[Java] Support enums in query params

[cbornet]

In Java generation, enums are generated when they are part of a definition object but not when they are a direct method property.
For instance:

"/pet/findByStatus": {
      "get": {
        "tags": [
          "pet"
        ],
        "summary": "Finds Pets by status",
        "description": "Finds Pets by status",
        "operationId": "findPetsByStatus",
        "produces": [
          "application/json",
          "application/xml"
        ],
        "parameters": [
          {
            "name": "status",
            "in": "query",
            "description": "Status value that for filter",
             "type": "string",
             "enum": ["available", "pending", "sold"]
            }
         ],
        "responses": {
          "200": {
            "description": "successful operation",
            "schema": {
              "type": "array",
              "items": {
                "$ref": "#/definitions/Pet"
              }
            }
          }
        }

Generates "status" as a String not as an enum.

[fehguy]

This can be tricky. When parameters are passed and mapped directly to enums, any name mismatching will cause big issues on the server side, typically things that you would need to handle on your own unless you're using some Jackson features for deserializing into enum properties

It is valid to expect, and probably worth keeping open and implementing, but it can have some negative side effects.

[vivin]

@fehguy This is an issue I am running into as well. It is tricky, to be sure! I'm using the Feign library and have a template that generates a client that works with Feign. In Feign you can specify custom "expanders" (basically serializers) to convert an arbitrary object into an representation that is query-string safe.

Currently it appears that the model built up by swagger-codegen ignores the enum information when building up a CodegenParameter instance. This is especially the case when the query parameter is an array of enums. It appears that all the data exposed by swagger is correctly recognized, but it is not attached to the CodegenParameter instance. For example, I have the following schema that describes a parameter:

{
    "collectionFormat": "multi",
    "enum": [
        "UNKNOWN",
        "CUSTOMER",
        "LEAD",
        "PARTNER",
        "COWORKER",
        "PERSONAL",
        "SYSTEM",
        "VISITOR",
        "ANONYMOUS",
        "VENDOR"
    ],
    "in": "query",
    "items": {
        "enum": [
            "UNKNOWN",
            "CUSTOMER",
            "LEAD",
            "PARTNER",
            "COWORKER",
            "PERSONAL",
            "SYSTEM",
            "VISITOR",
            "ANONYMOUS",
            "VENDOR"
        ],
        "type": "string"
    },
    "name": "classifications",
    "required": false,
    "type": "array"
}

Currently I just end up with the type as List<String> (which isn't necessarily incorrect), but I would like to have the model include the type information for enums so that I can generate my client as I see fit.

I'm not too familiar with the code, but I don't mind taking a stab at this - it might amount to simply attaching the data that swagger-codegen already finds, to the CodegenParameter object. This way we at least have access to the information.

Let me know what you think; thanks!

[vivin]

@wing328 Any thoughts on this? I could submit a PR for this too. I'll see if I can take a first pass at this sometime this week.

[wing328]

@vivin sorry that I missed your message. Yes, please submit a PR if you've time and we'll review accordingly.