SubClass annotations are missing from the base class

[yaelah]

In order for jackson ObjectMapper to deserialize correctly a base class to its subclasses, we need to include annotations defining all expected subclasses in the parent class.
The codegen currently does not generate those annotations.
For this example YAML:

definitions:
  Base:
    type: object
    discriminator: disc
    properties:
      disc:
        type: string
    required:
      - disc
  SubClass:
    allOf:
      - $ref: '#/definitions/Base'
        required:
          - field
        properties:
          field:
            type: string

Jackson expects these annotations to be produced, but they are not:
@JsonTypeInfo(use = JsonTypeInfo.Id.NAME, include = JsonTypeInfo.As.PROPERTY, property = "disc")
@JsonSubTypes({ @type(value = SubClass.class, name = "SubClass") })
public class Base {
}

[wing328]

@yaelah thanks for the detailed above. Would you have cycle to contribute a fix?

Here is a good starting point: https://github.com/swagger-api/swagger-codegen/blob/master/modules/swagger-codegen/src/main/resources/Java/model.mustache

[wing328]

Similar ticket: #1253

[yaelah]

I have a fix for this. What is the process for submitting patches? I have never submitted anything to this project before.
thanks.

[wing328]

Please refer to https://github.com/swagger-api/swagger-codegen/blob/master/CONTRIBUTING.md for contributing guidelines.

If you need help, just reply to let us know.

[yaelah]

Thanks for noticing my wrong email.

[alex-estela]

Hi guys,
Also currently facing this issue. Any news about the release of this fix ?
Thanks in advance

[triptec]

@wing328 Do you know anything about progress on this?

[wing328]

I believe @yaelah has fixed the issue in his branch but didn't have time to submit a PR yet.

Can you test his fix via https://github.com/yaelah/swagger-codegen to see if it resolves the issue for you?

If the fix works, I can cherry-pick the commits to submit a PR for the fix.

[szakrewsky]

I modified @yaelah post to work on the latest master branch. I have been using this for a while now, hope it helps. One issue I did notice is that it won't apply to sub-sub classes.

Update: the sub-sub class issue has now been fixed in my latest pr commit.

[wing328]

@szakrewsky I added a comment to the PR #4085. Please take a look when you've time.

[jeff9finger]

+1

[jeff9finger]

How is the discriminator value determined for the name property of the@JsonSubTypes.@Type annotation. It looks like the best that can be done without more information is to use the class name.

But if a discriminator property value for a sub class is not the name of the class, but another string, how can that be determined? This may need to be addressed in a new issue, but simply wanted to raise it here to see if anyone has ideas.

My idea is to add a codegen vendor extension that can be used to state the discriminator value in a subclass - maybe x-java-discriminator-value

definitions:
  Base:
    type: object
    discriminator: disc
    properties:
      disc:
        type: string
    required:
      - disc
  SubClass:
    x-java-discriminator-value: sub-class
    allOf:
      - $ref: '#/definitions/Base'
        required:
          - field
        properties:
          field:
            type: string 

[szakrewsky]

@jeff9finger I have seen a similar issue. The spec (http://swagger.io/specification/) for the discriminator field says

When used, the value MUST be the name of this schema or any schema that inherits it.

And since the class name is the schema name its all good.

[jeff9finger]

Using the classname is a good default (in fact, it looks to be the only thing that makes sense). But in some cases, the discriminator value needs to be something other than the class name.

I have used a vendor extension "x-discriminator-value" on each subclass. This value contains the value from the enumeration defined for that subclass. Changed the template to handle that value if it exists (otherwise, use name as before). This works perfectly for my. I'd like to see it in the master branch so I don't have to have my own templates.... Anyone else having this issue?

[boldtrn]

Isn't this fixed by #4085?

[wing328]

Isn't this fixed by #4085?

I think so.