rage-pdf: RAG-powered PDF Q&A System

This project implements a Retrieval-Augmented Generation (RAG) system that allows you to ask questions about any PDF documents placed in its data/ directory. It processes these PDFs, extracts their content, chunks them into manageable pieces, and stores them as semantic memories. When you ask a question, it retrieves relevant information from these memories and uses a Large Language Model (LLM) to generate an answer.

Key Features:

• PDF Ingestion: Automatically processes PDF files found in the data/ directory.
• Semantic Memory: Chunks PDF content and stores embeddings for efficient retrieval of relevant information.
• Interactive Q&A: Engage in a conversational Q&A with your documents via an interactive dialog. Type @exit or @quit to end the session.
• LLM Powered: Utilizes the DeepSeek API for generating responses. The system is designed to be easily refactored to integrate with any other LLM API (e.g., OpenAI, Gemini, Ollama) by modifying the DeepSeekLLMNode or introducing a new LLM node.

Current Configuration:

• LLM Provider: DeepSeek API. Ensure your DEEPSEEK_API_KEY is set in your .env file.
• Embedding Model: nomic-embed-text via Ollama. Ensure Ollama is running (ollama serve) and the model is pulled (ollama pull nomic-embed-text).
• PDF Processing: Requires poppler-utils to be installed on your system.

To install dependencies:

bun install

To run:

bun src/index.js

This project was created using bun init in bun v1.3.1. Bun is a fast all-in-one JavaScript runtime.