📘 RAG-Based PDF Chatbot (FAISS + LLaMA 3)

A Retrieval-Augmented Generation (RAG) chatbot that answers questions from PDF documents using semantic search and large language models.
Built from scratch using FAISS, Sentence Transformers, and Groq’s LLaMA-3.1 model.

This project is designed for internship and portfolio use, demonstrating real-world AI engineering skills.

---

🚀 Features

• 📄 Ingests one or multiple PDF documents
• ✂️ Smart text chunking with overlap
• 🔎 Semantic search using FAISS vector database
• 🧠 Context-aware answers using LLaMA-3.1
• 💬 Chat-style UI with conversation history
• 📌 Uses PDF context first, with intelligent fallback
• 🧾 (Optional) Source-aware responses

---

🧠 Architecture Overview

PDF Documents ↓ Text Extraction (PyPDF) ↓ Chunking with Overlap ↓ Embeddings (Sentence Transformers) ↓ FAISS Vector Database ↓ Top-K Context Retrieval + Reranking ↓ LLaMA-3.1 (Groq API) ↓ Final Answer (Chat UI)

---

🛠️ Tech Stack

• Programming Language: Python 3.13
• Embeddings: Sentence Transformers (all-MiniLM-L6-v2)
• Vector Database: FAISS
• LLM: LLaMA-3.1 via Groq API
• Frontend: Streamlit
• PDF Parsing: PyPDF

---

📂 Project Structure

RAGChatBot/ ├── backend/ │ ├── load_pdf.py # PDF text extraction │ ├── chunker.py # Text chunking logic │ ├── vector_store.py # FAISS index creation │ ├── rag_answer.py # Final RAG pipeline │ └── test_rag.py # Backend testing ├── data/ │ └── pdfs/ # Input PDF files ├── faiss_index/ │ ├── index.faiss # FAISS vector index │ └── chunks.txt # Stored text chunks ├── frontend/ │ └── app.py # Streamlit chat UI ├── requirements.txt └── README.md

---

⚙️ How to Run Locally

1️⃣ Clone the repository

git clone https://github.com/your-username/RAGChatBot.git
cd RAGChatBot


2️⃣ Create and activate virtual environment
python -m venv venv
venv\Scripts\activate

3️⃣ Install dependencies
pip install -r requirements.txt

4️⃣ Build the FAISS vector index
python backend/vector_store.py

5️⃣ Run the chatbot UI
python -m streamlit run frontend/app.py