Skip to content

hassanaiengineer/IntelliDocs

BranchesTags

Repository files navigation

🧠 Enhanced RAG System - IntelliDocs

A production-ready Retrieval-Augmented Generation (RAG) system that transforms your documents into intelligent conversations. Built with modern web technologies and AI capabilities for seamless document analysis.

Python 3.8+ FastAPI License: MIT

IntelliDocs Home

Important

Backend Code Folder: For production ready backend code folder,Follow on GitHub and message me on LinkedIn to receive the production-ready backend code.

✨ Features

🎯 Core Capabilities

  • Progressive File Upload: Upload up to 4 documents (PDF/DOCX) without losing existing files
  • Intelligent Chat: Ask questions and get answers based on your document content
  • Multi-Document Analysis: Query across multiple documents simultaneously
  • Semantic Search: Advanced retrieval with both vector similarity and keyword matching
  • Real-time Responses: Instant AI-powered answers with source attribution

🔧 Technical Features

  • Multiple LLM Providers: OpenAI, Anthropic Claude, Google Gemini support
  • Semantic Chunking: Intelligent text segmentation preserving context
  • Vector Database: ChromaDB for efficient similarity search
  • Session Management: Secure, temporary sessions with no data persistence
  • Caching System: Embedding caching for improved performance
  • Production Ready: Built with FastAPI, proper error handling, and logging

🎨 User Experience

  • Beautiful UI: Modern, responsive design with animations
  • Drag & Drop: Intuitive file upload experience
  • Real-time Feedback: Progress indicators and status updates
  • File Management: Add, remove, and manage documents easily
  • Source Attribution: See which documents provided each answer

🚀 Quick Start

Prerequisites

  • Python 3.8 or higher
  • 4GB+ RAM recommended
  • Modern web browser

Installation

  1. Clone the repository
git clone <repository-url>
cd enhanced_rag
  1. Make startup script executable
chmod +x start.sh
  1. Run the application
./start.sh

The script will:

  • Create a virtual environment
  • Install all dependencies
  • Initialize databases
  • Start the server on http://localhost:8000

Manual Setup (Alternative)

If you prefer manual setup:

# Create virtual environment
python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Start the server
uvicorn main:app --reload

🎮 Usage Guide

1. Welcome & Setup

  1. Open http://localhost:8000 in your browser
  2. Enter your name to create a session
  3. Choose an AI provider (OpenAI, Anthropic, or Gemini)
  4. Enter your API key and test the connection

2. Document Upload

  • Drag & Drop: Drop files onto the upload zone
  • Browse: Click to select files from your computer
  • Progressive Upload: Add new files without removing existing ones
  • Supported Formats: PDF, DOCX, DOC (up to 20MB each)

3. Chat Interface

  • Ask Questions: Type naturally about your documents
  • View Sources: See which documents provided each answer
  • File Management: Remove individual files or clear all
  • Session Limits: Up to 50 questions per session

4. Advanced Features

  • Multi-file Queries: Ask questions that span multiple documents
  • Contextual Answers: Get responses that combine information from different sources
  • Real-time Processing: See upload progress and typing indicators

🏗️ Architecture

Backend Components

backend/
├── api/           # FastAPI routes
│   ├── routes_upload.py    # File upload handling
│   ├── routes_rag.py       # Chat and Q&A
│   ├── routes_validate.py  # Authentication
│   └── routes_files.py     # File management
├── core/          # Core system
│   ├── config.py           # Configuration management
│   └── session.py          # Session handling
├── ingestion/     # Document processing
│   ├── document_processor.py  # PDF/DOCX extraction
│   └── chunker.py             # Semantic chunking
├── embedding/     # Vector embeddings
│   └── embedding.py        # Sentence transformers
├── vector/        # Vector storage
│   ├── vectorstore.py      # ChromaDB interface
│   └── retriever.py        # Advanced retrieval
└── llm/           # LLM integration
    └── provider.py         # Multi-provider support

Frontend Components

frontend/
├── welcome.html   # Landing page
├── auth.html      # API key setup
├── chat.html      # Main interface
└── static/
    ├── css/       # Styled components
    └── js/        # Interactive features

🔧 Configuration

Environment Variables

Create a .env file or modify the generated one:

# Server Settings
HOST=0.0.0.0
PORT=8000
WORKERS=1

# File Upload Limits
MAX_FILES_PER_SESSION=4
MAX_FILE_SIZE_MB=20

# Session Management
MAX_QUESTIONS_PER_SESSION=50
SESSION_TIMEOUT_HOURS=24

# Vector Database
CHROMA_DIR=./chroma_store

# Embedding Model
EMBEDDING_MODEL=all-MiniLM-L6-v2

# Text Processing
CHUNK_SIZE=1000
CHUNK_OVERLAP=200

Supported LLM Providers

OpenAI

  • Models: GPT-3.5 Turbo, GPT-4
  • API Key Format: sk-...
  • Get API Key: OpenAI Platform

