RAG-Chatbot

This repository contains a Retrieval-Augmented Generation (RAG) chatbot designed to assist users by answering questions based on the context of uploaded documents. It combines state-of-the-art technologies for document processing, embeddings generation, vector search, and natural language generation.

Features

• OCR and PDF Parsing: Extracts text from scanned or text-based PDFs using PyPDFLoader and pytesseract (OCR).
• Document Embeddings: Uses SentenceTransformer for generating embeddings of documents and queries.
• Vector Search: Implements FAISS for storing and retrieving document embeddings with similarity search.
• Query Refinement: An LLM (llama3.2) is used to refine user queries for better retrieval performance.
• Customizable Prompt Templates: Prompts can be tailored for context-specific assistance.
• Streamlit Interface: Allows Admin to upload PDFs, process them, and update the vector store.
• FastAPI Backend: Provides REST APIs for asking questions and refining queries.
• Sources for Answers: Responses include document names and page numbers for better traceability.

Technologies Used

Python Libraries

• LangChain: For document loaders, prompts, and retrieval-based question answering.
• SentenceTransformers: For generating embeddings.
• FAISS: For efficient similarity search.
• pytesseract: For OCR text extraction.
• FastAPI: For creating the API server.
• Streamlit: For creating the Admin interface.

Frontend

• Next.js: For building the frontend of the web application, providing a dynamic and interactive user interface.

Model:

• dunzhang/stella_en_1.5B_v5: for embeddings.
• llama-3.2: for query refinement and responses.

Clone the Repository

1. To clone the repository, run the following commands in your terminal:

git clone https://github.com/Shridhar7-8/RAG-Chatbot.git
cd rag-chatbot

2. Set up a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

Usage Instructions

Go to admin folder

1. Start the Streamlit app:

streamlit run admin.py

2. Open the app in your browser at http://localhost:8501.
3. Upload PDF files and update the vector store.
4. Once the vector store creation is complete, stop the script by exiting (Ctrl+C).

Start the FastAPI Backend

1. Navigate to the fast_server folder inside admin.

cd fast_server

2. Run main.py to start the FastAPI server.

Start the Frontend Chatbot

1. open a new terminal and Navigate to the chatbot directory.

cd chatbot

2. Install dependencies (if not already installed).

npm install

3. Start the development server.

npm run dev

4. Open the chatbot interface in your browser. The default URL is usually http://localhost:3000