def test_nogui_agent(agent_pos=None, coin_pos=None):
'''
Testing of the agent training with no gui of the game board
:return:
'''
world = get_fake_world(
) if not agent_pos and not coin_pos else get_fake_world(
agent_pos, coin_pos)
qagent.setup(world)
plt.ioff()
for episode in range(settings['n_rounds']):
print('EPISODE: ', episode)
for step in range(100):
try:
print('STEP: ', step)
qagent.act(world)
next_action = world.next_action
new_world = fake_environment_step(world, next_action)
qagent.reward_update(new_world)
world = new_world
except:
print(traceback.format_exc())
print('ERRROROR error happened!!!!!!')
qagent.end_of_episode(world)
world = fake_environment_reset(
world
) if not agent_pos and not coin_pos else fake_environment_reset(
world, agent_pos=agent_pos, coin_pos=coin_pos)
def test_gui_agent(agent_pos=None, coin_pos=None):
'''
Testing of the agente training with gui of the game board
:return:
'''
plt.ion()
world = get_fake_world(
) if not agent_pos and not coin_pos else get_fake_world(
agent_pos, coin_pos)
qagent.setup(world)
screen = RealWorld(world)
for episode in range(10):
print('EPISODE: ', episode)
for step in range(100):
print('STEP: ', step)
qagent.act(world)
next_action = world.next_action
new_world = fake_environment_step(world, next_action)
screen.animate(new_world)
time.sleep(0.01)
qagent.reward_update(new_world)
world = new_world
qagent.end_of_episode(world)
world = fake_environment_reset(
world
) if not agent_pos and not coin_pos else fake_environment_reset(
world, agent_pos=agent_pos, coin_pos=coin_pos)
def test_coin_zone_gen():
'''
Test of the coin zone generation and the states from it.
:return:
'''
world =get_fake_world()
qagent.setup(world)
qagent.act(world)
qagent.reward_update(world)
agent_world = world.game_wrapper.agent.game_abstraction
print('agenst world arena')
print(agent_world.map)
print('position of the robot: ')
print(world.game_state['self'])
def test_nogui_agent():
'''
Testing of the agent training with no gui of the game board
:return:
'''
world = get_fake_world()
qagent.setup(world)
for episode in range(settings['n_rounds']):
print('EPISODE: ', episode)
for step in range(100):
print('STEP: ', step)
qagent.act(world)
next_action = world.next_action
new_world = fake_environment_step(world,next_action)
qagent.reward_update(new_world)
world = new_world
qagent.end_of_episode(world)
world = fake_environment_reset(world)