def __str__(self):
"""
Used for displaying the grid on the command line
"""
grid_rep = []
grid_rep.append('%ix%i World' % (self.h, self.w))
grid_rep.append('\n\n\n')
for i, row in enumerate(self.grid):
for j, elem in enumerate(row):
if elem == Simulator.__wall_sym:
grid_rep.append(Color.red(' %s ' % elem))
else:
if (i,j) in self.absorb:
if self.absorb[(i, j)] > 0:
grid_rep.append(Color.green(' %s ' % Simulator.__goal_sym))
elif self.absorb[(i, j)] < 0:
grid_rep.append(Color.red(' %s ' % Simulator.__goal_sym))
else:
grid_rep.append(' %s ' % Simulator.__goal_sym)
elif (i, j) == self.state:
grid_rep.append(Color.yellow(' %s ' % Simulator.__agent_sym))
else:
grid_rep.append(' %s ' % elem)
grid_rep.append('\n')
grid_rep.append('\n\n\n')
return ''.join(grid_rep)
def __str__(self):
grid_rep = ''
for i,row in enumerate(self.grid):
for j,elem in enumerate(row):
if(elem == RoomRunner.__wall_sym):
grid_rep += Color.red(' %s ' % elem)
elif(elem == RoomRunner.__agent_sym):
grid_rep += Color.yellow(' %s ' % elem)
else:
if((i,j) in self.goals):
grid_rep += Color.green(' %s ' % elem)
else:
grid_rep += ' %s ' % elem
grid_rep += '\n'
return grid_rep
def watch_execution(sim, policy, state=None, maxsteps=500):
"""
Visually displays the grid and the agents movements in the terminal
as the agent executes the given policy
"""
absorb = False
sim.reset(state)
totreward = 0.0
steps = 0
while not absorb and steps < maxsteps:
steps += 1
os.system('cls' if os.name=='nt' else 'clear')
print sim
stateindex = sim.state_to_index(sim.state)
action = policy.select_action(stateindex)[0]
print "Step: %i State: %s" % (steps, sim.state)
print "------------"
print "Executing action %s (%i)" % (print_action(action), action)
print "Total Reward: %f" % totreward
print "Reward per Step: %f" % (totreward/steps)
sample = sim.execute(action)
totreward += sample.reward
absorb = sample.absorb
sleep(.4)
print
if absorb:
print Color.green("Agent Reached Goal State %s with reward of %f" % (sim.state, sample.reward))
success = True
else:
print Color.red("Agent did not reach a goal state within %i steps!" % maxsteps)
success = False
print "Total Reward: %f" % totreward
print "Reward per Step: %f" % (totreward/steps)
return success, totreward, steps, (totreward/steps)
import sys
from terminal import Color
def _pad(i):
i = str(i)
return i + ' ' * (5 - len(i))
print('System colors')
for i in range(16):
c = Color(' ')
c.bgcolor = i
if i == 8:
print('')
sys.stdout.write(str(c))
print('\n\nBackground')
for green in range(6):
for red in range(6):
for blue in range(6):
c = Color(' ')
c.bgcolor = 16 + red * 36 + green * 6 + blue
sys.stdout.write(str(c))
print('')
print('\nGrayscale')
for i in range(232, 256):
c = Color(' ')
c.bgcolor = i
sys.stdout.write(str(c))
import sys
from terminal import Color
def _pad(i):
i = str(i)
return i + ' ' * (5 - len(i))
print('System colors')
for i in range(16):
c = Color(' ')
c.bgcolor = i
if i == 8:
print('')
sys.stdout.write(str(c))
print('\n\nBackground')
for green in range(6):
for red in range(6):
for blue in range(6):
c = Color(' ')
c.bgcolor = 16 + red * 36 + green * 6 + blue
sys.stdout.write(str(c))
print('')
print('\nGrayscale')
for i in range(232, 256):
c = Color(' ')
c.bgcolor = i
sys.stdout.write(str(c))