Using Azure OpenAI with Python: A Step-by-Step Guide

Azure OpenAI offers powerful tools for integrating AI capabilities into your applications. In this blog post, I’ll guide you through setting up and using Azure OpenAI with Python. We will walk through the code snippet provided, explaining each part to help you understand and implement it effectively.

Prerequisites

Before you start, make sure you have:

1. An Azure account with OpenAI service enabled.
2. The openai Python package installed. You can install it using pip:

pip install openai

Detailed Explanation

Imports and Setup

import os
from openai import AzureOpenAI

• os module is used for interacting with the operating system. However, in this code snippet, it’s not explicitly used.
• AzureOpenAI is imported from the openai library to interact with Azure's OpenAI service.

Configuration

endpoint = "Your_Endpoint"  # https://xyz.azure.com/ 
key = "Your_Key"  # Your API key 
model_name = "gpt-4o"  # Your model name

• endpoint: Replace "Your_Endpoint" with the endpoint URL of your Azure OpenAI resource. It should look something like https://<your-resource-name>.openai.azure.com/.
• key: Replace "Your_Key" with your actual API key.
• model_name: Specify the model you want to use. Here, "gpt-4o" is the chosen model. Make sure it matches the model name you have access to.

Creating the Client

client = AzureOpenAI(azure_endpoint=endpoint,api_version="2024-02-01",api_key=key)

• An instance of AzureOpenAI is created with the specified endpoint, API version, and API key.

Making a Request

completion = client.chat.completions.create(
model=model_name,
messages=[{"role": "user",
    "content": "What is AI?",  # Your question can go here
  },], 
)

• The client.chat.completions.create method is called to generate a completion.
• model: The model you want to use (model_name).
• messages: A list of message objects. Each message is a dictionary with role (user or assistant) and content(the text of the message).