Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
fix(api): report batch count as the progress denominator in Vector (#43005)
huoex1222-hub1366cadbb4c9
diff --git a/api/core/rag/datasource/vdb/vector_factory.py b/api/core/rag/datasource/vdb/vector_factory.pyindex 5fbc55ffd91..0806dfb2745 100644--- a/api/core/rag/datasource/vdb/vector_factory.py+++ b/api/core/rag/datasource/vdb/vector_factory.py@@ -180,7 +180,7 @@ class Vector:start = time.time()logger.info("start embedding %s texts %s", len(texts), start)batch_size = 1000- total_batches = len(texts) + batch_size - 1+ total_batches = (len(texts) + batch_size - 1) // batch_sizefor i in range(0, len(texts), batch_size):batch = texts[i : i + batch_size]batch_start = time.time()@@ -197,7 +197,7 @@ class Vector:start = time.time()logger.info("start embedding %s files %s", len(file_documents), start)batch_size = 1000- total_batches = len(file_documents) + batch_size - 1+ total_batches = (len(file_documents) + batch_size - 1) // batch_sizefor i in range(0, len(file_documents), batch_size):batch = file_documents[i : i + batch_size]batch_start = time.time()diff --git a/api/tests/unit_tests/core/rag/datasource/vdb/test_vector_factory.py b/api/tests/unit_tests/core/rag/datasource/vdb/test_vector_factory.pyindex afa3b56d8fe..5ed354d15ac 100644--- a/api/tests/unit_tests/core/rag/datasource/vdb/test_vector_factory.py+++ b/api/tests/unit_tests/core/rag/datasource/vdb/test_vector_factory.py@@ -1,5 +1,6 @@import base64import json+import loggingimport sysimport typesfrom datetime import UTC, datetime@@ -358,6 +359,28 @@ def test_create_batches_texts_and_skips_empty_input(vector_factory_module):vector._vector_processor.create.assert_not_called()+def test_create_logs_batch_count_as_progress_denominator(vector_factory_module, caplog):+ """Progress logs must report the batch count, not a text count with the raw remainder."""+ vector = vector_factory_module.Vector.__new__(vector_factory_module.Vector)+ vector._embeddings = MagicMock()+ vector._embeddings.embed_documents.side_effect = [+ [[0.1] for _ in range(1000)],+ [[0.2] for _ in range(500)],+ ]+ vector._vector_processor = MagicMock()++ docs = [Document(page_content=f"doc-{i}", metadata={"doc_id": f"id-{i}"}) for i in range(1500)]++ with caplog.at_level(logging.INFO):+ vector.create(texts=docs)++ progress = [record.getMessage() for record in caplog.records if "Processing batch" in record.getMessage()]+ assert progress == [+ "Processing batch 1/2 (1000 texts)",+ "Processing batch 2/2 (500 texts)",+ ]++def test_create_skips_empty_text_documents_before_embedding(vector_factory_module):vector = vector_factory_module.Vector.__new__(vector_factory_module.Vector)vector._embeddings = MagicMock()