Bundled execution of the RAG corpus prep workflow — Take a raw document and turn it into a vector-DB-ready JSONL dataset, deterministically. Measures the corpus, token-counts it with the real OpenAI BPE, chunks at the right token boundary, attache...