目录
-
- [1. 代码](#1. 代码)
- [2. 代码解释](#2. 代码解释)
- [3. 类似的例子](#3. 类似的例子)
1. 代码
python
from dataclasses import dataclass
from autogen_core import AgentId, MessageContext, RoutedAgent, message_handler
@dataclass
class MyMessageType:
content: str
class MyAgent(RoutedAgent):
def __init__(self) -> None:
super().__init__("MyAgent")
@message_handler
async def handle_my_message_type(self, message: MyMessageType, ctx: MessageContext) -> None:
print(f"{self.id.type} received message: {message.content}")
python
from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.messages import TextMessage
from autogen_ext.models.openai import OpenAIChatCompletionClient
class MyAssistant(RoutedAgent):
def __init__(self, name: str) -> None:
super().__init__(name)
model_client = OpenAIChatCompletionClient(
model="GLM-4-Air-0111",
api_key = "your api key",
base_url="https://open.bigmodel.cn/api/paas/v4/",
model_capabilities={
"vision": True,
"function_calling": True,
"json_output": True,
}
)
self._delegate = AssistantAgent(name, model_client=model_client)
@message_handler
async def handle_my_message_type(self, message: MyMessageType, ctx: MessageContext) -> None:
print(f"{self.id.type} received message: {message.content}")
response = await self._delegate.on_messages(
[TextMessage(content=message.content, source="user")], ctx.cancellation_token
)
print(f"{self.id.type} responded: {response.chat_message.content}")
python
from autogen_core import SingleThreadedAgentRuntime
runtime = SingleThreadedAgentRuntime()
await MyAgent.register(runtime, "my_agent", lambda: MyAgent())
await MyAssistant.register(runtime, "my_assistant", lambda: MyAssistant("my_assistant"))AgentType(type='my_assistant')
python
runtime.start() # Start processing messages in the background.
await runtime.send_message(MyMessageType("Hello, World!"), AgentId("my_agent", "default"))
await runtime.send_message(MyMessageType("Hello, World!"), AgentId("my_assistant", "default"))
await runtime.stop() # Stop processing messages in the background.my_agent received message: Hello, World!
my_assistant received message: Hello, World!
my_assistant responded: Hello! How can I help you today?2. 代码解释
这段代码展示了如何使用AutoGen框架创建和运行两个代理（Agent），其中每个代理可以接收和响应特定类型的消息。以下是代码逻辑的解释：
第一部分：定义消息类型和代理
python
from dataclasses import dataclass
from autogen_core import AgentId, MessageContext, RoutedAgent, message_handler
@dataclass
class MyMessageType:
content: str
class MyAgent(RoutedAgent):
def __init__(self) -> None:
super().__init__("MyAgent")
@message_handler
async def handle_my_message_type(self, message: MyMessageType, ctx: MessageContext) -> None:
print(f"{self.id.type} received message: {message.content}")- 定义消息类型 ：使用
@dataclass定义了一个数据类MyMessageType，用于表示消息，其中包含一个content字段。 - 定义代理类 ：创建了一个名为
MyAgent的代理类，继承自RoutedAgent。在初始化方法中，调用父类初始化并传入代理名称。 - 消息处理方法 ：使用
@message_handler装饰器定义了一个异步方法handle_my_message_type，用于处理MyMessageType类型的消息。当接收到消息时，打印消息内容。
第二部分：定义助手代理
python
from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.messages import TextMessage
from autogen_ext.models.openai import OpenAIChatCompletionClient
class MyAssistant(RoutedAgent):
def __init__(self, name: str) -> None:
super().__init__(name)
model_client = OpenAIChatCompletionClient(
model="GLM-4-Air-0111",
api_key = "your api key",
base_url="https://open.bigmodel.cn/api/paas/v4/",
model_capabilities={
"vision": True,
"function_calling": True,
"json_output": True,
}
)
self._delegate = AssistantAgent(name, model_client=model_client)
@message_handler
async def handle_my_message_type(self, message: MyMessageType, ctx: MessageContext) -> None:
print(f"{self.id.type} received message: {message.content}")
response = await self._delegate.on_messages(
[TextMessage(content=message.content, source="user")], ctx.cancellation_token
)
print(f"{self.id.type} responded: {response.chat_message.content}")- 定义助手代理类 ：创建了一个名为
MyAssistant的代理类，继承自RoutedAgent。 - 初始化助手代理 ：在初始化方法中，创建了一个
OpenAIChatCompletionClient实例，用于与OpenAI的聊天模型进行交互，并初始化了一个AssistantAgent作为代理的委托对象。 - 消息处理方法 ：定义了一个异步方法
handle_my_message_type，用于处理接收到的消息。当接收到消息时，打印消息内容，并通过委托对象_delegate将消息发送给OpenAI聊天模型，打印模型的响应。
第三部分：注册和运行代理
python
from autogen_core import SingleThreadedAgentRuntime
runtime = SingleThreadedAgentRuntime()
await MyAgent.register(runtime, "my_agent", lambda: MyAgent())
await MyAssistant.register(runtime, "my_assistant", lambda: MyAssistant("my_assistant"))- 创建运行时环境 ：创建了一个
SingleThreadedAgentRuntime实例，用于管理代理的运行。 - 注册代理 ：使用
register方法注册了MyAgent和MyAssistant代理。
第四部分：发送和停止消息处理
python
runtime.start() # Start processing messages in the background.
