mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-07-22 20:43:59 +00:00
chore: default to pytest asyncio-mode=auto (#2730)
# What does this PR do? previously, developers who ran `./scripts/unit-tests.sh` would get `asyncio-mode=auto`, which meant `@pytest.mark.asyncio` and `@pytest_asyncio.fixture` were redundent. developers who ran `pytest` directly would get pytest's default (strict mode), would run into errors leading them to add `@pytest.mark.asyncio` / `@pytest_asyncio.fixture` to their code. with this change - - `asyncio_mode=auto` is included in `pyproject.toml` making behavior consistent for all invocations of pytest - removes all redundant `@pytest_asyncio.fixture` and `@pytest.mark.asyncio` - for good measure, requires `pytest>=8.4` and `pytest-asyncio>=1.0` ## Test Plan - `./scripts/unit-tests.sh` - `uv run pytest tests/unit`
This commit is contained in:
Matthew Farrellee
GitHub
parent
2ebc172f33
commit
30b2e6a495
35 changed files with 29 additions and 239 deletions
|
|
@ -9,7 +9,6 @@ from unittest.mock import AsyncMock, MagicMock, patch
|
|||
|
||||
import numpy as np
|
||||
import pytest
|
||||
import pytest_asyncio
|
||||
|
||||
from llama_stack.apis.files import Files
|
||||
from llama_stack.apis.inference import EmbeddingsResponse, Inference
|
||||
|
|
@ -91,13 +90,13 @@ def faiss_config():
|
|||
return config
|
||||
|
||||
|
||||
@pytest_asyncio.fixture
|
||||
@pytest.fixture
|
||||
async def faiss_index(embedding_dimension):
|
||||
index = await FaissIndex.create(dimension=embedding_dimension)
|
||||
yield index
|
||||
|
||||
|
||||
@pytest_asyncio.fixture
|
||||
@pytest.fixture
|
||||
async def faiss_adapter(faiss_config, mock_inference_api, mock_files_api) -> FaissVectorIOAdapter:
|
||||
# Create the adapter
|
||||
adapter = FaissVectorIOAdapter(config=faiss_config, inference_api=mock_inference_api, files_api=mock_files_api)
|
||||
|
|
@ -113,7 +112,6 @@ async def faiss_adapter(faiss_config, mock_inference_api, mock_files_api) -> Fai
|
|||
yield adapter
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_faiss_query_vector_returns_infinity_when_query_and_embedding_are_identical(
|
||||
faiss_index, sample_chunks, sample_embeddings, embedding_dimension
|
||||
):
|
||||
|
|
@ -136,7 +134,6 @@ async def test_faiss_query_vector_returns_infinity_when_query_and_embedding_are_
|
|||
assert response.chunks[1] == sample_chunks[1]
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_health_success():
|
||||
"""Test that the health check returns OK status when faiss is working correctly."""
|
||||
# Create a fresh instance of FaissVectorIOAdapter for testing
|
||||
|
|
@ -160,7 +157,6 @@ async def test_health_success():
|
|||
mock_index_flat.assert_called_once_with(128) # VECTOR_DIMENSION is 128
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_health_failure():
|
||||
"""Test that the health check returns ERROR status when faiss encounters an error."""
