<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> This repository offers sample code for building AI applications powered by data residing in Azure SQL Database. It targets developers and data scientists aiming to integrate AI capabilities like natural language querying, content analysis, and recommendation systems with their relational data. The samples demonstrate practical workflows and technical concepts, accelerating the development of intelligent, data-driven solutions on Azure.

How It Works

The project integrates Azure SQL Database with Azure AI services (Azure OpenAI, Cognitive Services, Promptflow) and open-source AI frameworks (LangChain, Semantic Kernel, Vanna.AI). Data from Azure SQL DB is processed using AI models for tasks including vector embedding generation, similarity search, natural language to SQL (NL2SQL) translation, and content moderation. This approach enables sophisticated AI features directly on or alongside structured data.

Quick Start & Requirements

Setup instructions vary per sample, directing users to individual READMEs or notebooks; GitHub Codespaces are an option. Key prerequisites include an Azure SQL Database instance and Azure AI service access (e.g., Azure OpenAI API keys). Specific samples may require Python environments with libraries like LangChain, Semantic Kernel, Vanna.AI, FAISS, or SciKit Learn. Representative links include the Vanna.AI sample notebook, the RAG T-SQL sample, and the LangChain samples directory.

Highlighted Details

• Retrieval Augmented Generation (RAG): End-to-end RAG patterns for chatbots and content finders using Azure SQL DB data with OpenAI and LangChain.
• Vector Search & Embeddings: Showcases similarity search via OpenAI embeddings and libraries like FAISS, plus Redis Vector Search integration.
• Natural Language to SQL (NL2SQL): Features samples using Vanna.AI and LangChain for AI-driven SQL query generation from natural language.
• AI-Powered T-SQL Development: Examples using prompts with Generative AI models (e.g., GPT-4) to develop and test T-SQL code.
• Content Safety & PII Analysis: T-SQL scripts leveraging Azure OpenAI for text content moderation and PII detection/redaction.

Maintenance & Community

Officially maintained by Microsoft as part of Azure Samples. Specific community channels (e.g., Discord, Slack) or a public roadmap are not detailed in the README.

Licensing & Compatibility

The README does not explicitly state a software license for the code samples, referencing only Microsoft's trademark and brand guidelines. Commercial use or closed-source integration compatibility is unspecified.

Limitations & Caveats

Some AI features, like Azure Cognitive Search's automatic chunking/vectorization, are in preview. Setup complexity and dependencies vary significantly across samples, requiring users to consult individual documentation. The absence of a clear code license may pose adoption challenges.