From bbe08d6d81c9d8aef87223082a70e2b8eaae1095 Mon Sep 17 00:00:00 2001 From: sigoden Date: Tue, 2 Jul 2024 09:20:40 +0800 Subject: refactor: improve RAG (#676) --- src/rag/mod.rs | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) (limited to 'src') diff --git a/src/rag/mod.rs b/src/rag/mod.rs index f5f68ae..e2f6ddd 100644 --- a/src/rag/mod.rs +++ b/src/rag/mod.rs @@ -316,8 +316,14 @@ impl Rag { self.data.chunk_overlap, &separator, ); + + let metadata = metadata + .iter() + .map(|(k, v)| format!("{k}: {v}\n")) + .collect::>() + .join(""); let split_options = SplitterChunkHeaderOptions::default().with_chunk_header(&format!( - "\npath: {path}\n\n" + "\npath: {path}\n{metadata}\n\n" )); let document = RagDocument::new(contents); let splitted_documents = splitter.split_documents(&[document], &split_options); @@ -354,7 +360,7 @@ impl Rag { self.data.add(next_file_id, files, document_ids, embeddings); self.data.document_paths = document_paths; - progress(&spinner, "Building database".into()); + progress(&spinner, "Building store".into()); self.hnsw = self.data.build_hnsw(); self.bm25 = self.data.build_bm25(); -- cgit v1.2.3