RAG-AGENT

A sophisticated multi-turn Retrieval-Augmented Generation (RAG) agent that combines intelligent document retrieval, web search capabilities, and an intuitive interface for seamless data ingestion and interaction.

📋 Table of Contents

• Overview
• Features
• Architecture
• Tech Stack
• Components
• Getting Started
• Usage
• Configuration

🎯 Overview

RAG-AGENT is an advanced AI-powered system designed to provide accurate, context-aware responses by combining:

• Retrieval-Augmented Generation: Leverages a knowledge base to ground responses in actual documents
• Web Search Integration: Extends knowledge beyond stored documents using Bing Search API
• Multi-turn Conversations: Maintains context across multiple interactions for coherent dialogue
• Document Management: Intelligent indexing and retrieval using Azure AI Search

This system is ideal for customer support, knowledge base querying, research assistance, and intelligent information retrieval tasks.

✨ Features

Core Capabilities

• Multi-turn Conversation Support

  • Maintains conversation history and context
  • Understands follow-up questions with proper context awareness
  • Supports back-and-forth dialogue sessions

• Dual Search Tools

  • Document Search: Query indexed documents via Azure AI Search
  • Web Search: Real-time web search powered by Bing Search API
  • Intelligent tool selection based on query type

• Document Management System

  • Upload and ingest documents (PDF, TXT, DOCX, etc.)
  • Real-time document preview before indexing
  • Automatic chunking and embedding generation
  • Vector-based semantic search

• Intuitive Web Interface

  • Clean, responsive UI for querying
  • Document upload and management dashboard
  • Data preview functionality
  • Conversation history tracking
  • Search result visualization

🏗️ Architecture

┌─────────────────────────────────────────────────────────────┐
│                    User Interface                            │
│            (Web Interface + Document Ingestion)              │
└────────────────────┬────────────────────────────────────────┘
                     │
┌────────────────────▼────────────────────────────────────────┐
│                  RAG Agent Engine                            │
│          (Multi-turn Conversation Manager)                   │
└────────┬───────────────────────┬───────────────────────┬────┘
         │                       │                       │
    ┌────▼─────┐         ┌──────▼──────┐        ┌──────▼──────┐
    │ Document │         │ Bing Search │        │   Context   │
    │  Search  │         │    Tool     │        │   Manager   │
    │(Azure AI │         │             │        │             │
    │ Search)  │         └─────────────┘        └─────────────┘
    └────┬─────┘
         │
┌────────▼────────────────────────────────────────────────────┐
│            Knowledge Base & Vector Store                     │
│           (Azure AI Search + Embeddings)                     │
└──────────────────────────────────────────────────────────────┘

🛠️ Tech Stack

Backend

• Language: Python / Node.js
• LLM: OpenAI GPT-4 / Azure OpenAI
• Vector Database: Azure AI Search
• Web Search: Bing Search API
• Framework: FastAPI / Express.js

Frontend

• Framework: React / Vue.js
• UI Components: Tailwind CSS / Material UI
• State Management: Redux / Pinia
• Real-time Updates: WebSockets

Cloud Infrastructure

• Azure AI Search: Document indexing and semantic search
• Azure OpenAI: Language model and embeddings
• Azure App Service: Backend deployment
• Azure Static Web Apps: Frontend hosting

📦 Components

1. RAG Agent Engine

The core intelligence system that orchestrates:

• Query interpretation and routing
• Tool selection and execution
• Response generation with source attribution
• Conversation state management

2. Document Ingestion Pipeline

Handles the document upload and processing workflow:

• File upload and validation
• Document parsing and chunking
• Embedding generation
• Index creation in Azure AI Search
• Preview generation

3. Search Tools

Document Search

• Queries indexed documents in Azure AI Search
• Performs semantic and keyword search
• Returns ranked results with relevance scores
• Provides source attribution

Web Search

• Integrates Bing Search API
• Retrieves real-time information
• Supplements document knowledge
• Tracks search sources

4. Web Interface

User-facing application providing:

• Query Interface: Multi-turn conversation view
• Document Management: Upload, preview, and delete documents
• Settings Panel: Configuration and preferences
• Results Display: Formatted responses with sources

5. Context Manager

Maintains conversation state:

• Conversation history
• Token management
• Context window optimization
• Session persistence

🚀 Getting Started

Prerequisites

• Python 3.9+ / Node.js 18+
• Azure subscription
• Azure OpenAI deployed instance
• Bing Search API key
• Azure AI Search service

Installation

1. Clone the repository

   git clone https://github.com/yourusername/RAG-AGENT.git
   cd RAG-AGENT

2. Backend Setup

   cd backend
   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate
   pip install -r requirements.txt

3. Frontend Setup

   cd frontend
   npm install

4. Configuration

   • Create .env file with required credentials:

   AZURE_OPENAI_ENDPOINT=your_endpoint
   AZURE_OPENAI_KEY=your_key
   AZURE_SEARCH_ENDPOINT=your_search_endpoint
   AZURE_SEARCH_KEY=your_search_key
   BING_SEARCH_KEY=your_bing_key
   

5. Start the Application

   # Terminal 1: Backend
   cd backend && python main.py
   
   # Terminal 2: Frontend
   cd frontend && npm start

💬 Usage

Query the RAG Agent

1. Open the Web Interface

   • Navigate to http://localhost:3000

2. Ask a Question

   • Type your query in the chat interface
   • The agent automatically selects appropriate tools
   • Review the response and source documents

3. Multi-turn Conversation

   • Follow up with related questions
   • Agent maintains conversation context
   • Reference previous interactions

Ingest Documents

1. Access Document Management

   • Click "Upload Documents" in the interface

2. Preview Documents

   • View document preview before ingestion
   • Verify content and formatting

3. Complete Ingestion

   • Confirm upload
   • Monitor ingestion progress
   • Documents indexed and searchable

⚙️ Configuration

Model Settings

• Temperature, top_p, max_tokens
• System prompts and instructions
• Response formatting

Search Settings

• Document search confidence threshold
• Web search result count
• Result ranking preferences

Tool Settings

• Enable/disable Bing search
• Document search only mode
• Hybrid search configuration

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📧 Support

For issues and questions, please open an issue on GitHub or contact the development team.