[Refactor] [6/N] to simplify the vLLM openai chat_completion serving architecture (#32240)

Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>

tests/entrypoints/openai/responses/test_errors.py

@@ -6,8 +6,8 @@ from unittest.mock import MagicMock
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import ErrorResponse
-from vllm.entrypoints.openai.serving_engine import GenerationError, OpenAIServing
+from vllm.entrypoints.openai.engine.protocol import ErrorResponse
+from vllm.entrypoints.openai.engine.serving import GenerationError, OpenAIServing
 
 
 @pytest.mark.asyncio

tests/entrypoints/openai/responses/test_function_call_parsing.py

@@ -7,7 +7,7 @@ import json
 import pytest
 from openai.types.responses import ResponseFunctionToolCall
 
-from vllm.entrypoints.openai.protocol import ResponsesRequest
+from vllm.entrypoints.openai.engine.protocol import ResponsesRequest
 
 
 def test_function_call_dict_converted_to_object():
@@ -253,7 +253,7 @@ def test_function_call_validation_failure_logs_debug(caplog):
     }
 
     # Mock the logger to verify debug was called
-    with patch("vllm.entrypoints.openai.protocol.logger") as mock_logger:
+    with patch("vllm.entrypoints.openai.engine.protocol.logger") as mock_logger:
         with pytest.raises(ValueError):
             ResponsesRequest(**request_data)
 

tests/entrypoints/openai/test_chat_error.py

@@ -9,8 +9,9 @@ from unittest.mock import AsyncMock, MagicMock
 import pytest
 
 from vllm.config.multimodal import MultiModalConfig
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest, ErrorResponse
-from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
+from vllm.entrypoints.openai.engine.protocol import ErrorResponse
 from vllm.entrypoints.openai.serving_models import BaseModelPath, OpenAIServingModels
 from vllm.outputs import CompletionOutput, RequestOutput
 from vllm.tokenizers import get_tokenizer

tests/entrypoints/openai/test_chat_template.py

@@ -5,7 +5,7 @@ import pytest
 
 from vllm.config import ModelConfig
 from vllm.entrypoints.chat_utils import apply_hf_chat_template, load_chat_template
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.tokenizers import get_tokenizer
 
 from ...models.registry import HF_EXAMPLE_MODELS

tests/entrypoints/openai/test_completion_error.py

@@ -9,7 +9,7 @@ from unittest.mock import AsyncMock, MagicMock
 import pytest
 
 from vllm.config.multimodal import MultiModalConfig
-from vllm.entrypoints.openai.protocol import CompletionRequest, ErrorResponse
+from vllm.entrypoints.openai.engine.protocol import CompletionRequest, ErrorResponse
 from vllm.entrypoints.openai.serving_completion import OpenAIServingCompletion
 from vllm.entrypoints.openai.serving_models import BaseModelPath, OpenAIServingModels
 from vllm.outputs import CompletionOutput, RequestOutput

tests/entrypoints/openai/test_gptoss_structural_tags_integration.py

@@ -8,7 +8,7 @@ from unittest.mock import Mock
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.engine.protocol import (
     StructuredOutputsParams,
 )
 from vllm.entrypoints.tool_server import ToolServer

tests/entrypoints/openai/test_lora_resolvers.py

@@ -9,7 +9,7 @@ from unittest.mock import AsyncMock, MagicMock
 import pytest
 
 from vllm.config.multimodal import MultiModalConfig
-from vllm.entrypoints.openai.protocol import CompletionRequest, ErrorResponse
+from vllm.entrypoints.openai.engine.protocol import CompletionRequest, ErrorResponse
 from vllm.entrypoints.openai.serving_completion import OpenAIServingCompletion
 from vllm.entrypoints.openai.serving_models import BaseModelPath, OpenAIServingModels
 from vllm.lora.request import LoRARequest

tests/entrypoints/openai/test_protocol.py

@@ -4,7 +4,10 @@ from openai_harmony import (
     Message,
 )
 
-from vllm.entrypoints.openai.protocol import serialize_message, serialize_messages
+from vllm.entrypoints.openai.engine.protocol import (
+    serialize_message,
+    serialize_messages,
+)
 
 
 def test_serialize_message() -> None:

tests/entrypoints/openai/test_serving_chat.py

@@ -11,14 +11,16 @@ import pytest_asyncio
 from openai import OpenAI
 
 from vllm.config.multimodal import MultiModalConfig
-from vllm.entrypoints.openai.parser.harmony_utils import get_encoding
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
     ChatCompletionResponse,
+)
+from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
+from vllm.entrypoints.openai.engine.protocol import (
     ErrorResponse,
     RequestResponseMetadata,
 )
-from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
+from vllm.entrypoints.openai.parser.harmony_utils import get_encoding
 from vllm.entrypoints.openai.serving_models import BaseModelPath, OpenAIServingModels
 from vllm.outputs import CompletionOutput, RequestOutput
 from vllm.tokenizers import get_tokenizer
@@ -1517,12 +1519,12 @@ class TestCreateRemainingArgsDelta:
 
