"name" and "description" should have min_length=1

[ahouseholder]

Extending a comment in #821

If you define them with out the $ref, please add a minLength of 1.

Originally posted by @tschmidtb51 in #821 (comment)

[ahouseholder]

The change to make would be adding Field(...,min_length=1) to both name and description in the ssvc._mixins._Base class.

However, my first attempt at this while working on #821 caused multiple tests in test.test_selections to fail because adding the min_length=1 causes pydantic to reject objects created without specifying name and description.

This will be a problem for both the Selection object and the MinimalDecisionPointValue object, as they both inherit from ssvc._mixins._Base. So whatever solution we come up with needs to happen in both places.

Warning

We could get around this by making them Optional[str] = ... but then that would mean you can create a DecisionPoint object with name=None, description=None which is undesirable. Use of Optional[] in our Selections object also messes with the desire to avoid AnyOf[...,null] in our JSON schemas. So that's out.

Tip

Now it may be possible to work around this in a @model_validator or @field_validator method, for example by setting name=" " to meet the min_length requirement, but then we'll have to also ensure that model_dump_json() on the resulting object doesn't include name=" " (when it should just not have a name key at all).

Note

In a Selection object, name and description must not be required, as our minimal requirement is just for namespace, key, version to be included. (This is distinct from the requirements on DecisionPoint objects, which do require name and description to be specified.

Tip

This might be addressable by making Selection inherit from a different mixin from _Base or somehow altering the fieldspec in theSelection object.

I would prefer, to the degree possible, to avoid building specialized workarounds to pydantic's limitations if we can.

[bernhardreiter]

This issue has two specific examples for #904.

[bernhardreiter]

Tried a few things and browsed the documentation, but so far did not find an obvious way of doing this. I ponder asking the pydantic people about this.