sfy-data: analysis endpoints for hm0, trajectory and netcdf-files

[gauteh]

We need backend endpoints for:

• hm0, tp, etc
• trajectory
• netcdf-files

so that these can be displayed on the dashboard and be forwarded to other services.

This should go in a separate server so that the data-collection can operate independently.

Some thoughts:

• The access-token should just be forwarded to the data-server, that way no additional authentication needs to be done.
• Make sure to use predictable queries, i.e. from a specific day to the last received entry. So that the queries may be cached.
• This does not need to be super-flexible, for more advanced stuff we work directly with the data.
• We have two types of buoys: OMB and SFY, ideally both should be supported.

There are two main paths that can be chosen:

• Re-implement the analysis in the server
• Invoke sfydata (from sfy-processing) or call using python API. I am leaning towards this since we then only need to maintain this one place later. The disadvantage is that it is much slower than rust (and more likely to contain vulnerabilities). This limits what we can load by default on the dashboard since it should still be quick to show.

Needs adaptation of the frontend: #62