[Bugfix] Realign Serving Tokenization Name to Fix Error Handling Routes (#4816)

Signed-off-by: Alex Brooks <albrooks@redhat.com>

tests/entrypoints/openai_api/test_serving_speech.py

@@ -2530,7 +2530,7 @@ def test_api_server_delete_voice_exception_returns_500(mocker: MockerFixture):
 def test_api_server_create_speech_without_handler_returns_404(mocker: MockerFixture):
     fake_base = _patch_api_server_base(mocker)
     raw_request = _make_api_server_request(None, path="/v1/audio/speech")
-    raw_request.app.state.openai_serving_tokenization = fake_base
+    raw_request.app.state.serving_tokenization = fake_base
     request = OpenAICreateSpeechRequest(input="Hello")
 
     response = asyncio.run(api_server_module.create_speech(request, raw_request))
@@ -2543,7 +2543,7 @@ def test_api_server_create_speech_without_handler_returns_404(mocker: MockerFixt
 def test_api_server_create_speech_batch_without_handler_returns_404(mocker: MockerFixture):
     fake_base = _patch_api_server_base(mocker)
     raw_request = _make_api_server_request(None, path="/v1/audio/speech/batch")
-    raw_request.app.state.openai_serving_tokenization = fake_base
+    raw_request.app.state.serving_tokenization = fake_base
     request = BatchSpeechRequest(items=[SpeechBatchItem(input="hi")])
 
     response = asyncio.run(api_server_module.create_speech_batch(request, raw_request))
@@ -2612,7 +2612,7 @@ def test_api_server_create_audio_generate_without_handler_returns_404(mocker: Mo
     fake_base = _patch_api_server_base(mocker)
     raw_request = _make_api_server_request(None, path="/v1/audio/generate")
     raw_request.app.state.openai_serving_audio_generate = None
-    raw_request.app.state.openai_serving_tokenization = fake_base
+    raw_request.app.state.serving_tokenization = fake_base
     request = OpenAICreateAudioGenerateRequest(input="a bird singing")
 
     response = asyncio.run(api_server_module.create_audio_generate(request, raw_request))

vllm_omni/entrypoints/openai/api_server.py

@@ -751,7 +751,7 @@ async def omni_init_app_state(
         state.diffusion_engine = engine_client
         state.openai_serving_models = _DiffusionServingModels(base_model_paths)
         # OMNI: tokenization endpoints are not supported in pure diffusion mode.
-        state.openai_serving_tokenization = None
+        state.serving_tokenization = None
 
         # Use for_diffusion method to create chat handler
         state.openai_serving_chat = OmniOpenAIServingChat.for_diffusion(
@@ -1014,7 +1014,7 @@ async def omni_init_app_state(
         if any(t in supported_tasks for t in ("embed", "score", "token_embed"))
         else None
     )
-    state.openai_serving_tokenization = ServingTokenization(
+    state.serving_tokenization = ServingTokenization(
         state.openai_serving_models,
         state.openai_serving_render,
         request_logger=request_logger,
@@ -1154,7 +1154,7 @@ async def create_chat_completion(request: ChatCompletionRequest, raw_request: Re
     metrics_header_format = raw_request.headers.get(ENDPOINT_LOAD_METRICS_FORMAT_HEADER_LABEL, "")
     handler = Omnichat(raw_request)
     if handler is None:
-        base_server = getattr(raw_request.app.state, "openai_serving_tokenization", None)
+        base_server = getattr(raw_request.app.state, "serving_tokenization", None)
         if base_server is None:
             raise HTTPException(
                 status_code=HTTPStatus.NOT_FOUND.value,
@@ -1245,7 +1245,7 @@ async def create_speech(request: OpenAICreateSpeechRequest, raw_request: Request
     """
     handler = Omnispeech(raw_request)
     if handler is None:
-        base_server = getattr(raw_request.app.state, "openai_serving_tokenization", None)
+        base_server = getattr(raw_request.app.state, "serving_tokenization", None)
         if base_server is None:
             raise HTTPException(
                 status_code=HTTPStatus.NOT_FOUND.value,
@@ -1283,7 +1283,7 @@ async def create_speech(request: OpenAICreateSpeechRequest, raw_request: Request
 async def create_speech_batch(request: BatchSpeechRequest, raw_request: Request):
     handler = Omnispeech(raw_request)
     if handler is None:
-        base_server = getattr(raw_request.app.state, "openai_serving_tokenization", None)
+        base_server = getattr(raw_request.app.state, "serving_tokenization", None)
         if base_server is None:
             raise HTTPException(
                 status_code=HTTPStatus.NOT_FOUND.value,
@@ -1326,7 +1326,7 @@ async def create_speech_batch(request: BatchSpeechRequest, raw_request: Request)
 async def create_audio_generate(request: OpenAICreateAudioGenerateRequest, raw_request: Request):
     handler = OmniAudioGenerate(raw_request)
     if handler is None:
-        base_server = getattr(raw_request.app.state, "openai_serving_tokenization", None)
+        base_server = getattr(raw_request.app.state, "serving_tokenization", None)
         if base_server is None:
             raise HTTPException(
                 status_code=HTTPStatus.NOT_FOUND.value,