enable azure for audio/whisper

[kristapratico]

Description

The Audio/Whisper APIs in the openai library were implemented to work specifically with openai endpoints, e.g. taking a positional param for model. This PR shows how we might change the code so we can use deployment_id with Azure, but preserve the positional parameter behavior for model, file, filename, etc.

Usage

import openai

openai.api_base = "<endpoint>"
openai.api_key = "<api-key>"
openai.api_type = "azure"
openai.api_version = "2023-09-01-preview"

transcribed = openai.Audio.transcribe(
    deployment_id="whisper-1",
    file=open("recording.m4a", "rb")
)

translated = openai.Audio.translate(
    deployment_id="whisper-1",
    file=open("recording.m4a", "rb")
)

Intellisense:

Intellisense shows both options for calling the APIs.