📚 RAG Document Explainer

A RAG-based document analysis application that allows users to upload a PDF, ask questions about its content, and receive grounded answers based on relevant document sections with source references and page numbers.

🚀 Features

• 📄 Upload and extract text from PDF documents while preserving page numbers
• ✂️ Split documents into smaller overlapping chunks
• 🔢 Generate semantic embeddings using Sentence Transformers
• 🔎 Perform semantic similarity search using FAISS
• 🤖 Generate grounded answers using the Google Gemini API
• 📌 Display relevant document excerpts with source page references
• 🛡️ Reduce unsupported answers by grounding responses in retrieved document context
• 🔍 Analyze documents to identify important sections and potentially notable clauses

🛠️ Technologies Used

• Python
• Streamlit
• Google Gemini API
• Sentence Transformers
• FAISS
• PyMuPDF

📁 Project Structure

rag-document-explainer/
│
├── app.py
├── .env
├── .gitignore
├── README.md
├── requirements.txt
└── utils/
    ├── pdf_reader.py
    ├── chunker.py
    ├── vector_store.py
    └── gemini_client.py

📁 Future Enhancements

• Support for multiple document uploads
• Conversational memory across questions
• Hybrid keyword + semantic retrieval
• Retrieval quality evaluation
• Support for additional document formats
• Improved document-level analysis
• Advanced retrieval techniques such as reranking