Multiple body-types in an endpoint with same model (referenced via `$ref`) results in ambiguity in generated code

Actual Behavior

When an endpoint has multiple body types, each referencing the same model (via a $ref key), the generator generates the same class name for each body types, which results then in the following generated code:

def _get_kwargs(
    # ...
):
    # ...

    if isinstance(body, WritableJournalEntryRequest):
        _json_body = body.to_dict()
 
        _kwargs["json"] = _json_body
        headers["Content-Type"] = "application/json"
    if isinstance(body, WritableJournalEntryRequest):
        _files_body = body.to_multipart()
 
        _kwargs["files"] = _files_body
        headers["Content-Type"] = "multipart/form-data"
 
    # ...

As you can see, both if-statement's conditions will resolve to True, therefore only the last body type will be used.

Desired Behavior

The body-type should be part of the generated class name to avoid such ambiguities.
Or an if/elseif/else should be used in the template (which would result in the first one being applied ONLY).

OpenAPI Spec File

The OpenAPI spec is from Netbox, but here is the relevant snippet:

"requestBody": {
    "content": {
        "application/json": {
            "schema": {
                "$ref": "#/components/schemas/WritableJournalEntryRequest"
            }
        },
        "multipart/form-data": {
            "schema": {
                "$ref": "#/components/schemas/WritableJournalEntryRequest"
            }
        }
    },
    "required": true
},

先追蹤 generator 如何處理 OpenAPI requestBody 的 content 項目，以及如何產生生成的 _get_kwargs 函式。使用提供的 JSON 片段重現問題，然後為 application/json 和 multipart/form-data 兩者新增 coverage；完成的標準是生成的 client 能選擇正確的 body 處理方式，且沒有含義不明的相同條件。