Create ReAct Agent
AgentScope provides out-of-the-box ReAct agent ReActAgent under agentscope.agent that can be used directly.
It supports the following features at the same time:
✨ Basic features
Support hooks around
reply,observe,print,_reasoningand_actingfunctionsSupport structured output
✋ Realtime Steering
Support user interrupt
Support customized interruption handling
🛠️ Tools
Support both sync/async tool functions
Support streaming tool response
Support stateful tools management
Support parallel tool calls
Support MCP server
💾 Memory
Support agent-controlled long-term memory management
Support static long-term memory management
Tip
Refer to the Agent section for more details about these features. In quickstart, we focus on how to create a ReAct agent and run it.
from agentscope.agent import ReActAgent, AgentBasefrom agentscope.formatter import DashScopeChatFormatterfrom agentscope.memory import InMemoryMemoryfrom agentscope.message import Msgfrom agentscope.model import DashScopeChatModelimport asyncioimport osfrom agentscope.tool import Toolkit, execute_python_code
Creating ReAct Agent
To improve the flexibility, the ReActAgent class exposes the following parameters in its constructor:
Parameter | Further Reading | Description |
|---|---|---|
| The name of the agent | |
| The system prompt of the agent | |
| The model used by the agent to generate responses | |
| The prompt construction strategy, should be consisted with the model | |
| The toolkit to register/call tool functions. | |
| The short-term memory used to store the conversation history | |
| The long-term memory | |
| The mode of the long-term memory:
| |
| Whether to enable the meta tool, which allows the agent to manage tools by itself | |
| Whether to allow parallel tool calls | |
| The maximum number of iterations for the agent to generate a response | |
| The plan notebook to manage the plans | |
| Whether to print the hint message generated by the plan notebook at each step |
Taking DashScope API as example, we create an agent object as follows:
async def creating_react_agent() -> None:"""Create a ReAct agent and run a simple task."""# Prepare toolstoolkit = Toolkit()toolkit.register_tool_function(execute_python_code)jarvis = ReActAgent(name="Jarvis",sys_prompt="You're a helpful assistant named Jarvis",model=DashScopeChatModel(model_name="qwen-max",api_key=os.environ["DASHSCOPE_API_KEY"],stream=True,enable_thinking=False,),formatter=DashScopeChatFormatter(),toolkit=toolkit,memory=InMemoryMemory(),)msg = Msg(name="user",content="Hi! Jarvis, run Hello World in Python.",role="user",)await jarvis(msg)asyncio.run(creating_react_agent())
Jarvis: {"type": "tool_use","id": "call_ce99a90230ee4aee8baae5","name": "execute_python_code","input": {"code": "print('Hello World')","timeout": 5}}system: {"type": "tool_result","id": "call_ce99a90230ee4aee8baae5","name": "execute_python_code","output": [{"type": "text","text": "<returncode>0</returncode><stdout>Hello World\n</stdout><stderr></stderr>"}]}Jarvis: The Python code has been executed successfully and it printed:```Hello World```
Creating From Scratch
You may want to create an agent from scratch, AgentScope provides two base classes for you to inherit from:
Class | Abstract Methods | Description |
|---|---|---|
| replyobservehandleinterrupt |
|
| replyobservehandle_interrupt_reasoning_acting | Add two abstract functions |
Please refer to the Agent section for more details about the agent class.
Taking the AgentBase class as an example, we can create a custom agent class by inheriting from it and implementing the reply method.
class MyAgent(AgentBase):"""A custom agent class"""def __init__(self) -> None:"""Initialize the agent"""super().__init__()self.name = "Friday"self.sys_prompt = "You're a helpful assistant named Friday."self.model = DashScopeChatModel(model_name="qwen-max",api_key=os.environ["DASHSCOPE_API_KEY"],stream=False,)self.formatter = DashScopeChatFormatter()self.memory = InMemoryMemory()async def reply(self, msg: Msg | list[Msg] | None) -> Msg:"""Reply to the message."""await self.memory.add(msg)# Prepare the promptprompt = await self.formatter.format([Msg("system", self.sys_prompt, "system"),*await self.memory.get_memory(),],)# Call the modelresponse = await self.model(prompt)msg = Msg(name=self.name,content=response.content,role="assistant",)# Record the response in memoryawait self.memory.add(msg)# Print the messageawait self.print(msg)return msgasync def observe(self, msg: Msg | list[Msg] | None) -> None:"""Observe the message."""# Store the message in memoryawait self.memory.add(msg)async def handle_interrupt(self) -> Msg:"""Postprocess the interrupt."""# Taking a fixed response as examplereturn Msg(name=self.name,content="I noticed you interrupted me, how can I help you?",role="assistant",)async def run_custom_agent() -> None:"""Run the custom agent."""agent = MyAgent()msg = Msg(name="user",content="Who are you?",role="user",)await agent(msg)asyncio.run(run_custom_agent())
Friday: I'm Friday, an AI assistant designed to help with information, tasks, and a wide range of other things. How can I assist you today?
Further Reading
Total running time of the script: (0 minutes 8.263 seconds)
Download Jupyter notebook: quickstart_agent.ipynb
Download Python source code: quickstart_agent.py