Month: November 2023

Setting up Azure OpenAI with PowerShell

If haven’t been living under a rock, you know that Azure OpenAI is a powerful tool that brings the cutting-edge capabilities of OpenAI’s models to the cloud, offering scalability, reliability, and integration with Azure’s vast ecosystem.

Because I am who I am we will use PowerShell to setup our Azure OpenAI instance. Whether you’re automating deployment or integrating Azure OpenAI into your existing infrastructure, PowerShell scripts can simplify the process. Let’s get started with a step-by-step guide to setting up your Azure OpenAI instance using PowerShell.

Prerequisites

Before we dive into the commands, ensure you have the following:

  • An Azure subscription. If you don’t have one, you can create a free account.
  • PowerShell installed on your system. If you’re on Windows, you’re probably already set. For Mac and Linux users, check out PowerShell Core.
  • The Azure PowerShell module installed. You can install it by running Install-Module -Name Az -AllowClobber -Scope CurrentUser in your PowerShell terminal.

Step 1: Log in to Azure

First things first, let’s log into Azure. Open your PowerShell terminal and run:

Connect-AzAccount

This command opens a login window where you can enter your Azure credentials. Once authenticated, you’re ready to proceed.

Step 2: Create a Resource Group

Azure OpenAI instances need to reside in a resource group, a container that holds related resources for an Azure solution. To create a new resource group, use:

New-AzResourceGroup -Name 'MyResourceGroup' -Location 'EastUS'

Replace 'MyResourceGroup' with your desired resource group name and 'EastUS' with your preferred location.

Step 3: Register the OpenAI Resource Provider

Before deploying Azure OpenAI, ensure your subscription is registered to use the OpenAI resource provider. Register it with:

powershell

Register-AzResourceProvider -ProviderNamespace 'Microsoft.OpenAI'

This command might take a few minutes. To check the status, you can run Get-AzResourceProvider -ProviderNamespace 'Microsoft.OpenAI'.

Step 4: Create an Azure OpenAI Instance

Now, the exciting part—creating your Azure OpenAI instance. Use the following command:

powershell

New-AzResource -ResourceGroupName 'MyResourceGroup' -ResourceType 'Microsoft.OpenAI/workspaces' -Name 'MyOpenAIInstance' -Location 'EastUS' -PropertyObject @{ sku = 'S0'; properties = @{ description = 'My Azure OpenAI instance for cool AI projects'; } }

Make sure to replace 'MyResourceGroup', 'MyOpenAIInstance', and 'EastUS' with your resource group name, desired OpenAI instance name, and location, respectively.

Step 5: Confirm Your Azure OpenAI Instance

To ensure everything went smoothly, you can list all OpenAI instances in your resource group:

powershell

Get-AzResource -ResourceGroupName 'MyResourceGroup' -ResourceType 'Microsoft.OpenAI/workspaces'

This command returns details about the OpenAI instances in your specified resource group, confirming the successful creation of your instance. Enjoy your brand new OpenAI instance!

Quick Dive: Integrating Logic Apps with Azure OpenAI

Let’s cut to the chase: Integrating Azure Logic Apps with Azure OpenAI unlocks a plethora of possibilities, from automating content creation to enhancing data analysis. Below is a step-by-step guide to melding these powerful tools.

Step 1: Set Up Azure OpenAI

First, you need an Azure OpenAI service instance. Go to the Azure Portal, search for Azure OpenAI Service, and create a new instance. Once deployed, grab your API key and endpoint URL from the resource management section.

Step 2: Create Your Logic App

Navigate back to the Azure Portal and create a new Logic App:

  • Choose your subscription and resource group.
  • Pick a region close to you for lower latency.
  • Name your Logic App.
  • Click “Review + create” and then “Create” after validation passes.

Step 3: Design Your Logic App Workflow

Once your Logic App is ready, it’s time to design the workflow:

  • Open your Logic App in the Azure Portal and go to the Logic App Designer.
  • Start with a common trigger like “When an HTTP request is received” if you want your Logic App to act based on external requests.
  • Add a new step by searching for “HTTP” in the actions list and choose the “HTTP – HTTP” action. This will be used to call the Azure OpenAI API.

Step 4: Configure the HTTP Action for Azure OpenAI

  • Method: POST
  • URI: Enter the endpoint URL of your Azure OpenAI service.
  • Headers: Add two headers:
    • Content-Type with the value application/json
    • Authorization with the value Bearer <Your Azure OpenAI API Key>
  • Body: Craft the JSON payload according to your task. For example, to generate text, your body might look like this:
{
  "prompt": "Write a brief about integrating Azure OpenAI with Logic Apps.",
  "temperature": 0.7,
  "max_tokens": 100
}

Step 5: Process the Response

After calling the Azure OpenAI API, you’ll want to handle the response:

  • Add a “Parse JSON” action to interpret the API response.
  • In the “Content” box, select the body of the HTTP action.
  • Define the schema based on the Azure OpenAI response format. For text generation, you’ll focus on extracting the generated text from the response.

Step 6: Add Final Actions

Decide what to do with the Azure OpenAI’s response. You could:

  • Send an email with the generated content.
  • Save the response to a database or a file in Azure Blob Storage.
  • Respond to the initial HTTP request with the generated content.

Step 7: Test Your Logic App

  • Save your Logic App and run a test by triggering it based on your chosen trigger method.
  • Monitor the run in the “Overview” section of your Logic App to ensure everything executes as expected.