Boost Your Chatbot with Azure Open AI Search

Table of Contents

1. Introduction
2. Setting Up the Azure Open AI Search API
3. Creating a Web Application
4. Calling the API from a Custom Application
5. Using GitHub Code Spaces for Faster Development
6. Creating a Simple Chatbot
7. Modifying the Post Request
8. Extracting Text with JSON Objects
9. Making API Calls and Receiving Responses
10. Deploying the Application to Azure App Service
11. Exploring the APIs and Backend Code
12. Testing the Application

Introduction

In this article, we will explore how to set up and use the Azure Open AI Search API to Create a web application. We will then learn how to call the API from a custom application and make use of GitHub Code Spaces for faster development. Additionally, we will create a simple chatbot and modify the post request to enhance its functionality. We'll also learn how to extract text using JSON objects and make API calls to receive responses. Finally, we will deploy the application to Azure App Service and explore the APIs and backend code. Let's get started!

Setting Up the Azure Open AI Search API

To begin, we need to set up the Azure Open AI Search API. We'll start by creating a web application that will generate a sample UI. Our task is to call the API that this web application is using from our custom applications. We can use GitHub Code Spaces to speed up our development process and access an online VS Code editor. By creating a virtual machine assigned to our GitHub account, we can run all the necessary commands and set up our development environment quickly. Once the web application is set up, we can proceed to call the API.

Creating a Web Application

Before we can call the API, we need to create a web application. The web application will be generated by the Azure Open AI Search API and provide us with a user interface to test the solution. We can modify the application as per our requirements and make use of the API it is calling in our custom applications. The API can be called from any application, as it is a public API. Once the web application is ready, we can move on to the next step.

Calling the API from a Custom Application

Now, it's time to call the API from our custom application. We already have our source code, so we can modify it to include the API call. We'll create a simple chatbot that sends a message to a method and then makes the API call. We can use a post request to send a URL and content to the API and receive a response in return. By extracting the text from the JSON object, we can return it as a STRING. With these modifications, our custom application is ready to call the API.

Using GitHub Code Spaces for Faster Development

GitHub Code Spaces is an excellent tool for fast and efficient development. It allows us to create a virtual machine assigned to our GitHub account, which runs all the necessary commands and sets up our development environment. By accessing an online VS Code editor, we can quickly modify our source code and test our application. GitHub Code Spaces is a convenient option, especially when we don't have access to a local development environment. Let's give it a try!

Creating a Simple Chatbot

A chatbot is a great addition to any application. In this section, we will create a simple chatbot that responds to user messages. We'll modify our existing code to include the chatbot functionality. Whenever a message is received, we'll pass it to a method to call the API and retrieve a response. By using the JSON object to extract the text, we can return the response as a string. With these changes, our chatbot is ready to Interact with users.

Modifying the Post Request

To improve the functionality of our chatbot, we'll modify the post request. The URL and content parameters will be sent with the API call. We'll also update the return Type to ensure that the response is returned as a string. Instead of using a model class, we'll directly extract the text from the JSON object, simplifying the process. With these changes, our chatbot will be more efficient and provide accurate responses.

Extracting Text with JSON Objects

When making API calls, it is essential to extract the required information from the JSON objects. In this section, we'll learn how to extract text from JSON objects and use it in our application. We'll use a JSON object to pull the required information and return it as a string. By understanding how the API call is made and how the response is received, we can efficiently extract the text and use it as needed.

Making API Calls and Receiving Responses

With our modifications in place, we can now make API calls and receive responses. We'll use the chat endpoint provided by the Azure Open AI Search API to send queries and retrieve responses. By preparing the JSON body with the required parameters, we can make accurate API calls. We can test our API calls using the inspect tool in the browser and ensure that the JSON body is correctly received on the backend side. With these steps, we'll be able to interact with the API and receive the desired responses.

Deploying the Application to Azure App Service

Once our application is ready, we can deploy it to Azure App Service for wider accessibility. We'll need to create the necessary resource groups, app services, and storage accounts. The Azure Open AI Search API and other services will be deployed to the app service. During the deployment process, we'll receive status updates on the creation of the required resources. Once the deployment is complete, we can access the deployed solution through the Azure app service URL.

Exploring the APIs and Backend Code

To understand the workings of our application in more Detail, let's explore the APIs and backend code. In the frontend source code, we can find the APIs for the DaVinci model and chat GPT model. These APIs provide different functionalities and responses. On the backend side, we have the app.py file, which handles the API calls and retrieves the responses. By exploring this code, we can gain insights into how the API calls are handled and how the responses are generated.

Testing the Application

With our application fully set up and deployed, it's time to put it to the test. We can use the ask endpoint to send queries and receive responses specific to the Context. We can input different questions and analyze the responses generated by the APIs. By comparing the responses to those obtained from the web application, we can evaluate the accuracy of our application. Testing is crucial to ensure that our application functions as expected and meets the requirements of our users.