Add provider headers to response.

[tomphilpotts]

I am working on creating a Rate Limiter for my project, but the response from Open AI's are not including the headers.

[ChrisB-TL]

There's some relatively new built in support for inspecting rate limits but not for OpenAI just yet.

It would be a great PR if you are up for adding it.

Outside of that we have been talking about making some more request and response data available on responses - possibly in the next few weeks.

[ChrisB-TL]

@tomphilpotts 0.49.0 has just been tagged which now supports OpenAI rate limits for text and structured (see docs page in prior post). Streaming and embeddings to follow soon (see #251).

The docs may not yet have updated when you see this, but nothing has changed other than the section on provider support.