[Frontend] Refine the entrypoint class's inheritance hierarchy. (#47498)

Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
This commit is contained in:
wang.yuqi
2026-07-03 10:50:06 +00:00
committed by GitHub
parent 18f658bb31
commit a14f57a3ac
23 changed files with 90 additions and 191 deletions
@@ -7,7 +7,7 @@ from unittest.mock import AsyncMock, Mock
import pytest
from vllm.entrypoints.speech_to_text.base.serving import OpenAISpeechToText
from vllm.entrypoints.speech_to_text.base.serving import SpeechToTextBaseServing
from vllm.entrypoints.speech_to_text.transcription.protocol import TranscriptionResponse
@@ -43,7 +43,7 @@ async def test_non_streaming_cancel_aborts_engine_requests(
is_tracing_enabled=AsyncMock(return_value=False),
)
server = OpenAISpeechToText.__new__(OpenAISpeechToText)
server = SpeechToTextBaseServing.__new__(SpeechToTextBaseServing)
server.engine_client = engine_client
server.task_type = "transcribe"
server.models = SimpleNamespace(model_name=lambda: "audio")
@@ -112,7 +112,7 @@ async def test_non_streaming_cancel_advances_all_chunk_generators():
{"prompt": "chunk-1"},
{"prompt": "chunk-2"},
]
server = OpenAISpeechToText.__new__(OpenAISpeechToText)
server = SpeechToTextBaseServing.__new__(SpeechToTextBaseServing)
server.engine_client = engine_client
server.task_type = "transcribe"
server.models = SimpleNamespace(model_name=lambda: "audio")
@@ -170,7 +170,7 @@ async def test_language_detection_cancel_aborts_engine_request():
abort=AsyncMock(),
)
server = OpenAISpeechToText.__new__(OpenAISpeechToText)
server = SpeechToTextBaseServing.__new__(SpeechToTextBaseServing)
server.engine_client = engine_client
server.asr_config = SimpleNamespace()
server.tokenizer = Mock()
@@ -25,7 +25,7 @@ from vllm.entrypoints.openai.engine.protocol import (
)
from vllm.entrypoints.openai.models.serving import OpenAIServingModels
from vllm.entrypoints.speech_to_text.base.serving import (
OpenAISpeechToText,
SpeechToTextBaseServing,
asr_inter_chunk_separator,
)
from vllm.entrypoints.speech_to_text.transcription.protocol import TranscriptionRequest
@@ -234,7 +234,9 @@ async def test_create_transcription_non_streaming_joins_chunks_by_language():
"vllm.model_executor.model_loader.get_model_cls",
return_value=_StubTranscriptionModel,
),
patch.object(OpenAISpeechToText, "_preprocess_speech_to_text", preprocess_mock),
patch.object(
SpeechToTextBaseServing, "_preprocess_speech_to_text", preprocess_mock
),
):
serving = OpenAIServingTranscription(engine_client, models, request_logger=None)