🧠 RAG-based Document Assistant

This is a lightweight Retrieval-Augmented Generation (RAG) app that lets users upload a PDF document and ask questions about its content. It's designed for domain-specific question answering using OpenAI's LLMs and ChromaDB for local vector search.

---

🚀 Features

• Upload any PDF (e.g. academic papers, contracts, manuals)
• Ask natural language questions about the document
• Uses SentenceTransformer embeddings and OpenAI's gpt-3.5-turbo for answers
• Built with Streamlit for quick and interactive UI
• Chunking with overlap to preserve semantic context
• Stores and queries vector representations locally via ChromaDB

---

📦 Dependencies

pip install -r requirements.txt

Project Structure

Running Locally

1. Clone and enter the repo git clone https://github.com/your-username/RAG-Based-Domain-Specific-Q-A-System.git cd RAG-Based-Domain-Specific-Q-A-System

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

3. Install dependencies pip install -r requirements.txt

4.Set up .env file OPENAI_API_KEY=your-openai-api-key-here

5.Run the app streamlit run app/main.py

🧾 Example Questions for a research paper

Once you upload a PDF, try asking:

• "What is the main contribution of this paper?"
• "Which algorithm is used?"
• "What are the limitations discussed?"
• "Explain the methodology"
• "List the datasets used"

💡 Notes

• Chunk size and overlap can be tweaked in processor.py
• Embedding model can be changed in vector_store.py
• Local vector DB (ChromaDB) is reset on reruns unless persisted in vector_db/

📬 Contact

Feel free to reach out if you want to deploy or scale this further.

• Name: Vraj Desai
• Email: vrajdhar@usc.edu
• Contact: +1 2136915656