def test_env_save_state(self):
env = SLGEnv(ExampleState(), LevelCondition(1), ExampleReward(), TimeHandler())
env.setup('test_client')
env.save_state()
import json
with open(get_data_path() + '/' + 'state.json', 'r+') as f:
usr_data = json.load(f)
self.assertEqual(ExampleState().get_state(usr_data), [1])
def main():
l = init_logger('ExampleAgent')
sh = ExampleState()
th = LevelCondition(2)
time_h = TimeHandler()
rh = ExampleReward()
env = SLGEnv(sh, th, rh, time_h)
env.setup(generate_user_name())
agent = ExampleAgent(env)
agent.run()
def main():
l = init_logger('MLPQAgent')
sess = tf.Session()
sh = ResourceOnly()
th = LevelCondition(3)
time_h = CountActionTime(60 * 60 * 24 * 5)
rh = StepTimeRelated(time_h)
env = SLGEnv(sh, th, rh, time_h)
agent = FAQAgent(env, 'MLPQAgent', MLP, sess)
agent.train(1000000, 20, 10, 10000)
def main():
l = init_logger('QUKAgent')
goal_time = 468000
sh = ResourceOnly()
th = LevelCondition(3)
time_h = CountActionTime(60 * 60 * 24 * 6)
rh = TotalTimeRelated(time_h, goal_time)
env = SLGEnv(sh, th, rh, time_h)
fa = MLP((len(sh) + 1, 20, 20, 1))
agent = FAQAgent(env, 'QUKAgent', fa)
agent.train(10000, 20, 1000)
def test_env_get_state(self):
env = SLGEnv(ExampleState(), LevelCondition(1), ExampleReward(), TimeHandler())
env.setup('test_client')
self.assertEqual(env.get_state(), [1])
def test_env_setup(self):
env = SLGEnv(ExampleState(), LevelCondition(1), ExampleReward(), TimeHandler())
self.assertTrue(env.setup('test_client'))
def test_env_init(self):
env = SLGEnv(ExampleState(), LevelCondition(1), ExampleReward(), TimeHandler())
self.assertTrue(True)