[Frontend] Support reasoning content for deepseek r1 (#12473)

Signed-off-by: Ce Gao <cegao@tensorchord.ai>
Co-authored-by: Rafael Vasquez <rafvasq21@gmail.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Co-authored-by: Michael Goin <mgoin@redhat.com>

vllm/entrypoints/openai/api_server.py

@@ -61,6 +61,7 @@ from vllm.entrypoints.openai.protocol import (ChatCompletionRequest,
                                               TokenizeRequest,
                                               TokenizeResponse,
                                               UnloadLoraAdapterRequest)
+from vllm.entrypoints.openai.reasoning_parsers import ReasoningParserManager
 # yapf: enable
 from vllm.entrypoints.openai.serving_chat import OpenAIServingChat
 from vllm.entrypoints.openai.serving_completion import OpenAIServingCompletion
@@ -771,6 +772,8 @@ async def init_app_state(
         return_tokens_as_token_ids=args.return_tokens_as_token_ids,
         enable_auto_tools=args.enable_auto_tool_choice,
         tool_parser=args.tool_call_parser,
+        enable_reasoning=args.enable_reasoning,
+        reasoning_parser=args.reasoning_parser,
         enable_prompt_tokens_details=args.enable_prompt_tokens_details,
     ) if model_config.runner_type == "generate" else None
     state.openai_serving_completion = OpenAIServingCompletion(
@@ -844,6 +847,13 @@ async def run_server(args, **uvicorn_kwargs) -> None:
         raise KeyError(f"invalid tool call parser: {args.tool_call_parser} "
                        f"(chose from {{ {','.join(valid_tool_parses)} }})")
 
+    valid_reasoning_parses = ReasoningParserManager.reasoning_parsers.keys()
+    if args.enable_reasoning \
+        and args.reasoning_parser not in valid_reasoning_parses:
+        raise KeyError(
+            f"invalid reasoning parser: {args.reasoning_parser} "
+            f"(chose from {{ {','.join(valid_reasoning_parses)} }})")
+
     # workaround to make sure that we bind the port before the engine is set up.
     # This avoids race conditions with ray.
     # see https://github.com/vllm-project/vllm/issues/8204