|
||||
# Create a fresh instance of FaissVectorIOAdapter for testing
|
||||
|
|
|
|||
|
|
@ -10,7 +10,6 @@ from typing import Any
|
|||
from unittest.mock import AsyncMock, MagicMock, patch
|
||||
|
||||
import pytest
|
||||
import pytest_asyncio
|
||||
|
||||
from llama_stack.apis.inference import EmbeddingsResponse, Inference
|
||||
from llama_stack.apis.vector_io import (
|
||||
|
|
@ -68,7 +67,7 @@ def mock_api_service(sample_embeddings):
|
|||
return mock_api_service
|
||||
|
||||
|
||||
@pytest_asyncio.fixture
|
||||
@pytest.fixture
|
||||
async def qdrant_adapter(qdrant_config, mock_vector_db_store, mock_api_service, loop) -> QdrantVectorIOAdapter:
|
||||
adapter = QdrantVectorIOAdapter(config=qdrant_config, inference_api=mock_api_service)
|
||||
adapter.vector_db_store = mock_vector_db_store
|
||||
|
|
@ -80,7 +79,6 @@ async def qdrant_adapter(qdrant_config, mock_vector_db_store, mock_api_service,
|
|||
__QUERY = "Sample query"
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
@pytest.mark.parametrize("max_query_chunks, expected_chunks", [(2, 2), (100, 60)])
|
||||
async def test_qdrant_adapter_returns_expected_chunks(
|
||||
qdrant_adapter: QdrantVectorIOAdapter,
|
||||
|
|
@ -111,7 +109,6 @@ def _prepare_for_json(value: Any) -> str:
|
|||
|
||||
|
||||
@patch("llama_stack.providers.utils.telemetry.trace_protocol._prepare_for_json", new=_prepare_for_json)
|
||||
@pytest.mark.asyncio
|
||||
async def test_qdrant_register_and_unregister_vector_db(
|
||||
qdrant_adapter: QdrantVectorIOAdapter,
|
||||
mock_vector_db,
|
||||
|
|
|
|||
|
|
@ -8,7 +8,6 @@ import asyncio
|
|||
|
||||
import numpy as np
|
||||
import pytest
|
||||
import pytest_asyncio
|
||||
|
||||
from llama_stack.apis.vector_io import Chunk, QueryChunksResponse
|
||||
from llama_stack.providers.inline.vector_io.sqlite_vec.sqlite_vec import (
|
||||
|
|
@ -34,7 +33,7 @@ def loop():
|
|||
return asyncio.new_event_loop()
|
||||
|
||||
|
||||
@pytest_asyncio.fixture
|
||||
@pytest.fixture
|
||||
async def sqlite_vec_index(embedding_dimension, tmp_path_factory):
|
||||
temp_dir = tmp_path_factory.getbasetemp()
|
||||
db_path = str(temp_dir / "test_sqlite.db")
|
||||
|
|
@ -43,14 +42,12 @@ async def sqlite_vec_index(embedding_dimension, tmp_path_factory):
|
|||
await index.delete()
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunk_metadata(sqlite_vec_index, sample_chunks_with_metadata, sample_embeddings_with_metadata):
|
||||
await sqlite_vec_index.add_chunks(sample_chunks_with_metadata, sample_embeddings_with_metadata)
|
||||
response = await sqlite_vec_index.query_vector(sample_embeddings_with_metadata[-1], k=2, score_threshold=0.0)
|
||||
assert response.chunks[0].chunk_metadata == sample_chunks_with_metadata[-1].chunk_metadata
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_full_text_search(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
query_string = "Sentence 5"
|
||||
|
|
@ -68,7 +65,6 @@ async def test_query_chunks_full_text_search(sqlite_vec_index, sample_chunks, sa
|
|||
assert len(response_no_results.chunks) == 0, f"Expected 0 results, but got {len(response_no_results.chunks)}"
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
||||
|
|
@ -90,7 +86,6 @@ async def test_query_chunks_hybrid(sqlite_vec_index, sample_chunks, sample_embed
|
|||
assert all(response.scores[i] >= response.scores[i + 1] for i in range(len(response.scores) - 1))
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_full_text_search_k_greater_than_results(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
# Re-initialize with a clean index
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
|
@ -103,7 +98,6 @@ async def test_query_chunks_full_text_search_k_greater_than_results(sqlite_vec_i
|
|||
assert any("Sentence 1 from document 0" in chunk.content for chunk in response.chunks), "Expected chunk not found"
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_chunk_id_conflict(sqlite_vec_index, sample_chunks, embedding_dimension):
|
||||
"""Test that chunk IDs do not conflict across batches when inserting chunks."""
|
||||
# Reduce batch size to force multiple batches for same document
|
||||
|
|
@ -134,7 +128,6 @@ async def sqlite_vec_adapter(sqlite_connection):
|
|||
await adapter.shutdown()
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_no_keyword_matches(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
"""Test hybrid search when keyword search returns no matches - should still return vector results."""
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
|
@ -163,7 +156,6 @@ async def test_query_chunks_hybrid_no_keyword_matches(sqlite_vec_index, sample_c
|
|||
assert all(response.scores[i] >= response.scores[i + 1] for i in range(len(response.scores) - 1))
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_score_threshold(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
"""Test hybrid search with a high score threshold."""
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
|
@ -185,7 +177,6 @@ async def test_query_chunks_hybrid_score_threshold(sqlite_vec_index, sample_chun
|
|||
assert len(response.chunks) == 0
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_different_embedding(
|
||||
sqlite_vec_index, sample_chunks, sample_embeddings, embedding_dimension
|
||||
):
|
||||
|
|
@ -211,7 +202,6 @@ async def test_query_chunks_hybrid_different_embedding(
|
|||
assert all(response.scores[i] >= response.scores[i + 1] for i in range(len(response.scores) - 1))
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_rrf_ranking(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
"""Test that RRF properly combines rankings when documents appear in both search methods."""
