[Bugfix][Frontend] Forward the reasoning wiring to the engine for batch chat completions (#59263)

Signed-off-by: apex-mochen <2756823972@qq.com>
Signed-off-by: sfeng33 <4florafeng@gmail.com>
Co-authored-by: Flora Feng <4florafeng@gmail.com>
This commit is contained in:
dingdangmao
2026-10-08 19:32:46 +00:00
committed by GitHub
co-authored by Flora Feng
parent e3a29aa9b3
commit bab0ac61f8
3 changed files with 139 additions and 13 deletions
@@ -169,6 +169,12 @@ class OpenAIServingChatBatch(OpenAIServingChat):
else await self._get_trace_headers(raw_request.headers)
)
session_id = self._get_session_id(single_request, raw_request)
chat_template_kwargs = self._effective_chat_template_kwargs(single_request)
parser = (
self._make_parser(single_request, tokenizer, chat_template_kwargs)
if self.parser_cls is not None
else None
)
generators.append(
self.engine_client.generate(
engine_prompt,
@@ -179,7 +185,12 @@ class OpenAIServingChatBatch(OpenAIServingChat):
priority=request.priority,
data_parallel_rank=data_parallel_rank,
session_id=session_id,
reasoning_ended=None,
**self._engine_reasoning_kwargs(
single_request,
parser,
chat_template_kwargs,
self._extract_prompt_components(engine_prompt).token_ids,
),
)
)