Merge tag 'v0.9.1' into v0.9.1-ori

tests/distributed/test_expert_parallel.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 from dataclasses import dataclass
 from typing import Literal, NamedTuple, Optional

tests/distributed/test_multi_node_assignment.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """Make sure ray assigns GPU workers to the correct node.
 
 Run:

tests/distributed/test_pipeline_parallel.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """
 WARNING: This test runs in both single-node (4 GPUs) and multi-node
  (2 node with 2 GPUs each) modes. If the test only uses 2 GPUs, it is
@@ -227,6 +228,7 @@ MULTIMODAL_MODELS = {
     "llava-hf/llava-onevision-qwen2-0.5b-ov-hf": PPTestSettings.fast(),
     "openbmb/MiniCPM-Llama3-V-2_5": PPTestSettings.fast(),
     "allenai/Molmo-7B-D-0924": PPTestSettings.fast(),
+    "AIDC-AI/Ovis2-1B": PPTestSettings.fast(),
     "microsoft/Phi-3.5-vision-instruct": PPTestSettings.fast(),
     "mistralai/Pixtral-12B-2409": PPTestSettings.fast(load_format="dummy"),
     "Qwen/Qwen-VL-Chat": PPTestSettings.fast(),

tests/distributed/test_pipeline_partition.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import os
 

tests/distributed/test_pp_cudagraph.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 from __future__ import annotations
 
 from typing import TYPE_CHECKING

tests/distributed/test_pynccl.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import multiprocessing
 import os

tests/distributed/test_same_node.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import os
 

tests/distributed/test_sequence_parallel.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """
 WARNING: This test runs in both single-node (4 GPUs) and multi-node
  (2 node with 2 GPUs each) modes. If the test only uses 2 GPUs, it is

tests/distributed/test_shm_broadcast.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import multiprocessing
 import random

tests/distributed/test_torchrun_example.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 # unit test for `examples/offline_inference/torchrun_example.py`
 import os

tests/distributed/test_utils.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import socket
 

tests/encoder_decoder/test_e2e_correctness.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """E2E tests to verify the correctness of the encoder-decoder framework
 
 Run `pytest tests/encoder_decoder/test_e2e_correctness.py`.

tests/engine/conftest.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 import pytest
 
 

tests/engine/test_arg_utils.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import json
 from argparse import ArgumentError, ArgumentTypeError
 from contextlib import nullcontext
 from dataclasses import dataclass, field
-from typing import Literal, Optional
+from typing import Annotated, Literal, Optional
 
 import pytest
 
 from vllm.config import CompilationConfig, config
 from vllm.engine.arg_utils import (EngineArgs, contains_type, get_kwargs,
-                                   get_type, is_not_builtin, is_type,
-                                   literal_to_kwargs, nullable_kvs,
+                                   get_type, get_type_hints, is_not_builtin,
+                                   is_type, literal_to_kwargs, nullable_kvs,
                                    optional_type, parse_type)
 from vllm.utils import FlexibleArgumentParser
 
@@ -159,6 +160,18 @@ def test_is_not_builtin(type_hint, expected):
     assert is_not_builtin(type_hint) == expected
 
 
+@pytest.mark.parametrize(
+    ("type_hint", "expected"), [
+        (Annotated[int, "annotation"], {int}),
+        (Optional[int], {int, type(None)}),
+        (Annotated[Optional[int], "annotation"], {int, type(None)}),
+        (Optional[Annotated[int, "annotation"]], {int, type(None)}),
+    ],
+    ids=["Annotated", "Optional", "Annotated_Optional", "Optional_Annotated"])
+def test_get_type_hints(type_hint, expected):
+    assert get_type_hints(type_hint) == expected
+
+
 def test_get_kwargs():
     kwargs = get_kwargs(DummyConfig)
     print(kwargs)

tests/engine/test_computed_prefix_blocks.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import pytest
 

tests/engine/test_executor.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import asyncio
 import os

tests/engine/test_multi_step_output_processor.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import random
 from unittest.mock import MagicMock

tests/engine/test_multiproc_workers.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import asyncio
 from concurrent.futures import ThreadPoolExecutor

tests/engine/test_options.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 from contextlib import nullcontext
 
 import pytest

tests/engine/test_short_mm_context.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import pytest