phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-03 18:00:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Fix max_tool_calls for openai and add integration tests for the feat

Browse source
This commit is contained in:
Shabana Baig 2025-11-18 16:12:31 -05:00
parent a3580e6bc0
commit 67d10a9c7c
3 changed files with 161 additions and 168 deletions
Download patch file Download diff file Expand all files Collapse all files

11
src/llama_stack/providers/inline/agents/meta_reference/responses/streaming.py
View file

@ -66,6 +66,7 @@ from llama_stack_api import (
OpenAIResponseUsage,
OpenAIResponseUsageInputTokensDetails,
OpenAIResponseUsageOutputTokensDetails,
OpenAIToolMessageParam,
WebSearchToolTypes,
)
@ -905,10 +906,16 @@ class StreamingResponseOrchestrator:
"""Coordinate execution of both function and non-function tool calls."""
# Execute non-function tool calls
for tool_call in non_function_tool_calls:
# Check if total calls made to built-in and mcp tools exceed max_tool_calls
# if total calls made to built-in and mcp tools exceed max_tool_calls
# then create a tool response message indicating the call was skipped
if self.max_tool_calls is not None and self.accumulated_builtin_tool_calls >= self.max_tool_calls:
logger.info(f"Ignoring built-in and mcp tool call since reached the limit of {self.max_tool_calls=}.")
break
skipped_call_message = OpenAIToolMessageParam(
content=f"Tool call skipped: maximum tool calls limit ({self.max_tool_calls}) reached.",
tool_call_id=tool_call.id,
)
next_turn_messages.append(skipped_call_message)
continue
# Find the item_id for this tool call
matching_item_id = None