我有一个从Ray框架调用PolicyClient的函数,然后调用 start_episode。现在我想为这个函数创建单元测试:
def start_episode_historical(obs, reward, action):
client = PolicyClient(
os.environ['url_ray_online_training'],
inference_mode="local")
episode_id = client.start_episode(training_enabled=True)
client.log_action(episode_id, obs, action)
client.log_returns(episode_id, reward)
client.end_episode(episode_id, obs)
client.update_policy_weights()
任何人都可以帮忙吗?