This repository provides a sample for building an enterprise chat copilot using Azure AI Studio. It targets developers and researchers looking to ground AI responses in custom data and APIs, offering a starting point for customized AI solutions.

How It Works

The sample leverages Azure AI services to create a RAG (Retrieval-Augmented Generation) pipeline. It involves building an Azure AI Search index with custom data, then using Azure OpenAI models for chat completions and embeddings. The core logic is orchestrated via Python scripts, with options to integrate with Promptflow or LangChain for different development workflows.

Quick Start & Requirements

• Install: pip install -r requirements.txt (after cloning and setting up a Python virtual environment).
• Prerequisites: Azure subscription, deployed Azure OpenAI chat and embedding models, Azure AI Search resource and index. Requires Python 3.x.
• Environment: GitHub Codespaces or a pre-configured VS Code development environment is recommended. Local setup involves cloning the repo and activating a virtual environment.
• Resources: Links to Azure AI Studio for resource creation and configuration are provided.

Highlighted Details

• Supports multiple implementation frameworks: Azure AI SDK, Promptflow, and LangChain.
• Includes steps for building a custom search index from local data files.
• Provides commands for running single questions, evaluating copilot performance, and deploying to an online endpoint.
• Offers guidance on customizing the development container for persistent package installations.

Maintenance & Community

This repository is deprecated and no longer maintained. A new repository for the latest Azure AI code-first experiences is linked: https://github.com/Azure-Samples/rag-data-openai-python-promptflow/tree/main. The project follows the Microsoft Open Source Code of Conduct.

Licensing & Compatibility

The repository does not explicitly state a license in the provided README text. Compatibility for commercial use or closed-source linking would require clarification on licensing terms.

Limitations & Caveats

The repository is marked as deprecated and no longer maintained. Features used are in preview and not recommended for production workloads. The quality and suitability of copilot responses are not guaranteed and require user validation. A RuntimeError warning during execution can be safely ignored.