Decide how to serve collections

[sturoscy-personal]

Collections are essentially projects and there won't be many of them initially. Possibly only CMIP6 and CMIP7. Right now, this repo stores them in schemas as json.

• Must be STAC compliant

[bstrdsmkr]

Just my 2 cents, the bridge API and other projects have settled around Pydantic for schema definitions so I'd prefer seeing these modeled there as well. It's probably worth extracting all those models into a separate reusable package as well

[sturoscy-personal]

@bstrdsmkr We're definitely looking into pydantic, but I did confirm with CEDA that they are serving collections via json.

[sturoscy-personal]

Another thing we need to check is whether or not to use upper or lower case on the collection name (cmip6 vs CMIP6). CEDA is using lowercase and we are using uppercase. I believe the uppercase is breaking the following STAC browser https://radiantearth.github.io/stac-browser/#/external/api.stac.esgf-west.org/collections/cmip6?.language=en while CEDA's is working https://radiantearth.github.io/stac-browser/#/external/api.stac.ceda.ac.uk/collections/cmip6?.language=en