When building RAG (Retrieval-Augmented Generation) applications, one of the most fundamental yet often overlooked components is the data indexing pipeline.