Complete Azure OpenAI Service Tutorial: Enterprise AI with GPT Models

Azure OpenAI Service provides REST API access to OpenAI's powerful language models including GPT-4, GPT-3.5-Turbo, and embedding models. It combines OpenAI's capabilities with Azure's enterprise security and compliance.

Why Azure OpenAI?

• Enterprise security: Azure security and compliance
• Data privacy: Your data stays in your Azure subscription
• Regional availability: Deploy in multiple Azure regions
• Integration: Native Azure service integration
• Responsible AI: Built-in content filtering

• GPT-4 and GPT-4 Turbo
• GPT-3.5-Turbo
• DALL-E 3
• Embeddings (text-embedding-ada-002)
• Whisper

Prerequisites

pip install openai azure-identity

Azure CLI
az login

Setup

1. Create Azure OpenAI Resource

from azure.mgmt.cognitiveservices import CognitiveServicesManagementClient
from azure.identity import DefaultAzureCredential

credential = DefaultAzureCredential()

client = CognitiveServicesManagementClient(
    credential=credential,
    subscriptionid="your-subscription-id"

)

Create resource
resource = client.accounts.begincreate(
    resourcegroupname="my-resource-group",
    accountname="my-openai-resource",

    account={
        "location": "eastus",
        "kind": "OpenAI",
        "sku": {"name": "S0"},
        "properties": {}
    }
).result()

print(f"Resource created: {resource.name}")

# Deploy GPT-4 model
deployment = client.deployments.begincreateorupdate(
    resourcegroupname="my-resource-group",
    accountname="my-openai-resource",

    deploymentname="gpt-4-deployment",
    deployment={
        "sku": {"name": "Standard", "capacity": 10},
        "properties": {
            "model": {
                "format": "OpenAI",
                "name": "gpt-4",
                "version": "0613"
            }
        }
    }
).result()

print(f"Deployment created: {deployment.name}")

3. Connect to Azure OpenAI

from openai import AzureOpenAI

client = AzureOpenAI(
    apikey="your-api-key",

    apiversion="2024-02-01",
    azureendpoint="https://my-openai-resource.openai.azure.com"

)

Or use Azure Identity
from azure.identity import DefaultAzureCredential, getbearertokenprovider

tokenprovider = getbearertokenprovider(
    DefaultAzureCredential(),
    "https://cognitiveservices.azure.com/.default"
)

client = AzureOpenAI(
    azureadtokenprovider=tokenprovider,
    apiversion="2024-02-01",

    azureendpoint="https://my-openai-resource.openai.azure.com"
)

Chat Completions

1. Basic Chat

response = client.chat.completions.create(
    model="gpt-4-deployment",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "What is machine learning?"}
    ]
)

print(response.choices[0].message.content)

2. Multi-turn Conversation

class ChatBot:
    def init(self, client, deploymentname, systemprompt):
        self.client = client
        self.deploymentname = deploymentname
        self.messages = [{"role": "system", "content": systemprompt}]


    def chat(self, usermessage):
        self.messages.append({"role": "user", "content": usermessage})


        response = self.client.chat.completions.create(
            model=self.deploymentname,
            messages=self.messages,
            temperature=0.7,
            maxtokens=1000

        )

        assistantmessage = response.choices[0].message.content
        self.messages.append({"role": "assistant", "content": assistantmessage})


        return assistantmessage