[Tool] adjust_request to reasoning parser, and Gemma4 fixes (#39027)

Signed-off-by: Ben Browning <bbrownin@redhat.com>
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
Ben Browning
2026-04-08 19:04:04 +00:00
committed by GitHub
co-authored by Claude Opus 4.6 Cursor
parent e24e0a43a4
commit 8477fe427d
12 changed files with 878 additions and 16 deletions
+2 -1
View File
@@ -170,7 +170,8 @@ class AnthropicServingMessages(OpenAIServingChat):
else:
cls._convert_message_content(msg, openai_msg, openai_messages)
openai_messages.append(openai_msg)
if not (msg.role == "user" and "content" not in openai_msg):
openai_messages.append(openai_msg)
@classmethod
def _convert_message_content(
+2
View File
@@ -379,6 +379,7 @@ async def init_app_state(
enable_auto_tools=args.enable_auto_tool_choice,
exclude_tools_when_tool_choice_none=args.exclude_tools_when_tool_choice_none,
tool_parser=args.tool_call_parser,
reasoning_parser=args.structured_outputs_config.reasoning_parser,
default_chat_template_kwargs=args.default_chat_template_kwargs,
log_error_stack=args.log_error_stack,
)
@@ -481,6 +482,7 @@ async def init_render_app_state(
enable_auto_tools=args.enable_auto_tool_choice,
exclude_tools_when_tool_choice_none=args.exclude_tools_when_tool_choice_none,
tool_parser=args.tool_call_parser,
reasoning_parser=args.structured_outputs_config.reasoning_parser,
default_chat_template_kwargs=args.default_chat_template_kwargs,
log_error_stack=args.log_error_stack,
)
@@ -594,6 +594,7 @@ class OpenAIServingResponses(OpenAIServing):
default_template_kwargs=None,
tool_dicts=tool_dicts,
tool_parser=self.parser.tool_parser_cls if self.parser else None,
reasoning_parser=self.parser.reasoning_parser_cls if self.parser else None,
)
return messages, engine_inputs
@@ -618,6 +619,7 @@ class OpenAIServingResponses(OpenAIServing):
default_template_kwargs=None,
tool_dicts=tool_dicts,
tool_parser=tool_parser,
reasoning_parser=self.parser.reasoning_parser_cls if self.parser else None,
)
return engine_inputs
+13
View File
@@ -44,6 +44,7 @@ from vllm.inputs import (
)
from vllm.logger import init_logger
from vllm.parser import ParserManager
from vllm.reasoning.abs_reasoning_parsers import ReasoningParser
from vllm.renderers import BaseRenderer, merge_kwargs
from vllm.renderers.inputs.preprocess import (
extract_prompt_components,
@@ -74,6 +75,7 @@ class OpenAIServingRender:
enable_auto_tools: bool = False,
exclude_tools_when_tool_choice_none: bool = False,
tool_parser: str | None = None,
reasoning_parser: str | None = None,
default_chat_template_kwargs: dict[str, Any] | None = None,
log_error_stack: bool = False,
) -> None:
@@ -94,6 +96,11 @@ class OpenAIServingRender:
enable_auto_tools=enable_auto_tools,
model_name=model_config.model,
)
self.reasoning_parser: type[ReasoningParser] | None = (
ParserManager.get_reasoning_parser(
reasoning_parser_name=reasoning_parser,
)
)
self.default_chat_template_kwargs: dict[str, Any] = (
default_chat_template_kwargs or {}
)
@@ -245,6 +252,7 @@ class OpenAIServingRender:
default_template_kwargs=self.default_chat_template_kwargs,
tool_dicts=tool_dicts,
tool_parser=tool_parser,
reasoning_parser=self.reasoning_parser,
)
else:
# For GPT-OSS.
@@ -503,6 +511,7 @@ class OpenAIServingRender:
default_template_kwargs: dict[str, Any] | None,
tool_dicts: list[dict[str, Any]] | None = None,
tool_parser: type[ToolParser] | None = None,
reasoning_parser: type[ReasoningParser] | None = None,
*,
skip_mm_cache: bool = False,
) -> tuple[list[ConversationMessage], list[EngineInput]]:
@@ -539,6 +548,10 @@ class OpenAIServingRender:
skip_mm_cache=skip_mm_cache,
)
if reasoning_parser is not None:
tokenizer = renderer.get_tokenizer()
request = reasoning_parser(tokenizer).adjust_request(request=request)
# tool parsing is done only if a tool_parser has been set and if
# tool_choice is not "none" (if tool_choice is "none" but a tool_parser
# is set, we want to prevent parsing a tool_call hallucinated by the LLM