Improve open api schema population

[Leinnan]

Resolves #4407

Besides that it also improves OpenAPI components generation flow, so for example when we have class like MicroserviceRuntimeMetadata with array field public List<FederationComponentMetadata> federatedComponents = new List<FederationComponentMetadata>(); it detects that it is a separate class that is and stores information about the fact that the schema should include type information for FederationComponentMetadata as well. In old version the schema for the field was storing only that:

          "federatedComponents": {
            "type": "array"
          }

Which only informs that there is array of anything. New version gives information about items type and also stores schema for that field type.

[allister-beamable]

The principle seems sound! I am not too familiar with the nuts and bolts of this code, so I defer to Chris and Dias for that part of the review :)

[github-actions]

Lightbeam link

[github-actions]

Lightbeam link

[DiasAtBeamable]

The code looks good to me. But I cannot confirm whether some of the changes will affect how the ClientCode is generated, primarily on the Unreal side. @PedroRauizBeamable has some Unreal project that would be good to test it. I have a sample C# Microservice Code with a lot of callables to test a bunch of types of usages. Let me know if you want it, I can send it to you.

[DiasAtBeamable]

Have you tested those changes in an Unreal and Unity project to see if all the ClientCode is being properly generated?

[Leinnan]

@DiasAtBeamable if you can send it it would be great

[github-actions]

Lightbeam link