Support making raw requests

[SorenDreano]

The official Python OpenAI client implements a "post()" method: in the SyncAPIClient class, with the OpenAI class being a subclass of SyncAPIClient.
This method is also implemented for the AsyncAPIClient and thus for the AsyncOpenAI class.

This post method is mainly a wrapper around httpx (as self.request is a direct call to httpx), which is a general-purpose HTTP client. It is very convenient to create fine-grained custom requests, port requests to other automated tested tools such as Postman and curl, and overall understand the HTTP requests being made.

def post(
    self,
    path: str,
    *,
    cast_to: Type[ResponseT],
    body: Body | None = None,
    options: RequestOptions = {},
    files: RequestFiles | None = None,
    stream: bool = False,
    stream_cls: type[_StreamT] | None = None,
) -> ResponseT | _StreamT:
    opts = FinalRequestOptions.construct(
        method="post", url=path, json_data=body, files=to_httpx_files(files), **options
    )
    return cast(ResponseT, self.request(cast_to, opts, stream=stream, stream_cls=stream_cls))

To reach feature-parity with the Python client, it would be very convenient to port this method to the Java client. This request is for example used in the vLLM documentation