def test_and_plot_policy(policy,
env,
as_goals=True,
visualize=True,
sampling_res=1,
n_traj=1,
max_reward=1,
itr=0,
report=None,
center=None,
limit=None,
bounds=None):
avg_totRewards, avg_success, states, spacing, avg_time = test_policy(
policy,
env,
as_goals,
visualize,
sampling_res=sampling_res,
n_traj=n_traj,
bounds=bounds)
obj = env
while not hasattr(obj, '_maze_id') and hasattr(obj, 'wrapped_env'):
obj = env.wrapped_env
maze_id = obj._maze_id
plot_heatmap(avg_success,
states,
spacing=spacing,
show_heatmap=False,
maze_id=maze_id,
center=center,
limit=limit)
reward_img = save_image()
plot_heatmap(avg_time,
states,
spacing=spacing,
show_heatmap=False,
maze_id=maze_id,
center=center,
limit=limit,
adaptive_range=True)
time_img = save_image()
mean_rewards = np.mean(avg_totRewards)
success = np.mean(avg_success)
with logger.tabular_prefix('Outer_'):
logger.record_tabular('iter', itr)
logger.record_tabular('MeanRewards', mean_rewards)
logger.record_tabular('Success', success)
# logger.dump_tabular(with_prefix=False)
if report is not None:
report.add_image(
reward_img,
'policy performance\n itr: {} \nmean_rewards: {} \nsuccess: {}'.
format(itr, mean_rewards, success))
report.add_image(time_img, 'policy time\n itr: {} \n'.format(itr))
return mean_rewards, success
def plot_generator_samples(generator, env=None, size=100, fname=None):
goals, _ = generator.sample_states(size)
if env is None:
limit = np.max(np.abs(goals))
else:
goals_bound = env.goal_bounds[:len(env.goal_bounds) // 2]
limit = np.max(goals_bound)
goals_dim = goals.shape[1]
if goals_dim == 2:
plt.scatter(goals[:, 0], goals[:, 1], s=10)
plt.axis('equal')
plt.xlim(-limit, limit)
plt.ylim(-limit, limit)
else:
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
p = ax.scatter(goals[:, 0], goals[:, 1], goals[:, 2], s=10)
ax.axis('equal')
ax.set_xlim(-limit, limit)
ax.set_ylim(-limit, limit)
ax.set_zlim(-limit, limit)
img = save_image(fname=fname)
return img
def plot_policy_performance(policy,
env,
horizon,
n_samples=200,
n_traj=10,
fname=None):
goals_dim = env.dim
goals_bound = env.goal_bounds[:len(env.goal_bounds) // 2]
goals = np.random.uniform(-goals_bound, goals_bound,
[n_samples, goals_dim])
limit = np.max(goals_bound)
success_rates = evaluate_states(goals,
env,
policy,
horizon,
n_traj=n_traj,
key='goal_reached',
aggregator=(np.max, np.mean))
plt.clf()
if goals_dim == 2:
plt.scatter(goals[:, 0],
goals[:, 1],
c=success_rates,
s=10,
cmap='plasma')
plt.colorbar()
plt.axis('equal')
plt.xlim(-limit, limit)
plt.ylim(-limit, limit)
else:
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
p = ax.scatter(goals[:, 0],
goals[:, 1],
goals[:, 2],
c=success_rates,
s=10,
cmap='plasma')
fig.colorbar(p)
ax.axis('equal')
ax.set_xlim(-limit, limit)
ax.set_ylim(-limit, limit)
ax.set_zlim(-limit, limit)
img = save_image(fname=fname)
return img
def plot_policy_means(policy,
env,
sampling_res=2,
report=None,
center=None,
limit=None): # only for start envs!