     def test_preserves_id_type_name(self):
         """Test that id, type, and name are preserved from original delta."""
-        from vllm.entrypoints.openai.protocol import (
+        from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
+        from vllm.entrypoints.openai.engine.protocol import (
             DeltaFunctionCall,
             DeltaMessage,
             DeltaToolCall,
         )
-        from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
 
         original_delta = DeltaMessage(
             tool_calls=[
@@ -1552,12 +1554,12 @@ class TestCreateRemainingArgsDelta:
 
     def test_matches_by_index(self):
         """Test that the correct tool call is matched by index."""
-        from vllm.entrypoints.openai.protocol import (
+        from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
+        from vllm.entrypoints.openai.engine.protocol import (
             DeltaFunctionCall,
             DeltaMessage,
             DeltaToolCall,
         )
-        from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
 
         original_delta = DeltaMessage(
             tool_calls=[
@@ -1588,12 +1590,12 @@ class TestCreateRemainingArgsDelta:
 
     def test_no_matching_tool_call(self):
         """Test graceful handling when no matching tool call is found."""
-        from vllm.entrypoints.openai.protocol import (
+        from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
+        from vllm.entrypoints.openai.engine.protocol import (
             DeltaFunctionCall,
             DeltaMessage,
             DeltaToolCall,
         )
-        from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
 
         original_delta = DeltaMessage(
             tool_calls=[
@@ -1620,8 +1622,8 @@ class TestCreateRemainingArgsDelta:
 
     def test_function_is_none(self):
         """Test handling when original tool call has no function."""
-        from vllm.entrypoints.openai.protocol import DeltaMessage, DeltaToolCall
-        from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
+        from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
+        from vllm.entrypoints.openai.engine.protocol import DeltaMessage, DeltaToolCall
 
         original_delta = DeltaMessage(
             tool_calls=[

tests/entrypoints/openai/test_serving_chat_stream_harmony.py

@@ -9,7 +9,7 @@ from unittest.mock import patch
 
 import pytest
 
-from vllm.entrypoints.openai.serving_chat_stream_harmony import (
+from vllm.entrypoints.openai.chat_completion.stream_harmony import (
     extract_harmony_streaming_delta,
 )
 
@@ -82,7 +82,7 @@ class TestExtractHarmonyStreamingDelta:
         assert tools_streamed is False
 
     @pytest.mark.parametrize("channel", ["commentary", "analysis"])
-    @patch("vllm.entrypoints.openai.serving_chat_stream_harmony.make_tool_call_id")
+    @patch("vllm.entrypoints.openai.chat_completion.stream_harmony.make_tool_call_id")
     def test_new_tool_call(self, mock_make_tool_call_id, channel):
         """Test new tool call creation when recipient changes."""
         mock_make_tool_call_id.return_value = "call_test123"

tests/entrypoints/openai/test_serving_engine.py

@@ -8,7 +8,7 @@ from unittest.mock import Mock
 import pytest
 
 from vllm.config import ModelConfig
-from vllm.entrypoints.openai.serving_engine import OpenAIServing
+from vllm.entrypoints.openai.engine.serving import OpenAIServing
 from vllm.entrypoints.openai.serving_models import OpenAIServingModels
 from vllm.tokenizers.mistral import MistralTokenizer
 

tests/entrypoints/openai/test_serving_models.py

@@ -8,7 +8,7 @@ import pytest
 
 from vllm.config import ModelConfig
 from vllm.engine.protocol import EngineClient
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.engine.protocol import (
     ErrorResponse,
     LoadLoRAAdapterRequest,
     UnloadLoRAAdapterRequest,

tests/entrypoints/openai/test_serving_responses.py

@@ -14,7 +14,7 @@ from openai.types.responses.tool import (
 )
 
 from vllm.entrypoints.context import ConversationContext
-from vllm.entrypoints.openai.protocol import ErrorResponse, ResponsesRequest
+from vllm.entrypoints.openai.engine.protocol import ErrorResponse, ResponsesRequest
 from vllm.entrypoints.openai.serving_responses import (
     OpenAIServingResponses,
     _extract_allowed_tools_from_mcp_requests,

tests/entrypoints/openai/tool_parsers/test_gigachat3_tool_parser.py

@@ -9,7 +9,7 @@ from tests.entrypoints.openai.tool_parsers.utils import (
     run_tool_extraction,
     run_tool_extraction_streaming,
 )
-from vllm.entrypoints.openai.protocol import FunctionCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers import ToolParser, ToolParserManager
 

tests/entrypoints/openai/tool_parsers/test_hermes_tool_parser.py

@@ -5,7 +5,7 @@ import json
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.hermes_tool_parser import Hermes2ProToolParser
 

tests/entrypoints/openai/tool_parsers/test_hunyuan_a13b_tool_parser.py

@@ -11,7 +11,7 @@ from tests.entrypoints.openai.tool_parsers.utils import (
     run_tool_extraction,
     run_tool_extraction_streaming,
 )
-from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall, ToolCall
 from vllm.tool_parsers import ToolParser, ToolParserManager
 