await runtime.send_message(MyMessageType("Hello, World!"), AgentId("my_agent", "default"))
await runtime.send_message(MyMessageType("Hello, World!"), AgentId("my_assistant", "default"))
await runtime.stop() # Stop processing messages in the background.- 启动消息处理 ：调用
runtime.start()启动后台消息处理。 - 发送消息 ：使用
runtime.send_message向两个代理发送MyMessageType类型的消息。 - 停止消息处理 ：调用
runtime.stop()停止后台消息处理。
总结
这段代码展示了如何使用AutoGen框架创建两个代理，其中MyAgent用于接收和打印消息，MyAssistant用于接收消息并通过OpenAI聊天模型生成响应。通过运行时环境SingleThreadedAgentRuntime，可以启动、发送消息和停止代理的消息处理。
3. 类似的例子
python
from dataclasses import dataclass
from autogen_core import AgentId, MessageContext, RoutedAgent, message_handler
# Define a new message type that contains a number and the operation to be performed
@dataclass
class MathOperationMessage:
number1: float
number2: float
operation: str # operation can be 'add', 'subtract', 'multiply', 'divide'
class MathAgent(RoutedAgent):
def __init__(self) -> None:
super().__init__("MathAgent")
@message_handler
async def handle_math_operation_message(self, message: MathOperationMessage, ctx: MessageContext) -> None:
if message.operation == "add":
result = message.number1 + message.number2
elif message.operation == "subtract":
result = message.number1 - message.number2
elif message.operation == "multiply":
result = message.number1 * message.number2
elif message.operation == "divide":
if message.number2 != 0:
result = message.number1 / message.number2
else:
result = "Error: Division by zero"
else:
result = "Error: Unknown operation"
print(f"{self.id.type} received message: {message}")
print(f"{self.id.type} performed operation: {message.operation} on {message.number1} and {message.number2}")
print(f"{self.id.type} result: {result}")
# Define another assistant agent that can respond with basic information
from autogen_agentchat.agents import AssistantAgent
from autogen_agentchat.messages import TextMessage
from autogen_ext.models.openai import OpenAIChatCompletionClient
class MathAssistant(RoutedAgent):
def __init__(self, name: str) -> None:
super().__init__(name)
model_client = OpenAIChatCompletionClient(
model="GLM-4-Air-0111",
api_key="your api key", # Replace with your actual API key
base_url="https://open.bigmodel.cn/api/paas/v4/",
model_capabilities={
"vision": True,
"function_calling": True,
"json_output": True,
}
)
self._delegate = AssistantAgent(name, model_client=model_client)
@message_handler
async def handle_math_message(self, message: MathOperationMessage, ctx: MessageContext) -> None:
print(f"{self.id.type} received message: {message}")
response = await self._delegate.on_messages(
[TextMessage(content=f"Performing operation: {message.operation} on {message.number1} and {message.number2}", source="user")],
ctx.cancellation_token
)
print(f"{self.id.type} responded: {response.chat_message.content}")
# Set up the runtime and agents
from autogen_core import SingleThreadedAgentRuntime
runtime = SingleThreadedAgentRuntime()
await MathAgent.register(runtime, "math_agent", lambda: MathAgent())
await MathAssistant.register(runtime, "math_assistant", lambda: MathAssistant("math_assistant"))
# Start processing messages
runtime.start()
# Send a message to the MathAgent for processing
await runtime.send_message(MathOperationMessage(3, 5, "add"), AgentId("math_agent", "default"))
await runtime.send_message(MathOperationMessage(10, 2, "subtract"), AgentId("math_agent", "default"))
# Send a message to the MathAssistant to provide context about the operation
await runtime.send_message(MathOperationMessage(7, 4, "multiply"), AgentId("math_assistant", "default"))
# Stop the runtime
await runtime.stop()math_agent received message: MathOperationMessage(number1=3, number2=5, operation='add')
math_agent performed operation: add on 3 and 5
math_agent result: 8
math_agent received message: MathOperationMessage(number1=10, number2=2, operation='subtract')
math_agent performed operation: subtract on 10 and 2
math_agent result: 8
math_assistant received message: MathOperationMessage(number1=7, number2=4, operation='multiply')
math_assistant responded: ```
7 * 4 = 28
```
TERMINATE