Local_Pdf_Chat_RAG
LocalPdfChatRAG is a native Python implementation of a Retrieval-Augmented Generation (RAG) framework designed for educational purposes and practical deployment. It enables users to build a local intelligent Q&A system that processes various document formats including PDF, TXT, DOCX, MD, HTML, CSV, and Excel files. The system features a hybrid retrieval strategy combining FAISS vector search with BM25 keyword search to enhance recall and accuracy. It supports advanced optimization techniques such as cross-encoder reranking and recursive retrieval to refine results. Users can deploy the system with local Ollama models or cloud-based SiliconFlow APIs, featuring automatic backend detection and failover mechanisms. An additional optional module integrates SerpAPI for real-time web search enhancement. The application includes a user-friendly Gradio-based interactive web interface that visualizes document chunking and supports multi-model switching. The modular architecture separates core pipeline stages like docum