Python: Introducing Local MCP Servers (#389)

* mcp parts

* mcp parts 2

* removed structured output in favor of handling in chatresponse, mcp as AITool and running samples

* updated naming

* fixed test

python/samples/getting_started/agents/openai_chat_client/openai_chat_client_with_local_mcp.py

@@ -0,0 +1,76 @@
+# Copyright (c) Microsoft. All rights reserved.
+
+import asyncio
+
+from agent_framework import ChatClientAgent, McpStreamableHttpTool
+from agent_framework.openai import OpenAIChatClient
+
+
+async def mcp_tools_on_run_level() -> None:
+    """Example showing MCP tools defined when running the agent."""
+    print("=== Tools Defined on Run Level ===")
+
+    # Tools are provided when running the agent
+    # This means we have to ensure we connect to the MCP server before running the agent
+    # and pass the tools to the run method.
+    async with (
+        McpStreamableHttpTool(
+            name="Microsoft Learn MCP",
+            url="https://learn.microsoft.com/api/mcp",
+        ) as mcp_server,
+        ChatClientAgent(
+            chat_client=OpenAIChatClient(),
+            name="DocsAgent",
+            instructions="You are a helpful assistant that can help with microsoft documentation questions.",
+        ) as agent,
+    ):
+        # First query
+        query1 = "How to create an Azure storage account using az cli?"
+        print(f"User: {query1}")
+        result1 = await agent.run(query1, tools=mcp_server)
+        print(f"{agent.name}: {result1}\n")
+        print("\n=======================================\n")
+        # Second query
+        query2 = "What is Microsoft Semantic Kernel?"
+        print(f"User: {query2}")
+        result2 = await agent.run(query2, tools=mcp_server)
+        print(f"{agent.name}: {result2}\n")
+
+
+async def mcp_tools_on_agent_level() -> None:
+    """Example showing tools defined when creating the agent."""
+    print("=== Tools Defined on Agent Level ===")
+
+    # Tools are provided when creating the agent
+    # The agent can use these tools for any query during its lifetime
+    # The agent will connect to the MCP server through its context manager.
+    async with OpenAIChatClient().create_agent(
+        name="DocsAgent",
+        instructions="You are a helpful assistant that can help with microsoft documentation questions.",
+        tools=McpStreamableHttpTool(  # Tools defined at agent creation
+            name="Microsoft Learn MCP",
+            url="https://learn.microsoft.com/api/mcp",
+        ),
+    ) as agent:
+        # First query
+        query1 = "How to create an Azure storage account using az cli?"
+        print(f"User: {query1}")
+        result1 = await agent.run(query1)
+        print(f"{agent.name}: {result1}\n")
+        print("\n=======================================\n")
+        # Second query
+        query2 = "What is Microsoft Semantic Kernel?"
+        print(f"User: {query2}")
+        result2 = await agent.run(query2)
+        print(f"{agent.name}: {result2}\n")
+
+
+async def main() -> None:
+    print("=== OpenAI Chat Client Agent with MCP Tools Examples ===\n")
+
+    await mcp_tools_on_agent_level()
+    await mcp_tools_on_run_level()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

python/samples/getting_started/agents/openai_responses_client/openai_responses_client_with_local_mcp.py