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
|
@ -236,7 +226,6 @@ async def test_query_chunks_hybrid_rrf_ranking(sqlite_vec_index, sample_chunks,
|
|||
assert all(response.scores[i] >= response.scores[i + 1] for i in range(len(response.scores) - 1))
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_score_selection(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
||||
|
|
@ -284,7 +273,6 @@ async def test_query_chunks_hybrid_score_selection(sqlite_vec_index, sample_chun
|
|||
assert response.scores[0] == pytest.approx(2.0 / 61.0, rel=1e-6) # Should behave like RRF
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_mixed_results(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
"""Test hybrid search with documents that appear in only one search method."""
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
|
@ -313,7 +301,6 @@ async def test_query_chunks_hybrid_mixed_results(sqlite_vec_index, sample_chunks
|
|||
assert "document-2" in doc_ids # From keyword search
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_weighted_reranker_parametrization(
|
||||
sqlite_vec_index, sample_chunks, sample_embeddings
|
||||
):
|
||||
|
|
@ -369,7 +356,6 @@ async def test_query_chunks_hybrid_weighted_reranker_parametrization(
|
|||
)
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_rrf_impact_factor(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
"""Test RRFReRanker with different impact factors."""
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
|
@ -401,7 +387,6 @@ async def test_query_chunks_hybrid_rrf_impact_factor(sqlite_vec_index, sample_ch
|
|||
assert response.scores[0] == pytest.approx(2.0 / 101.0, rel=1e-6)
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_edge_cases(sqlite_vec_index, sample_chunks, sample_embeddings):
|
||||
await sqlite_vec_index.add_chunks(sample_chunks, sample_embeddings)
|
||||
|
||||
|
|
@ -445,7 +430,6 @@ async def test_query_chunks_hybrid_edge_cases(sqlite_vec_index, sample_chunks, s
|
|||
assert len(response.chunks) <= 100
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_hybrid_tie_breaking(
|
||||
sqlite_vec_index, sample_embeddings, embedding_dimension, tmp_path_factory
|
||||
):
|
||||
|
|
|
|||
|
|
@ -25,12 +25,10 @@ from llama_stack.providers.remote.vector_io.milvus.milvus import VECTOR_DBS_PREF
|
|||
# -v -s --tb=short --disable-warnings --asyncio-mode=auto
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_initialize_index(vector_index):
|
||||
await vector_index.initialize()
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_add_chunks_query_vector(vector_index, sample_chunks, sample_embeddings):
|
||||
vector_index.delete()
|
||||
vector_index.initialize()
|
||||
|
|
@ -40,7 +38,6 @@ async def test_add_chunks_query_vector(vector_index, sample_chunks, sample_embed
|
|||
vector_index.delete()
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_chunk_id_conflict(vector_index, sample_chunks, embedding_dimension):
|
||||
embeddings = np.random.rand(len(sample_chunks), embedding_dimension).astype(np.float32)
|
||||
await vector_index.add_chunks(sample_chunks, embeddings)
|
||||
|
|
@ -54,7 +51,6 @@ async def test_chunk_id_conflict(vector_index, sample_chunks, embedding_dimensio
|
|||
assert len(contents) == len(set(contents))
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_initialize_adapter_with_existing_kvstore(vector_io_adapter):
|
||||
key = f"{VECTOR_DBS_PREFIX}db1"
|
||||
dummy = VectorDB(
|
||||
|
|
@ -65,7 +61,6 @@ async def test_initialize_adapter_with_existing_kvstore(vector_io_adapter):
|
|||
await vector_io_adapter.initialize()
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_persistence_across_adapter_restarts(vector_io_adapter):
|
||||
await vector_io_adapter.initialize()
|
||||
dummy = VectorDB(
|
||||
|
|
@ -79,7 +74,6 @@ async def test_persistence_across_adapter_restarts(vector_io_adapter):
|
|||
await vector_io_adapter.shutdown()
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_register_and_unregister_vector_db(vector_io_adapter):
|
||||
unique_id = f"foo_db_{np.random.randint(1e6)}"
|
||||
dummy = VectorDB(
|
||||
|
|
@ -92,14 +86,12 @@ async def test_register_and_unregister_vector_db(vector_io_adapter):
|
|||
assert dummy.identifier not in vector_io_adapter.cache
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_unregistered_raises(vector_io_adapter):
|
||||
