ai-agent-book 精选快照（<2MB 代码与文档，来自 github.com/bojieli/ai-agent-book）

chapter9/gaia-experience/AWorld/tests/gym_demo/README.md

@@ -0,0 +1,3 @@
+# gym demo
+
+Example of gym operation.

chapter9/gaia-experience/AWorld/tests/gym_demo/__init__.py

@@ -0,0 +1,2 @@
+# coding: utf-8
+# Copyright (c) 2025 inclusionAI.

chapter9/gaia-experience/AWorld/tests/gym_demo/agent.py

@@ -0,0 +1,34 @@
+# coding: utf-8
+# Copyright (c) 2025 inclusionAI.
+
+from typing import Any, Dict, Union, List
+
+from examples.common.tools.common import Tools
+from aworld.config.conf import AgentConfig, ConfigDict
+from aworld.agents.llm_agent import Agent
+from aworld.core.common import Observation, ActionModel
+
+
+class GymDemoAgent(Agent):
+    """Example agent"""
+
+    def __init__(self, conf: Union[Dict[str, Any], ConfigDict, AgentConfig], **kwargs):
+        super(GymDemoAgent, self).__init__(conf=conf, **kwargs)
+
+    def policy(self, observation: Observation, info: Dict[str, Any] = {}, **kwargs) -> Union[
+        List[ActionModel], None]:
+        import numpy as np
+
+        env_id = observation.info.get('env_id')
+        if env_id and env_id != 'CartPole-v1':
+            raise ValueError("Unsupported env")
+
+        res = np.random.randint(2)
+        action = [ActionModel(agent_name=self.id(), tool_name=Tools.GYM.value, action_name="play", params={"result": res})]
+        if observation.info.get("done"):
+            self._finished = True
+        return action
+
+    async def async_policy(self, observation: Observation, info: Dict[str, Any] = {}, **kwargs) -> Union[
+        List[ActionModel], None]:
+        return self.policy(observation, info, **kwargs)

chapter9/gaia-experience/AWorld/tests/gym_demo/run.py

@@ -0,0 +1,23 @@
+# coding: utf-8
+# Copyright (c) 2025 inclusionAI.
+import asyncio
+
+from aworld.config import RunConfig
+
+from aworld.config.conf import AgentConfig
+from aworld.core.task import Task
+from aworld.runner import Runners
+from tests.gym_demo.agent import GymDemoAgent as GymAgent
+
+
+async def main():
+    agent = GymAgent(name="gym_agent", conf=AgentConfig(), tool_names=["gym"], feedback_tool_result=True)
+    # It can also be used `ToolFactory` for simplification.
+    task = Task(agent=agent,
+                tools_conf={"gym": {"env_id": "CartPole-v1", "render_mode": "human", "render": True, "use_async": True}})
+    res = await Runners.run_task(task=task, run_conf=RunConfig())
+
+
+if __name__ == "__main__":
+    # We use it as a showcase to demonstrate the framework's scalability.
+    asyncio.run(main())