xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-11 06:45:01 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] [issue-21565] Fix the incompatibility issue with stream and named function calling when Thinking is disabled (#21573)

Browse Source 
Signed-off-by: wangzi <3220100013@zju.edu.cn>
Co-authored-by: wangzi <3220100013@zju.edu.cn>
This commit is contained in:
Hongsheng Liu 2025-07-28 13:43:50 +08:00 committed by GitHub
parent 3ea57a56d9
commit 7656cf4cf3
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 23 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
tests/entrypoints/openai/test_completion_with_function_calling.py
View File

@ -1,6 +1,8 @@
# SPDX-License-Identifier: Apache-2.0 # SPDX-License-Identifier: Apache-2.0
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
from typing import Union
import openai # use the official client for correctness check import openai # use the official client for correctness check
import pytest import pytest
import pytest_asyncio import pytest_asyncio
@ -40,10 +42,17 @@ async def client(server):
@pytest.mark.asyncio @pytest.mark.asyncio
@pytest.mark.parametrize("model_name", [MODEL_NAME]) @pytest.mark.parametrize("model_name", [MODEL_NAME])
@pytest.mark.parametrize("stream", [True, False]) @pytest.mark.parametrize("stream", [True, False])
@pytest.mark.parametrize("tool_choice", ["auto", "required"]) @pytest.mark.parametrize("tool_choice", [
"auto", "required", {
"type": "function",
"function": {
"name": "get_current_weather"
}
}
])
@pytest.mark.parametrize("enable_thinking", [True, False]) @pytest.mark.parametrize("enable_thinking", [True, False])
async def test_function_tool_use(client: openai.AsyncOpenAI, model_name: str, async def test_function_tool_use(client: openai.AsyncOpenAI, model_name: str,
stream: bool, tool_choice: str, stream: bool, tool_choice: Union[str, dict],
enable_thinking: bool): enable_thinking: bool):
tools = [ tools = [
{ {

17
vllm/entrypoints/openai/serving_chat.py
View File

@ -623,7 +623,7 @@ class OpenAIServingChat(OpenAIServing):
# handle streaming deltas for tools with named tool_choice # handle streaming deltas for tools with named tool_choice
if tool_choice_function_name: if tool_choice_function_name:
if (self.reasoning_parser if (self.reasoning_parser and not reasoning_end_arr[i]
and not reasoning_parser.is_reasoning_end( and not reasoning_parser.is_reasoning_end(
previous_token_ids)): previous_token_ids)):
assert reasoning_parser is not None assert reasoning_parser is not None
@ -637,11 +637,18 @@ class OpenAIServingChat(OpenAIServing):
current_token_ids, current_token_ids,
output.token_ids, output.token_ids,
)) ))
# When encountering think end id in delta_token_ids, # When encountering think end id in delta_token_ids
# process the `content`. Only keep 'content', # or think end id in prompt_token_ids
# remove 'reasoning_content' # i.e {"enable_thinking": False},
# set reasoning status to end.
# Only keep 'content', remove 'reasoning_content'.
if reasoning_parser.is_reasoning_end( if reasoning_parser.is_reasoning_end(
list(output.token_ids)): list(output.token_ids)) or \
(res.prompt_token_ids and
reasoning_parser.is_reasoning_end(
list(res.prompt_token_ids)
)):
reasoning_end_arr[i] = True
if delta_message and delta_message.content: if delta_message and delta_message.content:
# This need to be added to next `delta_text` # This need to be added to next `delta_text`
current_text = delta_message.content current_text = delta_message.content