[Refactor] [8/N] to simplify the vLLM openai responsesapi_serving architecture (#32260)

Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
This commit is contained in:
Chauncey
2026-01-14 07:26:24 +00:00
committed by GitHub
parent 6388b50058
commit 9312a6c03a
21 changed files with 754 additions and 674 deletions
@@ -7,7 +7,7 @@ import json
import pytest
from openai.types.responses import ResponseFunctionToolCall
from vllm.entrypoints.openai.engine.protocol import ResponsesRequest
from vllm.entrypoints.openai.responses.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.engine.protocol.logger") as mock_logger:
with patch("vllm.entrypoints.openai.responses.protocol.logger") as mock_logger:
with pytest.raises(ValueError):
ResponsesRequest(**request_data)
+1 -1
View File
@@ -4,7 +4,7 @@ from openai_harmony import (
Message,
)
from vllm.entrypoints.openai.engine.protocol import (
from vllm.entrypoints.openai.responses.protocol import (
serialize_message,
serialize_messages,
)
@@ -14,8 +14,9 @@ from openai.types.responses.tool import (
)
from vllm.entrypoints.context import ConversationContext
from vllm.entrypoints.openai.engine.protocol import ErrorResponse, ResponsesRequest
from vllm.entrypoints.openai.serving_responses import (
from vllm.entrypoints.openai.engine.protocol import ErrorResponse
from vllm.entrypoints.openai.responses.protocol import ResponsesRequest
from vllm.entrypoints.openai.responses.serving import (
OpenAIServingResponses,
_extract_allowed_tools_from_mcp_requests,
extract_tool_types,