RAG-chatbot

A Retrieval-Augmented Generation (RAG) based chatbot that answers questions using information from your PDF documents and displays the source of the response (PDF name and page number).

---

Setup

1️⃣ Install Dependencies

pip install -r requirements.txt

2️⃣ Set Environment Variable

Create a .env file in the root directory:

# Cloud LLM API Key
GEMINI_API_KEY=your_gemini_api_key # Provide your cloud provider's API key

---

Steps to Run

1️⃣ Ingest Documents

Create a folder named pdfs in the root directory (if it does not exist). Place all PDFs inside the pdfs/ folder, then run:

python 1_ingest.py

This will:

• Load all PDFs
• Split text into chunks
• Generate embeddings
• Store them in the vector database

2️⃣ Run Chatbot UI

streamlit run 2_streamlit_app.py

This will:

• Launch the Streamlit chatbot interface
• Internally use rag.py for retrieval and response generation
• Display the source of each answer (PDF name and page number)

---

🖥️ Streamlit Demo

Part 1

Part 2

---

Notes

• Make sure ingestion is completed before running the chatbot.
• Add new PDFs to the pdfs/ folder and rerun 1_ingest.py to update embeddings.
• Embeddings are currently stored locally in the root directory. The storage layer can be replaced with any vector database (FAISS, Chroma, Pinecone, etc.).
• The project currently uses Gemini via GEMINI_API_KEY. The cloud model provider can be changed as needed.

---