 

tests/entrypoints/openai/tool_parsers/test_llama3_json_tool_parser.py

@@ -5,7 +5,7 @@ from unittest.mock import MagicMock, patch
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import ExtractedToolCallInformation
+from vllm.entrypoints.openai.engine.protocol import ExtractedToolCallInformation
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers.llama_tool_parser import Llama3JsonToolParser
 

tests/entrypoints/openai/tool_parsers/test_llama4_pythonic_tool_parser.py

@@ -9,7 +9,7 @@ from tests.entrypoints.openai.tool_parsers.utils import (
     run_tool_extraction,
     run_tool_extraction_streaming,
 )
-from vllm.entrypoints.openai.protocol import FunctionCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers import ToolParser, ToolParserManager
 

tests/entrypoints/openai/tool_parsers/test_olmo3_tool_parser.py

@@ -9,7 +9,7 @@ from tests.entrypoints.openai.tool_parsers.utils import (
     run_tool_extraction,
     run_tool_extraction_streaming,
 )
-from vllm.entrypoints.openai.protocol import FunctionCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers import ToolParser, ToolParserManager
 

tests/entrypoints/openai/tool_parsers/test_pythonic_tool_parser.py

@@ -9,7 +9,7 @@ from tests.entrypoints.openai.tool_parsers.utils import (
     run_tool_extraction,
     run_tool_extraction_streaming,
 )
-from vllm.entrypoints.openai.protocol import FunctionCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall
 from vllm.tokenizers import TokenizerLike
 from vllm.tool_parsers import ToolParser, ToolParserManager
 

tests/entrypoints/openai/tool_parsers/utils.py

@@ -3,8 +3,8 @@
 
 from collections.abc import Iterable
 
-from vllm.entrypoints.openai.protocol import (
-    ChatCompletionRequest,
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.engine.protocol import (
     DeltaMessage,
     ExtractedToolCallInformation,
     FunctionCall,

tests/entrypoints/openai/utils.py

@@ -4,11 +4,13 @@ import json
 from collections.abc import AsyncGenerator
 from typing import Any
 
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionResponse,
     ChatCompletionResponseChoice,
     ChatCompletionStreamResponse,
     ChatMessage,
+)
+from vllm.entrypoints.openai.engine.protocol import (
     UsageInfo,
 )
 

tests/reasoning/test_base_thinking_reasoning_parser.py

@@ -5,7 +5,7 @@ import pytest
 from transformers import AutoTokenizer
 
 from tests.reasoning.utils import run_reasoning_extraction
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.reasoning.basic_parsers import BaseThinkingReasoningParser
 
 

tests/reasoning/test_deepseekv3_reasoning_parser.py

@@ -4,7 +4,8 @@
 import pytest
 from transformers import AutoTokenizer
 
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest, DeltaMessage
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.engine.protocol import DeltaMessage
 from vllm.reasoning.deepseek_r1_reasoning_parser import DeepSeekR1ReasoningParser
 from vllm.reasoning.deepseek_v3_reasoning_parser import DeepSeekV3ReasoningParser
 from vllm.reasoning.identity_reasoning_parser import IdentityReasoningParser

tests/reasoning/utils.py

@@ -1,8 +1,8 @@
 # SPDX-License-Identifier: Apache-2.0
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
-
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest, DeltaMessage
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.engine.protocol import DeltaMessage
 from vllm.reasoning import ReasoningParser
 from vllm.tokenizers.mistral import MistralTokenizer
 

tests/tool_parsers/test_ernie45_moe_tool_parser.py

@@ -7,8 +7,8 @@ from collections.abc import Generator
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import (
-    ChatCompletionRequest,
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.engine.protocol import (
     DeltaMessage,
     FunctionCall,
     ToolCall,

tests/tool_parsers/test_functiongemma_tool_parser.py

@@ -5,7 +5,7 @@ from unittest.mock import MagicMock
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 from vllm.tool_parsers.functiongemma_tool_parser import FunctionGemmaToolParser
 
 

tests/tool_parsers/test_glm4_moe_tool_parser.py

@@ -6,7 +6,7 @@ import json
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall, ToolCall
 from vllm.tokenizers import get_tokenizer
 from vllm.tool_parsers.glm4_moe_tool_parser import (
     Glm4MoeModelToolParser,

tests/tool_parsers/test_jamba_tool_parser.py

@@ -8,7 +8,7 @@ import partial_json_parser
 import pytest
 from partial_json_parser.core.options import Allow
 
-from vllm.entrypoints.openai.protocol import DeltaMessage, FunctionCall, ToolCall
+from vllm.entrypoints.openai.engine.protocol import DeltaMessage, FunctionCall, ToolCall
 from vllm.tokenizers import TokenizerLike, get_tokenizer
 from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
 from vllm.tool_parsers.jamba_tool_parser import JambaToolParser

tests/tool_parsers/test_kimi_k2_tool_parser.py

@@ -6,7 +6,7 @@ import json
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall, ToolCall
 from vllm.tokenizers import get_tokenizer
 from vllm.tool_parsers.kimi_k2_tool_parser import KimiK2ToolParser
 

tests/tool_parsers/test_minimax_tool_parser.py

@@ -7,8 +7,10 @@ from typing import Any
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionToolsParam,
+)
+from vllm.entrypoints.openai.engine.protocol import (
     FunctionCall,
     ToolCall,
 )

tests/tool_parsers/test_mistral_tool_parser.py

@@ -11,7 +11,7 @@ from mistral_common.protocol.instruct.request import InstructRequest
 from mistral_common.protocol.instruct.tool_calls import FunctionCall, ToolCall
 from partial_json_parser.core.options import Allow
 
-from vllm.entrypoints.openai.protocol import DeltaMessage, DeltaToolCall
+from vllm.entrypoints.openai.engine.protocol import DeltaMessage, DeltaToolCall
 from vllm.tokenizers import TokenizerLike, get_tokenizer
 from vllm.tokenizers.detokenizer_utils import detokenize_incrementally
 from vllm.tokenizers.mistral import MistralTokenizer

tests/tool_parsers/test_openai_tool_parser.py

@@ -14,7 +14,7 @@ from openai_harmony import (
     load_harmony_encoding,
 )
 
-from vllm.entrypoints.openai.protocol import FunctionCall, ToolCall
+from vllm.entrypoints.openai.engine.protocol import FunctionCall, ToolCall
 from vllm.tokenizers import get_tokenizer
 from vllm.tool_parsers.openai_tool_parser import OpenAIToolParser
 

tests/tool_parsers/test_qwen3coder_tool_parser.py

@@ -6,9 +6,11 @@ from collections.abc import Generator
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
     ChatCompletionToolsParam,
+)
+from vllm.entrypoints.openai.engine.protocol import (
     DeltaMessage,
     FunctionCall,
     ToolCall,

tests/tool_parsers/test_seed_oss_tool_parser.py

@@ -7,9 +7,11 @@ from collections.abc import Generator
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
     ChatCompletionToolsParam,
+)
+from vllm.entrypoints.openai.engine.protocol import (
     DeltaMessage,
     FunctionCall,
     ToolCall,

tests/tool_parsers/test_xlam_tool_parser.py

@@ -6,8 +6,8 @@ from collections.abc import Generator
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import (
-    ChatCompletionRequest,
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.engine.protocol import (
     DeltaMessage,
     FunctionCall,
     ToolCall,

tests/tool_use/test_chat_completion_request_validations.py

@@ -3,7 +3,7 @@
 
 import pytest
 
-from vllm.entrypoints.openai.protocol import ChatCompletionRequest
+from vllm.entrypoints.openai.chat_completion.protocol import ChatCompletionRequest
 
 
 def test_chat_completion_request_with_no_tools():

tests/tool_use/test_tool_choice_required.py

@@ -8,10 +8,10 @@ import pytest
 import regex as re
 from pydantic import TypeAdapter
 
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionToolsParam,
 )
-from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
+from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
 from vllm.tool_parsers.utils import get_json_schema_from_tools
 
 pytestmark = pytest.mark.cpu_test

tests/v1/engine/test_async_llm.py

@@ -11,12 +11,14 @@ from vllm import SamplingParams
 from vllm.assets.image import ImageAsset
 from vllm.config import VllmConfig
 from vllm.engine.arg_utils import AsyncEngineArgs
-from vllm.entrypoints.openai.protocol import (
+from vllm.entrypoints.openai.chat_completion.protocol import (
     ChatCompletionRequest,
     ChatCompletionResponse,
+)
+from vllm.entrypoints.openai.chat_completion.serving import OpenAIServingChat
+from vllm.entrypoints.openai.engine.protocol import (
     ErrorResponse,
 )
-from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
 from vllm.entrypoints.openai.serving_models import BaseModelPath, OpenAIServingModels
 from vllm.inputs import PromptType
 from vllm.outputs import RequestOutput