[Doc] Reorganize online pooling APIs (#11172)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

vllm/entrypoints/openai/api_server.py

@@ -406,7 +406,7 @@ async def create_embedding(request: EmbeddingRequest, raw_request: Request):
     assert_never(generator)
 
 
-@router.post("/v1/score")
+@router.post("/score")
 async def create_score(request: ScoreRequest, raw_request: Request):
     handler = score(raw_request)
     if handler is None:
@@ -423,6 +423,15 @@ async def create_score(request: ScoreRequest, raw_request: Request):
     assert_never(generator)
 
 
+@router.post("/v1/score")
+async def create_score_v1(request: ScoreRequest, raw_request: Request):
+    logger.warning(
+        "To indicate that Score API is not part of standard OpenAI API, we "
+        "have moved it to `/score`. Please update your client accordingly.")
+
+    return await create_score(request, raw_request)
+
+
 if envs.VLLM_TORCH_PROFILER_DIR:
     logger.warning(
         "Torch Profiler is enabled in the API server. This should ONLY be "

vllm/entrypoints/openai/protocol.py

@@ -812,10 +812,11 @@ class ScoreRequest(OpenAIBaseModel):
     text_2: Union[List[str], str]
     truncate_prompt_tokens: Optional[Annotated[int, Field(ge=1)]] = None
 
-    # doc: begin-chat-embedding-pooling-params
+    # doc: begin-score-pooling-params
     additional_data: Optional[Any] = None
-    # doc: end-chat-embedding-pooling-params
+    # doc: end-score-pooling-params
 
+    # doc: begin-score-extra-params
     priority: int = Field(
         default=0,
         description=(
@@ -823,6 +824,8 @@ class ScoreRequest(OpenAIBaseModel):
             "default: 0). Any priority other than 0 will raise an error "
             "if the served model does not use priority scheduling."))
 
+    # doc: end-score-extra-params
+
     def to_pooling_params(self):
         return PoolingParams(additional_data=self.additional_data)