Azure Function App using Vector Search

This project implements two Azure Function APIs:

1. IndexDocuments: Accepts a document link (PDF or DOCX) and indexes it into Azure AI search.
2. QueryKnowledgeBase: Queries the indexed documents and retrieves relevant information using OpenAI's GPT model.

Features

IndexDocuments:

• Reads documents from Azure Blob Storage.
• Splits document content into chunks.
• Generates embeddings using OpenAI's text-embedding-3-small model.
• Indexes chunks into Azure AI search.
• Handles errors and returns a status message.

QueryKnowledgeBase:

• Takes a user query and converts it into a vector embedding.
• Searches through the indexed content using vector search.
• Uses GPT to return relevant information or null if no relevant information is found.

Technologies Used

• Azure Functions
• Azure AI Search
• Azure OpenAI
• Python
• Postman

Environment Variables

Ensure the following environment variables are set:

• AZURE_OPENAI_ENDPOINT: Azure OpenAI API endpoint.
• AZURE_OPENAI_API_KEY: Azure OpenAI API key.
• AZURE_OPENAI_API_VERSION: OpenAI API version.
• AZURE_SEARCH_ENDPOINT: Azure Search API endpoint.
• AZURE_SEARCH_KEY: Azure Search API key.
• SYSTEM_MESSAGE_TEMPLATE: The system message template for GPT responses.

Installation

1. Clone this repository:

   git clone https://github.com/ShashankGowni/Azure-Function-App-using-Vector-search

2. Navigate to the project folder:

   cd Azure-Function-App-using-Vector-search

3. Install the required dependencies:

   pip install -r requirements.txt

4. Set the necessary environment variables:

   • For local development, create a .env file or set them in the terminal.

Running Locally

To run the function app locally, use the Azure Functions Core Tools:

func start

Output

API Usage

1. IndexDocuments API

   • Request Type: POST
   • Endpoint: http://localhost:<port>/api/IndexDocuments
   • Request Body:

     {
       "doc_link": "https://yourblobstorageurl.com/yourfile.pdf"
     }

   • Success Response:

     {
       "status": "COMPLETED",
       "error": null
     }

   • Failure Response:

     {
       "status": "FAILED",
       "error": "Document format is not supported."
     }

2. QueryKnowledgeBase API

   • Request Type: GET
   • Endpoint: http://localhost:<port>/api/QueryKnowledgeBase?query=What%20is%20Azure&index_name=your_index_name
   • Success Response:

     {
       "response": "Azure is a cloud computing service from Microsoft.",
       "error": null
     }

   • Failure Response:

     {
       "response": null,
       "error": null
     }

Output Screenshot (References)

**output in the vscode **

** output in PostMan **