def do_step(self, state):
""" Do an agent step """
state = list(state)
# Get the matchbox for this state
matchbox = self.get_matchbox(state)
# Play
marble, new_state = self.play(matchbox)
# Store this move for learning
self.moves.append((marble, matchbox))
# Some debugging output
print_state([state, new_state])
print
# Return new state to environment
action = Action()
action.intArray = new_state
return action
def do_step(self, state):
""" Do an agent step """
state = list(state)
# Get the matchbox for this state
matchbox = self.get_matchbox(state)
# Play
marble, new_state = self.play(matchbox)
# Store this move for learning
self.moves.append((marble, matchbox))
# Some debugging output
print_state([state, new_state])
print
# Return new state to environment
action = Action()
action.intArray = new_state
return action
def print_moves(self):
""" Print the agent's moves so far """
marbles = [("(%d)" % (m[0])).center(5) for m in self.moves]
print " ".join(marbles)
print_state([m[1].state for m in self.moves])
def print_moves(self):
""" Print the agent's moves so far """
marbles = [("(%d)" % (m[0])).center(5) for m in self.moves]
print " ".join(marbles)
print_state([m[1].state for m in self.moves])