📚 RAG Chat: Document-based Question Answering System

Overview

This Streamlit application implements a Retrieval-Augmented Generation (RAG) system for intelligent document-based question answering, enabling users to upload PDFs and interactively query their contents.

🌟 Features

• PDF document upload and processing
• Advanced text chunking and embedding
• Vector storage using Pinecone
• AI-powered question answering with Mistral
• Interactive chat interface

🛠 Technologies Used

• Streamlit
• Pinecone
• LangChain
• Mistral AI
• HuggingFace Embeddings

🚀 How It Works

Document Processing

• Upload PDF files through Streamlit interface
• Extract and chunk text using advanced splitters
• Generate high-dimensional embeddings
• Store vectorized documents in Pinecone index

Question Answering Pipeline

• Retrieve contextually relevant document chunks
• Generate precise answers using Mistral AI
• Provide source document references

📦 Dependencies

streamlit
pinecone-client
langchain
transformers
mistralai

🔧 Configuration

Required API Keys

• Pinecone API Key
• Mistral AI API Key

Embedding Model

• Model: BAAI/bge-large-en-v1.5
• Dimensions: 1024
• Device: CPU/CUDA

💻 Usage Instructions

1. Upload PDF documents
2. Click "Process Documents"
3. Ask questions in chat interface
4. Receive AI-generated answers with source references

🔍 Example Workflow

User uploads research papers ➡️ Documents are chunked and embedded ➡️ User asks: "What are the key findings?" ➡️ AI retrieves relevant sections ➡️ Generates comprehensive answer

🔒 Security Notes

• Secrets managed via Streamlit
• Temporary file handling
• Secure API key management

🚧 Potential Improvements

• Multi-language support
• Enhanced embedding models
• More granular source tracking
• Advanced filtering options

📝 License

[MIT]

👥 Contributors

• [Gauri Sharan]

🙏 Acknowledgements

• Streamlit Community
• Pinecone
• Mistral AI
• LangChain Team