[Refactor] [10/N] to simplify the vLLM openai completion serving architecture (#32369)

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

vllm/entrypoints/serve/tokenize/api_router.py

@@ -9,10 +9,10 @@ from fastapi.exceptions import RequestValidationError
 from fastapi.responses import JSONResponse
 from typing_extensions import assert_never
 
-from vllm.entrypoints.openai.api_server import validate_json_request
 from vllm.entrypoints.openai.engine.protocol import (
     ErrorResponse,
 )
+from vllm.entrypoints.openai.utils import validate_json_request
 from vllm.entrypoints.serve.tokenize.protocol import (
     DetokenizeRequest,
     DetokenizeResponse,

vllm/entrypoints/serve/tokenize/serving.py

@@ -13,7 +13,7 @@ from vllm.entrypoints.openai.engine.protocol import (
     ErrorResponse,
 )
 from vllm.entrypoints.openai.engine.serving import OpenAIServing
-from vllm.entrypoints.openai.serving_models import OpenAIServingModels
+from vllm.entrypoints.openai.models.serving import OpenAIServingModels
 from vllm.entrypoints.renderer import RenderConfig
 from vllm.entrypoints.serve.tokenize.protocol import (
     DetokenizeRequest,

vllm/entrypoints/utils.py

@@ -31,11 +31,13 @@ if TYPE_CHECKING:
     from vllm.entrypoints.openai.chat_completion.protocol import (
         ChatCompletionRequest,
     )
-    from vllm.entrypoints.openai.engine.protocol import (
+    from vllm.entrypoints.openai.completion.protocol import (
         CompletionRequest,
+    )
+    from vllm.entrypoints.openai.engine.protocol import (
         StreamOptions,
     )
-    from vllm.entrypoints.openai.serving_models import LoRAModulePath
+    from vllm.entrypoints.openai.models.protocol import LoRAModulePath
 else:
     ChatCompletionRequest = object
     CompletionRequest = object
@@ -281,7 +283,7 @@ def should_include_usage(
 def process_lora_modules(
     args_lora_modules: list[LoRAModulePath], default_mm_loras: dict[str, str] | None
 ) -> list[LoRAModulePath]:
-    from vllm.entrypoints.openai.serving_models import LoRAModulePath
+    from vllm.entrypoints.openai.models.serving import LoRAModulePath
 
     lora_modules = args_lora_modules
     if default_mm_loras: