README.md

metadata

title: RAG System Demo
emoji: 🤗
colorFrom: blue
colorTo: purple
sdk: streamlit
sdk_version: 1.28.1
app_file: app.py
pinned: false

🤗 Enterprise RAG System Demo - Hugging Face Edition

A simplified, demonstration version of the Enterprise RAG System that uses Hugging Face models instead of OpenAI/Anthropic APIs. Perfect for showcasing RAG capabilities and deploying on Hugging Face Spaces for free!

🚀 Live Demo

Try it on Hugging Face Spaces →

✨ What This Demo Shows

This streamlined version demonstrates the complete RAG pipeline of the original Enterprise system:

🔄 Complete RAG Workflow

• Document Upload & Processing - PDF, DOCX, TXT, CSV support
• Text Chunking & Embeddings - Smart text splitting with sentence-transformers
• Vector Search - ChromaDB for semantic similarity search
• AI-Powered Q&A - Hugging Face transformers for response generation
• Source Attribution - Shows which documents were used for answers

🏗️ Enterprise Features (Simplified)

• Multi-format Support - Same document types as the full system
• Hybrid Search - Semantic search with relevance scoring
• Real-time Chat Interface - Interactive Q&A with chat history
• Source Tracking - See which documents contributed to each answer
• Responsive UI - Clean, professional interface built with Streamlit

🔧 Technical Architecture

Models Used (Hugging Face)

• Text Generation: microsoft/DialoGPT-small or distilgpt2 (lightweight)
• Embeddings: sentence-transformers/all-MiniLM-L6-v2 (384 dimensions)
• Vector Store: ChromaDB (in-memory for demo)

Key Differences from Original

Original Enterprise System	This Demo Version
OpenAI/Anthropic APIs	Hugging Face Transformers
PostgreSQL + Redis	In-memory storage
JWT Authentication	No authentication (demo)
Microservices Architecture	Single Streamlit app
Production deployment	Hugging Face Spaces

📦 Installation & Setup

Option 1: Run Locally

# Clone this demo
git clone <repo-url>
cd rag-demo-hf

# Install dependencies
pip install -r requirements.txt

# Run the app
streamlit run app.py

Option 2: Deploy on Hugging Face Spaces

1. Create a new Space on Hugging Face
2. Select Streamlit as the SDK
3. Upload these files:
   • app.py
   • requirements.txt
   • README.md
4. Your demo will be live at https://huggingface.co/spaces/YOUR_USERNAME/SPACE_NAME

🎯 How to Use

1. Upload Documents

• Click "Upload Documents" in the sidebar
• Select PDF, DOCX, TXT, or CSV files
• Click "Process Documents" to add them to the knowledge base

2. Ask Questions

• Type your question in the chat input
• The system will:
  • Search for relevant document chunks
  • Generate an AI-powered response
  • Show source documents used

3. Review Sources

• See relevance scores for each source
• Click on document excerpts to see the content used
• Source attribution shows which documents contributed to answers

📊 Performance & Capabilities

What Works Great

• ✅ Document Processing - Handles multiple formats reliably
• ✅ Semantic Search - Finds relevant content accurately
• ✅ Source Attribution - Clear tracking of information sources
• ✅ User Experience - Intuitive chat interface

Demo Limitations (vs. Full System)

• 🔄 Text Generation - Smaller models, shorter responses
• 💾 Memory - In-memory storage (resets on restart)
• 🔐 Security - No authentication or user management
• 📈 Scale - Optimized for demonstration, not production

🔄 Original vs Demo Comparison

Original Enterprise RAG System

• Purpose: Production-ready enterprise solution
• Models: OpenAI GPT-4, Anthropic Claude
• Infrastructure: PostgreSQL, Redis, Docker
• Features: Authentication, rate limiting, monitoring
• Deployment: Self-hosted or cloud production

This Demo Version

• Purpose: Showcase RAG capabilities and architecture
• Models: Open-source Hugging Face models
• Infrastructure: Streamlit + in-memory storage
• Features: Core RAG functionality only
• Deployment: Hugging Face Spaces (free)

🚀 Key Demo Features

1. Real RAG Pipeline

Document Upload → Text Extraction → Chunking → Embeddings → Vector Store → Search → LLM Response

2. Interactive Experience

• Upload your own documents
• Ask domain-specific questions
• See exactly which sources were used
• Get contextual, relevant answers

3. Educational Value

• See RAG in action with your own data
• Understand the workflow from documents to answers
• Explore source attribution and relevance scoring
• Experience the full user journey

🎯 Perfect For

• 🎓 Learning RAG concepts - See how it works end-to-end
• 💼 Client demonstrations - Show RAG capabilities quickly
• 🔬 Prototyping - Test ideas before building production systems
• 📚 Education - Teach AI and NLP concepts interactively

🔗 Related Projects

• Original Enterprise RAG System - Full production system
• Hugging Face Transformers - Model library used
• ChromaDB - Vector database
• Streamlit - Web framework

📝 License

MIT License - Feel free to use this demo for learning, presentations, or as a starting point for your own RAG systems!

---

🤗 Built with Hugging Face • Powered by Open Source AI

This demo shows that you can build powerful RAG systems using entirely open-source components!