from snake_game import Snake
import time
import random
snake = Snake(20, True)
debug = False
best_score = 0
while 1:
time.sleep(0.03)
if(debug):
move = snake.do_move(0)
else:
move = snake.do_move(random.randint(0,4))
print(move)
if(move['reward'] < 0):
snake.reset()
elif(snake.get_score() > best_score):
best_score = snake.get_score()
bestScore = score
print('New Best Score : ' + str(bestScore) + ' a ' +
str(datetime.datetime.now()))
agent.save()
if (reward < -1):
#print('Total reward : ' + str(lastTotalReward))
lastTotalReward = 0
snake.reset()
iteration += 1
observation = snake.get_observation()
state_control = observation['state_control']
snake_location = observation['snake_location']
state = get_state(grid_size, snake_location, state_control)
action = agent.pickAction(state, True)
new_observation = snake.do_move(agent.getSelectedAction(action))
reward = new_observation['reward']
lastTotalReward += reward
new_state_control = new_observation['state_control']
new_snake_location = new_observation['snake_location']
next_state = get_state(grid_size, new_snake_location, new_state_control)
agent.train(state, action, reward, next_state, True)
print('Best Score : ' + str(bestScore))
