agent-framework/python/samples/02-agents/middleware/session_behavior_middleware.py

# Copyright (c) Microsoft. All rights reserved.

import asyncio
from collections.abc import Awaitable, Callable
from typing import Annotated

from agent_framework import (
    AgentContext,
    InMemoryHistoryProvider,
    tool,
)
from agent_framework.azure import AzureOpenAIChatClient
from azure.identity import AzureCliCredential
from pydantic import Field

"""
Thread Behavior MiddlewareTypes Example

This sample demonstrates how middleware can access and track session state across multiple agent runs.
The example shows:

- How AgentContext.session property behaves across multiple runs
- How middleware can access conversation history through the session
- The timing of when session messages are populated (before vs after call_next() call)
- How to track session state changes across runs

Key behaviors demonstrated:
1. First run: context.messages is populated, context.session is initially empty (before call_next())
2. After call_next(): session contains input message + response from agent
3. Second run: context.messages contains only current input, session contains previous history
4. After call_next(): session contains full conversation history (all previous + current messages)
"""


# NOTE: approval_mode="never_require" is for sample brevity. Use "always_require" in production; see samples/02-agents/tools/function_tool_with_approval.py and samples/02-agents/tools/function_tool_with_approval_and_sessions.py.
@tool(approval_mode="never_require")
def get_weather(
    location: Annotated[str, Field(description="The location to get the weather for.")],
) -> str:
    """Get the weather for a given location."""
    from random import randint

    conditions = ["sunny", "cloudy", "rainy", "stormy"]
    return f"The weather in {location} is {conditions[randint(0, 3)]} with a high of {randint(10, 30)}°C."


async def thread_tracking_middleware(
    context: AgentContext,
    call_next: Callable[[], Awaitable[None]],
) -> None:
    """MiddlewareTypes that tracks and logs session behavior across runs."""
    session_message_count = 0
    if context.session:
        memory_state = context.session.state.get(InMemoryHistoryProvider.DEFAULT_SOURCE_ID, {})
        session_message_count = len(memory_state.get("messages", []))

    print(f"[MiddlewareTypes pre-execution] Current input messages: {len(context.messages)}")
    print(f"[MiddlewareTypes pre-execution] Session history messages: {session_message_count}")

    # Call call_next to execute the agent
    await call_next()

    # Check session state after agent execution
    updated_session_message_count = 0
    if context.session:
        memory_state = context.session.state.get(InMemoryHistoryProvider.DEFAULT_SOURCE_ID, {})
        updated_session_message_count = len(memory_state.get("messages", []))

    print(f"[MiddlewareTypes post-execution] Updated session messages: {updated_session_message_count}")


async def main() -> None:
    """Example demonstrating session behavior in middleware across multiple runs."""
    print("=== Session Behavior MiddlewareTypes Example ===")

    # For authentication, run `az login` command in terminal or replace AzureCliCredential with preferred
    # authentication option.
    agent = AzureOpenAIChatClient(credential=AzureCliCredential()).as_agent(
        name="WeatherAgent",
        instructions="You are a helpful weather assistant.",
        tools=get_weather,
        middleware=[thread_tracking_middleware],
    )

    # Create a session that will persist messages between runs
    session = agent.create_session()

    print("\nFirst Run:")
    query1 = "What's the weather like in Tokyo?"
    print(f"User: {query1}")
    result1 = await agent.run(query1, session=session)
    print(f"Agent: {result1.text}")

    print("\nSecond Run:")
    query2 = "How about in London?"
    print(f"User: {query2}")
    result2 = await agent.run(query2, session=session)
    print(f"Agent: {result2.text}")


if __name__ == "__main__":
    asyncio.run(main())