[Misc] Cleanup: use consistent pytest-mock in unit tests (#2698)

Signed-off-by: yuanheng <jonathan.zhaoyh@gmail.com>

tests/engine/test_arg_utils.py

@@ -6,7 +6,7 @@ explicitly patch values that differ from vLLM.
 
 import argparse
 import inspect
-from unittest.mock import Mock
+from types import SimpleNamespace
 
 import pytest
 from pydantic import ValidationError
@@ -102,7 +102,7 @@ def test_qwen3_tts_codec_frame_rate_patching():
     vllm_config = EngineArgs().create_model_config()
 
     # Create a mock talking config with a dummy value for position_id_per_seconds
-    mock_talker_config = Mock()
+    mock_talker_config = SimpleNamespace()
     mock_talker_config.position_id_per_seconds = 12.3
     vllm_config.hf_config.talker_config = mock_talker_config
 
@@ -146,13 +146,12 @@ def test_stage_specific_text_config_override():
     # Switch the created hf text config with a mock whose
     # values we want to pull through the text config helper
     stage_text_config = vllm_config.hf_text_config
-    vllm_config.hf_text_config = Mock()
+    vllm_config.hf_text_config = SimpleNamespace()
     stage_text_config.sliding_window = 4096
     stage_text_config.attention_chunk_size = 2048
 
     # Move the stage config's text config getter & thinker config
-    mock_stage_config = Mock()
-    mock_stage_config.get_text_config.return_value = stage_text_config
+    mock_stage_config = SimpleNamespace(get_text_config=lambda: stage_text_config)
     vllm_config.hf_config.thinker_config = mock_stage_config
 
     # Ensure that create from a vLLM config correctly pulls the

tests/engine/test_async_omni_engine_input.py

@@ -1,6 +1,5 @@
-from unittest.mock import Mock
-
 import pytest
+from pytest_mock import MockerFixture
 from vllm.sampling_params import SamplingParams
 from vllm.v1.engine import EngineCoreRequest
 
@@ -24,18 +23,18 @@ def _make_engine_core_request() -> EngineCoreRequest:
     )
 
 
-def test_build_add_request_message_preserves_additional_information():
+def test_build_add_request_message_preserves_additional_information(mocker: MockerFixture):
     engine = object.__new__(AsyncOmniEngine)
     params = SamplingParams(max_tokens=8)
     engine.default_sampling_params_list = [params]
     engine.stage_metadata = [{"stage_type": "llm"}]
     engine.supported_tasks = ("speech",)
 
-    input_processor = Mock()
+    input_processor = mocker.Mock()
     input_processor.process_inputs.return_value = _make_engine_core_request()
     engine.input_processor = input_processor
 
-    output_processor = Mock()
+    output_processor = mocker.Mock()
     engine.output_processors = [output_processor]
 
     prompt = {
@@ -63,18 +62,18 @@ def test_build_add_request_message_preserves_additional_information():
     output_processor.add_request.assert_called_once()
 
 
-def test_build_add_request_message_with_resumable_streaming():
+def test_build_add_request_message_with_resumable_streaming(mocker: MockerFixture):
     engine = object.__new__(AsyncOmniEngine)
     params = SamplingParams(max_tokens=8)
     engine.default_sampling_params_list = [params]
     engine.stage_metadata = [{"stage_type": "llm"}]
     engine.supported_tasks = ("generate",)
 
-    input_processor = Mock()
+    input_processor = mocker.Mock()
     input_processor.process_inputs.return_value = _make_engine_core_request()
     engine.input_processor = input_processor
 
-    output_processor = Mock()
+    output_processor = mocker.Mock()
     engine.output_processors = [output_processor]
 
     msg = engine._build_add_request_message(

tests/engine/test_async_omni_engine_outputs.py

@@ -5,36 +5,36 @@ subsequent attempts to collect output raise RuntimeError.
 """
 
 import queue
-from unittest.mock import MagicMock
 
 import pytest
+from pytest_mock import MockerFixture
 
 from vllm_omni.engine.async_omni_engine import AsyncOmniEngine
 
 pytestmark = [pytest.mark.core_model, pytest.mark.cpu]
 
 
-def _make_engine(output_queue, *, thread_alive: bool = True) -> AsyncOmniEngine:
+def _make_engine(output_queue, mocker: MockerFixture, *, thread_alive: bool = True) -> AsyncOmniEngine:
     """Create an AsyncOmniEngine bypassing __init__."""
     engine = object.__new__(AsyncOmniEngine)
     engine.output_queue = output_queue
-    engine.orchestrator_thread = MagicMock(
-        is_alive=MagicMock(return_value=thread_alive),
+    engine.orchestrator_thread = mocker.MagicMock(
+        is_alive=mocker.MagicMock(return_value=thread_alive),
     )
     return engine
 
 
-def test_try_get_output_raises_after_orchestrator_dies():
+def test_try_get_output_raises_after_orchestrator_dies(mocker: MockerFixture):
     """Draining remaining results then hitting an empty queue with a dead
     orchestrator must raise RuntimeError so callers know the pipeline is gone."""
-    mock_queue = MagicMock()
+    mock_queue = mocker.MagicMock()
     # First call succeeds; second call finds the queue empty.
     mock_queue.sync_q.get.side_effect = [
         {"type": "output", "request_id": "r1"},
         queue.Empty,
     ]
 
-    engine = _make_engine(mock_queue, thread_alive=True)
+    engine = _make_engine(mock_queue, mocker, thread_alive=True)
 
     # Collect the one buffered result.
     assert engine.try_get_output()["request_id"] == "r1"
@@ -47,15 +47,15 @@ def test_try_get_output_raises_after_orchestrator_dies():
 
 
 @pytest.mark.asyncio
-async def test_try_get_output_async_raises_after_orchestrator_dies():
+async def test_try_get_output_async_raises_after_orchestrator_dies(mocker: MockerFixture):
     """Same scenario as above but for the async variant."""
-    mock_queue = MagicMock()
+    mock_queue = mocker.MagicMock()
     mock_queue.sync_q.get_nowait.side_effect = [
         {"type": "output", "request_id": "r1"},
         queue.Empty,
     ]
 
-    engine = _make_engine(mock_queue, thread_alive=True)
+    engine = _make_engine(mock_queue, mocker, thread_alive=True)
 
     assert (await engine.try_get_output_async())["request_id"] == "r1"