apunkt/SurfSense
1
0
Fork 0
mirror of https://github.com/MODSetter/SurfSense.git synced 2026-05-21 18:55:16 +02:00
Code Issues Projects Releases Packages Wiki Activity

refactor: update test for index() to use chunk_text_hybrid

Browse source
This commit is contained in:
Anish Sarkar 2026-05-13 00:22:43 +05:30
parent 2402b730fa
commit 9b926b3133
1 changed files with 16 additions and 6 deletions
Download patch file Download diff file Expand all files Collapse all files

22
surfsense_backend/tests/unit/indexing_pipeline/test_index_batch_parallel.py
View file

@ -37,7 +37,11 @@ def _make_orm_doc(connector_doc, doc_id):
async def test_index_calls_embed_and_chunk_via_to_thread( async def test_index_calls_embed_and_chunk_via_to_thread(
pipeline, make_connector_document, monkeypatch pipeline, make_connector_document, monkeypatch
): ):
"""index() runs embed_texts and chunk_text via asyncio.to_thread, not blocking the loop.""" """index() runs the chunker and embed_texts via asyncio.to_thread, not blocking the loop.
The default (non-code) path uses ``chunk_text_hybrid`` so Markdown tables stay
intact (see issue #1334); ``chunk_text`` is reserved for the code-chunker branch.
"""
to_thread_calls = [] to_thread_calls = []
original_to_thread = asyncio.to_thread original_to_thread = asyncio.to_thread
@ -51,11 +55,11 @@ async def test_index_calls_embed_and_chunk_via_to_thread(
"app.indexing_pipeline.indexing_pipeline_service.summarize_document", "app.indexing_pipeline.indexing_pipeline_service.summarize_document",
AsyncMock(return_value="Summary."), AsyncMock(return_value="Summary."),
) )
mock_chunk = MagicMock(return_value=["chunk1"]) mock_chunk_hybrid = MagicMock(return_value=["chunk1"])
mock_chunk.__name__ = "chunk_text" mock_chunk_hybrid.__name__ = "chunk_text_hybrid"
monkeypatch.setattr( monkeypatch.setattr(
"app.indexing_pipeline.indexing_pipeline_service.chunk_text", "app.indexing_pipeline.indexing_pipeline_service.chunk_text_hybrid",
mock_chunk, mock_chunk_hybrid,
) )
mock_embed = MagicMock( mock_embed = MagicMock(
side_effect=lambda texts: [[0.1] * _EMBEDDING_DIM for _ in texts] side_effect=lambda texts: [[0.1] * _EMBEDDING_DIM for _ in texts]
@ -65,6 +69,11 @@ async def test_index_calls_embed_and_chunk_via_to_thread(
"app.indexing_pipeline.indexing_pipeline_service.embed_texts", "app.indexing_pipeline.indexing_pipeline_service.embed_texts",
mock_embed, mock_embed,
) )
# Bypass set_committed_value, which requires a real ORM instance (not MagicMock).
monkeypatch.setattr(
"app.indexing_pipeline.indexing_pipeline_service.attach_chunks_to_document",
MagicMock(),
)
connector_doc = make_connector_document( connector_doc = make_connector_document(
document_type=DocumentType.GOOGLE_GMAIL_CONNECTOR, document_type=DocumentType.GOOGLE_GMAIL_CONNECTOR,
@ -77,8 +86,9 @@ async def test_index_calls_embed_and_chunk_via_to_thread(
await pipeline.index(document, connector_doc, llm=MagicMock()) await pipeline.index(document, connector_doc, llm=MagicMock())
assert "chunk_text" in to_thread_calls assert "chunk_text_hybrid" in to_thread_calls
assert "embed_texts" in to_thread_calls assert "embed_texts" in to_thread_calls
assert document.status == DocumentStatus.ready()
def _mock_session_factory(orm_docs_by_id): def _mock_session_factory(orm_docs_by_id):