[feature] Surface apim-request-id

[iseabock]

Motivation and Context

Surface apim-request-id for better debugging in streaming and non-streaming responses

Description

Added a headers key with a dictionary containing the key/value for apim-request-id. Currently, this is viewable in the result object in the frontend. The task however, requires that this value be in the HAR file. I'm still trying to get that to show up.

Contribution Checklist

• I have built and tested the code locally and in a deployed app
• For frontend changes, I have pulled the latest code from main, built the frontend, and committed all static files.
• This is a change for all users of this app. No code or asset is specific to my use case or my organization.
• I didn't break any existing functionality 😄

[sarah-widder]

@iseabock what about non-streaming?

[iseabock]

what about non-streaming?

I'm not sure how to retrieve the apim-request-id for non-streaming. For some reason, send_chat_request in complete_chat_request doesn't return a header.

[abhahn]

what about non-streaming?

I'm not sure how to retrieve the apim-request-id for non-streaming. For some reason, send_chat_request in complete_chat_request doesn't return a header.

what about non-streaming?

I'm not sure how to retrieve the apim-request-id for non-streaming. For some reason, send_chat_request in complete_chat_request doesn't return a header.

I think the problem with the non-streaming response is that it is being returned as a plain ChatCompletion object, which doesn't include the raw response.

I got some inspiration from this StackOverflow thread to try using with_raw_response, and running in the debugger found that the response from that does have the headers (including apim-request-id), but the content would need to be deserialized to ChatCompletion before returning it. I'm not sure what other implications there are (particularly for streaming) but maybe you could try it out to see? https://stackoverflow.com/questions/77605691/how-to-get-headers-from-openai-api