# All-in-RAG

A full-stack, hands-on tutorial on retrieval-augmented generation for large-model application developers, covering data preparation, index construction, advanced retrieval, generation and evaluation.

All-in-RAG is a full-stack tutorial on retrieval-augmented generation aimed at developers building large-model applications, teaching a systematic path from fundamentals to production practice rather than a scattered set of unconnected tips. It moves through RAG concepts and use cases, a complete data-preparation pipeline covering document loading, cleaning and text chunking, index construction and optimization covering vector and multimodal embeddings and vector-database construction, advanced retrieval techniques including hybrid dense-and-sparse retrieval, query construction and translating natural language into SQL, and generation integration and evaluation methodology.

Later chapters walk through complete, practical RAG projects built from the earlier material, including one built on a knowledge graph, and an extra-chapter section invites community-contributed material on related RAG and vector-database topics. It assumes basic Python, basic Docker and basic Linux command-line skills, and covers both text and image retrieval.

Website: https://datawhalechina.github.io/all-in-rag/