Add pydocstyle to the ruff rules (#52136)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: Harry Mellor <hej.mellor@gmail.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Harry Mellor
2026-09-16 16:44:29 +00:00
committed by GitHub
co-authored by Claude Opus 5
parent 804e5377aa
commit 8c1557a79c
2002 changed files with 15024 additions and 14269 deletions
+7 -12
View File
@@ -268,8 +268,7 @@ def test_save_and_load_slice(monkeypatch: pytest.MonkeyPatch):
def test_cache_load_returns_tuple_consistency_tuple_output(
monkeypatch: pytest.MonkeyPatch,
):
"""
Test that cache loading correctly handles models that return tuples.
"""Test that cache loading correctly handles models that return tuples.
This verifies that when a model returns a tuple of tensors, the output
type is preserved as a tuple between fresh compilation and cache load.
@@ -335,8 +334,7 @@ def test_cache_load_returns_tuple_consistency_tuple_output(
@pytest.mark.skipif(not is_torch_equal_or_newer("2.10.0"), reason="requires torch 2.10")
def test_shape_env(monkeypatch: pytest.MonkeyPatch):
"""
Test that the shape environment is correctly serialized and preserved
"""Test that the shape environment is correctly serialized and preserved
when loading from cache.
"""
with monkeypatch.context() as m:
@@ -374,8 +372,7 @@ def test_shape_env(monkeypatch: pytest.MonkeyPatch):
def test_partition_wrapper_applied_on_aot_load(
monkeypatch: pytest.MonkeyPatch, vllm_tmp_cache: Path, mocker
):
"""
Test that partition wrappers are applied when loading AOT cached functions.
"""Test that partition wrappers are applied when loading AOT cached functions.
This test verifies the fix for GitHub issue #31439 where AOT compile
caused 2x latency regression when use_inductor_graph_partition=True.
@@ -498,14 +495,12 @@ def test_standalone_compile_correctness():
@pytest.mark.skipif(not is_torch_equal_or_newer("2.10.0"), reason="requires torch 2.10")
@create_new_process_for_each_test("spawn")
def test_gpt2_cache_hit(monkeypatch: pytest.MonkeyPatch):
"""
Test that compiling gpt2 twice results in a cache hit.
"""Test that compiling gpt2 twice results in a cache hit.
Counter values are read from the EngineCore subprocess via
``LLM.collective_rpc`` so the test works under default V1
multiprocessing (no shared memory between test and engine).
"""
from vllm import LLM
def _snap(self):
@@ -641,7 +636,7 @@ class TestStandaloneCompiledArtifacts:
@patch("torch._inductor.standalone_compile.AOTCompiledArtifact.deserialize")
def test_load_all_success(self, mock_deserialize):
"""Test successful loading of all artifacts"""
"""Test successful loading of all artifacts."""
cache = StandaloneCompiledArtifacts()
mock_artifact1 = Mock()
@@ -658,7 +653,7 @@ class TestStandaloneCompiledArtifacts:
@patch("torch._inductor.standalone_compile.AOTCompiledArtifact.deserialize")
def test_load_all_already_loaded(self, mock_deserialize):
"""Test that load_all skips if already loaded"""
"""Test that load_all skips if already loaded."""
cache = StandaloneCompiledArtifacts()
mock_artifact = Mock()
@@ -671,7 +666,7 @@ class TestStandaloneCompiledArtifacts:
@patch("torch._inductor.standalone_compile.AOTCompiledArtifact.deserialize")
def test_get_loaded_artifact(self, mock_deserialize):
"""Test retrieving loaded artifacts"""
"""Test retrieving loaded artifacts."""
cache = StandaloneCompiledArtifacts()
mock_artifact = Mock()