Report item schemas in API and use them to dynamically create UIs

[ml-evs]

General idea: we want to dynamically report the item types that a datalab supports. These might be defined outside of this repository at the deployment level, e.g., a datalab instance defines a new custom field on top of our Sample type. The server should then report the JSONSchema for the Sample type at runtime on the server.

This will involve 3 main steps:

• Add an API endpoints for reporting schemas #940
• Use dynamic schema info in UI #952
• Then actually extend our pydantic models with ways to hint about desired UI. This ties in with custom types needing custom components, e.g., date time.
• Document all this!

Points that come to mind:

• determining/reporting type hierarchy
• versioning of schemas
• generally need to think about how we support custom schemas in datalab, of which this is a important part but not the whole story
• this approach, if successful, could also then be used for blocks