@@ -0,0 +1,87 @@
+# Copyright (c) Microsoft. All rights reserved.
+
+import asyncio
+
+from agent_framework import ChatClientAgent, McpStreamableHttpTool
+from agent_framework.openai import OpenAIResponsesClient
+
+
+async def streaming_with_mcp(show_raw_stream: bool = False) -> None:
+    """Example showing tools defined when creating the agent.
+
+    If you want to access the full stream of events that has come from the model, you can access it,
+    through the raw_representation. You can view this, by setting the show_raw_stream parameter to True.
+    """
+    print("=== Tools Defined on Agent Level ===")
+
+    # Tools are provided when creating the agent
+    # The agent can use these tools for any query during its lifetime
+    async with ChatClientAgent(
+        chat_client=OpenAIResponsesClient(),
+        name="DocsAgent",
+        instructions="You are a helpful assistant that can help with microsoft documentation questions.",
+        tools=McpStreamableHttpTool(  # Tools defined at agent creation
+            name="Microsoft Learn MCP",
+            url="https://learn.microsoft.com/api/mcp",
+        ),
+    ) as agent:
+        # First query
+        query1 = "How to create an Azure storage account using az cli?"
+        print(f"User: {query1}")
+        print(f"{agent.name}: ", end="")
+        async for chunk in agent.run_streaming(query1):
+            if show_raw_stream:
+                print("Streamed event: ", chunk.raw_representation.raw_representation)  # type:ignore
+            elif chunk.text:
+                print(chunk.text, end="")
+        print("")
+        print("\n=======================================\n")
+        # Second query
+        query2 = "What is Microsoft Semantic Kernel?"
+        print(f"User: {query2}")
+        print(f"{agent.name}: ", end="")
+        async for chunk in agent.run_streaming(query2):
+            if show_raw_stream:
+                print("Streamed event: ", chunk.raw_representation.raw_representation)  # type:ignore
+            elif chunk.text:
+                print(chunk.text, end="")
+        print("\n\n")
+
+
+async def run_with_mcp() -> None:
+    """Example showing tools defined when creating the agent."""
+    print("=== Tools Defined on Agent Level ===")
+
+    # Tools are provided when creating the agent
+    # The agent can use these tools for any query during its lifetime
+    async with ChatClientAgent(
+        chat_client=OpenAIResponsesClient(),
+        name="DocsAgent",
+        instructions="You are a helpful assistant that can help with microsoft documentation questions.",
+        tools=McpStreamableHttpTool(  # Tools defined at agent creation
+            name="Microsoft Learn MCP",
+            url="https://learn.microsoft.com/api/mcp",
+        ),
+    ) as agent:
+        # First query
+        query1 = "How to create an Azure storage account using az cli?"
+        print(f"User: {query1}")
+        result1 = await agent.run(query1)
+        print(f"{agent.name}: {result1}\n")
+        print("\n=======================================\n")
+        # Second query
+        query2 = "What is Microsoft Semantic Kernel?"
+        print(f"User: {query2}")
+        result2 = await agent.run(query2)
+        print(f"{agent.name}: {result2}\n")
+
+
+async def main() -> None:
+    print("=== OpenAI Responses Client Agent with Function Tools Examples ===\n")
+
+    await run_with_mcp()
+    await streaming_with_mcp()
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

python/samples/getting_started/chat_client/azure_responses_client.py

@@ -4,9 +4,10 @@ import asyncio
 from random import randint
 from typing import Annotated
 
+from agent_framework import ChatResponse
 from agent_framework.azure import AzureResponsesClient
 from azure.identity import DefaultAzureCredential
-from pydantic import Field
+from pydantic import BaseModel, Field
 
 
 def get_weather(
@@ -17,20 +18,28 @@ def get_weather(
     return f"The weather in {location} is {conditions[randint(0, 3)]} with a high of {randint(10, 30)}°C."
 
 
+class OutputStruct(BaseModel):
+    """Structured output for weather information."""
+
+    location: str
+    weather: str
+
+
 async def main() -> None:
     client = AzureResponsesClient(ad_credential=DefaultAzureCredential())
     message = "What's the weather in Amsterdam and in Paris?"
-    stream = False
+    stream = True
     print(f"User: {message}")
     if stream:
-        print("Assistant: ", end="")
-        async for chunk in client.get_streaming_response(message, tools=get_weather):
-            if str(chunk):
-                print(str(chunk), end="")
-        print("")
+        response = await ChatResponse.from_chat_response_generator(
+            client.get_streaming_response(message, tools=get_weather, response_format=OutputStruct),
+            output_format_type=OutputStruct,
+        )
+        print(f"Assistant: {response.value}")
+
     else:
-        response = await client.get_response(message, tools=get_weather)
-        print(f"Assistant: {response}")
+        response = await client.get_response(message, tools=get_weather, response_format=OutputStruct)
+        print(f"Assistant: {response.value}")
 
 
 if __name__ == "__main__":