RAG_langchain

Why we included this project

This repo works through the whole retrieval-augmented generation stack on LangChain, notebook by notebook, so it reads more like a course than a single demo. You get document loading and parsing, chunking, embedding choice, vector database comparison, reranking, and evaluation, with each stage shown in enough detail to see how it affects retrieval quality. The notebooks make concrete calls too: Chroma versus FAISS versus Milvus, metadata filtering, hybrid dense-plus-sparse retrieval, and even Graph RAG and long-context setups. Newcomers can follow the beginner-to-advanced path to get oriented, while people already building systems can jump straight to the reranker, evaluation, or agent notebooks and borrow working patterns. It is a learning resource rather than something you deploy, but the code is concrete enough to lift pieces from for real projects.