agent/agent-framework
2
0
Fork 0
mirror of https://github.com/microsoft/agent-framework.git synced 2026-06-16 21:04:09 +08:00
Code Issues Packages Projects Releases Wiki Activity

Python: Introducing UserInputRequest and Response types and HostedMcpTool (#405)

Browse Source 
* initial work on User Approval (and hosted mcp to validate)

* small update to the comments in the sample

* enable local MCP tools in chatClient get methods

* working streaming and improved setup

* fix for pyright

* updated create_approval -> create_response method

* added tests

* updated HostedMcpTool and addressed feedback

* update type name

* naming updates

* small docstring update

* mypy fix

* fixes and updates

* fixes for responses

* fix int tests

* removed broken tests

* updated test running

* removed specific content check on websearch

* increased timeout

* split slow foundry test

* don't parallel run samples

* add dist load to unit tests

---------

Co-authored-by: Eric Zhu <ekzhu@users.noreply.github.com>
This commit is contained in:
Eduard van Valkenburg
2025-09-10 15:37:34 +02:00
committed by GitHub
Unverified
parent 947f2bf642
commit 6aa746d891
21 changed files with 1186 additions and 447 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/packages/foundry/tests/test_foundry_chat_client.py
-40
View File
@@ -22,7 +22,6 @@ from agent_framework import (
Role,
TextContent,
UriContent,
ai_function,
)
from agent_framework import __version__ as AF_VERSION
from agent_framework.exceptions import ServiceInitializationError
@@ -933,42 +932,3 @@ async def test_foundry_chat_client_agent_level_tool_persistence():
assert second_response.text is not None
# Should use the agent-level weather tool again
assert any(term in second_response.text.lower() for term in ["miami", "sunny", "25"])
@skip_if_foundry_integration_tests_disabled
async def test_foundry_chat_client_run_level_tool_isolation():
"""Test that run-level tools are isolated to specific runs and don't persist with FoundryChatClient."""
# Counter to track how many times the weather tool is called
call_count = 0
@ai_function
async def get_weather_with_counter(location: Annotated[str, "The location as a city name"]) -> str:
"""Get the current weather in a given location."""
nonlocal call_count
call_count += 1
return f"The weather in {location} is sunny and 25°C."
async with ChatAgent(
chat_client=FoundryChatClient(async_credential=AzureCliCredential()),
instructions="You are a helpful assistant.",
) as agent:
# First run - use run-level tool
first_response = await agent.run(
"What's the weather like in Chicago?",
tools=[get_weather_with_counter], # Run-level tool
)
assert isinstance(first_response, AgentRunResponse)
assert first_response.text is not None
# Should use the run-level weather tool (call count should be 1)
assert call_count == 1
assert any(term in first_response.text.lower() for term in ["chicago", "sunny", "25"])
# Second run - run-level tool should NOT persist (key isolation test)
second_response = await agent.run("What's the weather like in Miami?")
assert isinstance(second_response, AgentRunResponse)
assert second_response.text is not None
# Should NOT use the weather tool since it was only run-level in previous call
# Call count should still be 1 (no additional calls)
assert call_count == 1