Example to use a single AAD token for multiple Azure OpenAI resources

azure_openai_aad_access.py

import os
import requests
from datetime import datetime
from azure.identity import DefaultAzureCredential, ChainedTokenCredential, ManagedIdentityCredential

# see https://learn.microsoft.com/en-us/azure/developer/python/sdk/authentication-on-premises-apps?tabs=azure-portal
# this Service Principal has the Cognitive Services OpenAI User role on the AOAI resources
os.environ["AZURE_CLIENT_ID"] = "xxxxx"
os.environ["AZURE_TENANT_ID"] = "xxxxx"
os.environ["AZURE_CLIENT_SECRET"] = "xxxxx"

# Define strategy which potential authentication methods should be tried to gain an access token
credential = ChainedTokenCredential(ManagedIdentityCredential(), DefaultAzureCredential())
access_token = credential.get_token("https://cognitiveservices.azure.com/.default")

# convert unix timestamp access_token.expires_on to datetime
print(datetime.fromtimestamp(access_token.expires_on))

token = access_token.token

endpoints = [
    "https://xxxxx.openai.azure.com/",
    "https://xxxxx.openai.azure.com/"
]

API_VERSION = "2023-05-15"
DEPLOYMENT_NAME = "gpt-35-turbo"

headers = {
    "Content-Type": "application/json",
    "Authorization": f"Bearer {token}"
}

data = {"max_tokens": 1, "messages":[{"role": "system", "content": ""},{"role": "user", "content": "Hi"}]}

for endpoint in endpoints:
    url = f"{endpoint}/openai/deployments/{DEPLOYMENT_NAME}/chat/completions?api-version={API_VERSION}"
    response = requests.post(url, headers=headers, json=data)
    print(response.json())