Possible path towards improvements of tasks JSON schemas

[tcompa]

Here is how I propose we could proceed in view of improvements of the JSON Schema generation tools:

1. Test JSON-Schema generation on multiple packages #757 (low-effort)
2. Relaxing pydantic requirements #760
3. Switch to Pydantic v2 #592
4. Allow conditional fields & either/or logic #756
5. Allow developers to influence schema building in finer detail #755

It'd be great to work on tasks 1 and 2 first, if possible, so that we have a bit more solid foundations to move on.

[jluethi]

Sounds reasonable.

4 & 5 were not added as something we need to urgently add, but as goals we should work towards.

Step 5 would then be work towards #200 (major topics here aren't just schemas, but also what OME-Zarr metadata we store where & when)