This repository provides code samples for implementing vector search capabilities within Azure AI Search. It targets developers working with Python, C#, REST, and JavaScript, offering practical examples for integrating vector embeddings into search applications, particularly for AI-driven scenarios like Retrieval-Augmented Generation (RAG).

How It Works

The samples demonstrate various approaches to vector search, including direct API calls for creating and querying vector indexes, as well as integrated solutions where Azure OpenAI is used for generating embeddings during both indexing and query time. This integrated approach simplifies the workflow by automating the vectorization process within Azure AI Search's indexing pipeline. Some samples also explore performance optimizations like quantization and selective vector storage.

Quick Start & Requirements

• Install/Run: Samples are provided as code (Python notebooks, .NET console apps, Java console apps, JavaScript files). Execution typically involves setting up an Azure AI Search service and an Azure OpenAI service.
• Prerequisites: Azure subscription, Azure AI Search service, Azure OpenAI service, relevant SDKs (Python, .NET, Java, JavaScript), and potentially Azure Storage.
• Resources: Requires configuration of Azure services. Links to official documentation for Azure AI Search and Azure OpenAI are provided.

Highlighted Details

• Demonstrates vector search with text and image content.
• Includes integrated vectorization using Azure OpenAI within the indexing pipeline.
• Explores quantization and storage options for optimizing vector index size.
• Features end-to-end RAG solutions and a ChatGPT retrieval plugin.

Maintenance & Community

This repository is maintained by Azure. Links to Azure AI Search and Azure OpenAI documentation are provided.

Licensing & Compatibility

Samples are typically provided under a permissive license (e.g., MIT), but users must adhere to the terms of use for Azure services. Compatibility is focused on the Azure ecosystem.

Limitations & Caveats

The repository notes breaking changes between REST API versions (e.g., 2023-07-01-Preview to newer versions), which also affect SDK beta packages. Users should consult migration guidance for these changes. Preview features are subject to supplemental terms.