states, empty_spaces, spacing = find_empty_spaces(
env, sampling_res=sampling_res)
goal = env.current_goal
observations = [
np.concatenate([
state, [
0,
] * (env.observation_space.flat_dim - len(state) - len(goal)), goal
]) for state in states
]
actions, agent_infos = policy.get_actions(observations)
vecs = agent_infos['mean']
vars = [np.exp(log_std) * 0.25 for log_std in agent_infos['log_std']]
ells = [
patches.Ellipse(state, width=vars[i][0], height=vars[i][1], angle=0)
for i, state in enumerate(states)
]
fig = plt.figure()
ax = fig.add_subplot(111)
for e in ells:
ax.add_artist(e)
e.set_alpha(0.2)
plt.scatter(*goal, color='r', s=100)
Q = plt.quiver(states[:, 0],
states[:, 1],
vecs[:, 0],
vecs[:, 1],
units='xy',
angles='xy',
scale_units='xy',
scale=1) # , np.linalg.norm(vars * 4)
qk = plt.quiverkey(Q,
0.8,
0.85,
1,
r'1 Nkg',
labelpos='E',
coordinates='figure')
# cb = plt.colorbar(Q)
vec_img = save_image()
if report is not None:
report.add_image(vec_img, 'policy mean')
def plot_regions_states(self, maze_id=0, report=None):
fig, ax = plt.subplots()
states_per_reg = [len(region.states) for region in self.regions]
states_per_reg_lims = (min(states_per_reg), max(states_per_reg))
normal = pylab.Normalize(*states_per_reg_lims)
colors = pylab.cm.BuGn(normal(states_per_reg))
for region, color in zip(self.regions, colors):
lengths = region.max_border - region.min_border
ax.add_patch(
patches.Rectangle(region.min_border,
*lengths,
fill=True,
edgecolor='k',
facecolor=color))
cax, _ = cbar.make_axes(ax)
print("the interest lims are: ", states_per_reg_lims)
cb2 = cbar.ColorbarBase(cax, cmap=pylab.cm.BuGn, norm=normal)
ax.set_xlim(self.state_bounds[0][0], self.state_bounds[1][0])
ax.set_ylim(self.state_bounds[0][1], self.state_bounds[1][1])
if maze_id == 0:
ax.add_patch(
patches.Rectangle((-3, -3),
10,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-3, -1),
2,
6,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-3, 5),
10,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((5, -1),
2,
6,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-1, 1),
4,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
elif maze_id == 11:
ax.add_patch(
patches.Rectangle((-7, 5),
14,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((5, -5),
2,
10,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-7, -7),
14,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-7, -5),
2,
10,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-1, 1),
6,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-3, -3),
2,
6,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-1, -3),
4,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
regions_fig = save_image(fig)
if report is None:
return regions_fig
else:
report.add_image(
regions_fig,
'States per region\nTotal number of states: {}'.format(
sum([len(region.states) for region in self.regions])))
def plot_regions_interest(self, maze_id=0, report=None):
fig, ax = plt.subplots()
interests = self.compute_all_interests()
interest_lims = (min(interests), max(interests))
normal = pylab.Normalize(*interest_lims)
colors = pylab.cm.YlOrRd(normal(interests))
for region, color in zip(self.regions, colors):
lengths = region.max_border - region.min_border
ax.add_patch(
patches.Rectangle(region.min_border,
*lengths,
fill=True,
edgecolor='k',
facecolor=color))
cax, _ = cbar.make_axes(ax)
print("the interest lims are: ", interest_lims)
cb2 = cbar.ColorbarBase(cax, cmap=pylab.cm.YlOrRd, norm=normal)
ax.set_xlim(self.state_bounds[0][0], self.state_bounds[1][0])
ax.set_ylim(self.state_bounds[0][1], self.state_bounds[1][1])
if maze_id == 0:
ax.add_patch(
patches.Rectangle((-3, -3),
10,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-3, -1),
2,
6,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-3, 5),
10,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((5, -1),
2,
6,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-1, 1),
4,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
elif maze_id == 11:
ax.add_patch(
patches.Rectangle((-7, 5),
14,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((5, -5),
2,
10,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-7, -7),
14,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-7, -5),
2,
10,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-1, 1),
6,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-3, -3),
2,
6,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
ax.add_patch(
patches.Rectangle((-1, -3),
4,
2,
fill=True,
edgecolor="none",
facecolor='0.4',
alpha=0.3))
regions_fig = save_image(fig)
if report is None:
return regions_fig
else:
report.add_image(
regions_fig,
'Interest per region:\nthe number of regions is: {}'.format(
len(self.regions)))