Anthropic

  • Models: Claude 3 Haiku, Claude 3 Sonnet
  • API Key Format: sk-ant-...
  • Get API Key: Anthropic Console

Google Gemini

🚀 Production Deployment

Using Docker (Recommended)

FROM python:3.9-slim

WORKDIR /app
COPY requirements.txt .
RUN pip install -r requirements.txt

COPY . .
EXPOSE 8000

CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8000"]

Using Gunicorn

gunicorn main:app -w 4 -k uvicorn.workers.UvicornWorker --bind 0.0.0.0:8000

Environment-Specific Settings

Development

export RELOAD=true
export LOG_LEVEL=debug

Production

export WORKERS=4
export LOG_LEVEL=warning
export SECRET_KEY=your-production-secret
export ALLOWED_ORIGINS=https://yourdomain.com

🧪 Advanced Usage

Custom Embedding Models

Modify backend/core/config.py:

EMBEDDING_MODEL = "sentence-transformers/all-mpnet-base-v2"
# or
EMBEDDING_MODEL = "sentence-transformers/all-distilroberta-v1"

Custom Chunking Strategy

Modify backend/ingestion/chunker.py:

chunker = SemanticChunker(
    chunk_size=1500,      # Larger chunks
    overlap=300,          # More overlap
)

API Integration

import requests

# Upload files
files = {'files': open('document.pdf', 'rb')}
response = requests.post(
    'http://localhost:8000/api/upload/files',
    data={'session_id': 'your-session-id'},
    files=files
)

# Ask questions
response = requests.post(
    'http://localhost:8000/api/rag/ask',
    json={'question': 'What is the main topic?'},
    headers={
        'X-Session-ID': 'your-session-id',
        'X-API-Key': 'your-api-key',
        'X-Provider': 'openai'
    }
)

🛠️ Development

Project Structure

  • main.py - FastAPI application entry point
  • start.sh - Development startup script
  • requirements.txt - Python dependencies
  • backend/ - Server-side logic
  • frontend/ - Client-side interface

Adding New Features

  1. Backend: Add routes in backend/api/
  2. Frontend: Modify HTML/CSS/JS in frontend/
  3. Database: Update models in backend/core/

Running Tests

pytest tests/ -v

📊 Performance

Benchmarks

  • File Processing: ~2-5 seconds per MB
  • Query Response: ~1-3 seconds
  • Memory Usage: ~500MB base + ~100MB per session
  • Concurrent Users: 10+ (single worker)

Optimization Tips

  1. Use smaller embedding models for faster processing
  2. Increase chunk overlap for better retrieval
  3. Use multiple workers for production
  4. Enable response caching for repeated queries

🔒 Security

Data Privacy

  • No Persistent Storage: Files and conversations are session-only
  • API Key Security: Keys stored in browser session only
  • Encrypted Communication: HTTPS recommended for production

Security Headers

# Add to main.py for production
from fastapi.middleware.trustedhost import TrustedHostMiddleware

app.add_middleware(TrustedHostMiddleware, allowed_hosts=["example.com"])

🐛 Troubleshooting

Common Issues

1. Import Errors

# Ensure virtual environment is activated
source venv/bin/activate
pip install -r requirements.txt

2. ChromaDB Issues

# Clear vector database
rm -rf chroma_store/
# Restart application

3. Memory Issues

# Reduce chunk size in config.py
CHUNK_SIZE = 500
# Use smaller embedding model
EMBEDDING_MODEL = "all-MiniLM-L6-v2"

4. File Upload Failures

  • Check file size limits (20MB max)
  • Verify file format (PDF/DOCX only)
  • Ensure sufficient disk space

Debug Mode

export LOG_LEVEL=debug
export RELOAD=true
./start.sh

📝 API Documentation

Once running, visit:

  • Interactive Docs: http://localhost:8000/docs
  • ReDoc: http://localhost:8000/redoc
  • OpenAPI JSON: http://localhost:8000/openapi.json

🤝 Contributing

  1. Fork the repository
  2. Create a feature branch
  3. Make your changes
  4. Add tests
  5. Submit a pull request

Code Style

black backend/ frontend/
flake8 backend/

📄 License

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

🙏 Acknowledgments

  • FastAPI for the excellent web framework
  • ChromaDB for vector storage
  • Sentence Transformers for embeddings
  • Tailwind CSS for beautiful styling
  • OpenAI/Anthropic/Google for AI capabilities

🤝 Let's Connect

If you found this template helpful or want to discuss AI systems, feel free to reach out:

Built with ❤️ for intelligent document analysis

For questions, issues, or contributions, please visit our GitHub repository.

About

A production-ready Retrieval-Augmented Generation (RAG) system that transforms your documents into intelligent conversations.

Resources

Readme
Activity

Stars

3 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

 
 
 

Contributors

Languages