fake_emb = np.zeros(8, dtype=np.float32)
|
||||
with pytest.raises(ValueError):
|
||||
await vector_io_adapter.query_chunks("no_such_db", fake_emb)
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_insert_chunks_calls_underlying_index(vector_io_adapter):
|
||||
fake_index = AsyncMock()
|
||||
vector_io_adapter._get_and_cache_vector_db_index = AsyncMock(return_value=fake_index)
|
||||
|
|
@ -110,7 +102,6 @@ async def test_insert_chunks_calls_underlying_index(vector_io_adapter):
|
|||
fake_index.insert_chunks.assert_awaited_once_with(chunks)
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_insert_chunks_missing_db_raises(vector_io_adapter):
|
||||
vector_io_adapter._get_and_cache_vector_db_index = AsyncMock(return_value=None)
|
||||
|
||||
|
|
@ -118,7 +109,6 @@ async def test_insert_chunks_missing_db_raises(vector_io_adapter):
|
|||
await vector_io_adapter.insert_chunks("db_not_exist", [])
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_calls_underlying_index_and_returns(vector_io_adapter):
|
||||
expected = QueryChunksResponse(chunks=[Chunk(content="c1")], scores=[0.1])
|
||||
fake_index = AsyncMock(query_chunks=AsyncMock(return_value=expected))
|
||||
|
|
@ -130,7 +120,6 @@ async def test_query_chunks_calls_underlying_index_and_returns(vector_io_adapter
|
|||
assert response is expected
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_query_chunks_missing_db_raises(vector_io_adapter):
|
||||
vector_io_adapter._get_and_cache_vector_db_index = AsyncMock(return_value=None)
|
||||
|
||||
|
|
@ -138,7 +127,6 @@ async def test_query_chunks_missing_db_raises(vector_io_adapter):
|
|||
await vector_io_adapter.query_chunks("db_missing", "q", None)
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_save_openai_vector_store(vector_io_adapter):
|
||||
store_id = "vs_1234"
|
||||
openai_vector_store = {
|
||||
|
|
@ -155,7 +143,6 @@ async def test_save_openai_vector_store(vector_io_adapter):
|
|||
assert vector_io_adapter.openai_vector_stores[openai_vector_store["id"]] == openai_vector_store
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_update_openai_vector_store(vector_io_adapter):
|
||||
store_id = "vs_1234"
|
||||
openai_vector_store = {
|
||||
|
|
@ -172,7 +159,6 @@ async def test_update_openai_vector_store(vector_io_adapter):
|
|||
assert vector_io_adapter.openai_vector_stores[openai_vector_store["id"]] == openai_vector_store
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_delete_openai_vector_store(vector_io_adapter):
|
||||
store_id = "vs_1234"
|
||||
openai_vector_store = {
|
||||
|
|
@ -188,7 +174,6 @@ async def test_delete_openai_vector_store(vector_io_adapter):
|
|||
assert openai_vector_store["id"] not in vector_io_adapter.openai_vector_stores
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_load_openai_vector_stores(vector_io_adapter):
|
||||
store_id = "vs_1234"
|
||||
openai_vector_store = {
|
||||
|
|
@ -204,7 +189,6 @@ async def test_load_openai_vector_stores(vector_io_adapter):
|
|||
assert loaded_stores[store_id] == openai_vector_store
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_save_openai_vector_store_file(vector_io_adapter, tmp_path_factory):
|
||||
store_id = "vs_1234"
|
||||
file_id = "file_1234"
|
||||
|
|
@ -226,7 +210,6 @@ async def test_save_openai_vector_store_file(vector_io_adapter, tmp_path_factory
|
|||
await vector_io_adapter._save_openai_vector_store_file(store_id, file_id, file_info, file_contents)
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_update_openai_vector_store_file(vector_io_adapter, tmp_path_factory):
|
||||
store_id = "vs_1234"
|
||||
file_id = "file_1234"
|
||||
|
|
@ -260,7 +243,6 @@ async def test_update_openai_vector_store_file(vector_io_adapter, tmp_path_facto
|
|||
assert loaded_contents != file_info
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_load_openai_vector_store_file_contents(vector_io_adapter, tmp_path_factory):
|
||||
store_id = "vs_1234"
|
||||
file_id = "file_1234"
|
||||
|
|
@ -284,7 +266,6 @@ async def test_load_openai_vector_store_file_contents(vector_io_adapter, tmp_pat
|
|||
assert loaded_contents == file_contents
|
||||
|
||||
|
||||
@pytest.mark.asyncio
|
||||
async def test_delete_openai_vector_store_file_from_storage(vector_io_adapter, tmp_path_factory):
|
||||
store_id = "vs_1234"
|
||||
file_id = "file_1234"
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue