one comment on the code, and I have additional questions about deployment - is there anything that needs to be done on the spin side to make sure there's no problem creating the static file directory?

[pkalita-lbl]

is there anything that needs to be done on the spin side to make sure there's no problem creating the static file directory?

Not that I'm aware of, but we should find out as soon as this goes to dev if there is.

[naglepuff]

Looks good. I tested this locally by visiting some of the /static urls and confirming that I got some neat json files as a response, and checking that the DH works correctly (i.e. expected columns, validation). As far as I could tell, things appear to be in working order.