def _check_for_episode_termination(self, reset_rules_status, agents_info_map):
'''Check for whether a episode should be terminated
Args:
reset_rules_status: dictionary of reset rules status with key as reset rule names and value as
reset rule bool status
agents_info_map: dictionary of agents info map with key as agent name and value as agent info
Returns:
tuple (string, bool, bool): episode status, pause flag, and done flag
'''
episode_status = EpisodeStatus.get_episode_status(reset_rules_status)
pause = False
done = False
# Note: check EPISODE_COMPLETE as the first item because agent might crash
# at the finish line.
if EpisodeStatus.EPISODE_COMPLETE.value in reset_rules_status and \
reset_rules_status[EpisodeStatus.EPISODE_COMPLETE.value]:
done = True
episode_status = EpisodeStatus.EPISODE_COMPLETE.value
elif EpisodeStatus.CRASHED.value in reset_rules_status and \
reset_rules_status[EpisodeStatus.CRASHED.value]:
# check crash with all other objects besides static obstacle
crashed_object_name = agents_info_map[self._agent_name_][AgentInfo.CRASHED_OBJECT_NAME.value]
if 'obstacle' not in crashed_object_name:
current_progress = agents_info_map[self._agent_name_][AgentInfo.CURRENT_PROGRESS.value]
crashed_object_progress = agents_info_map[crashed_object_name]\
[AgentInfo.CURRENT_PROGRESS.value]
if current_progress < crashed_object_progress:
done, pause = self._check_for_phase_change()
else:
done, pause = self._check_for_phase_change()
elif any(reset_rules_status.values()):
done, pause = self._check_for_phase_change()
return episode_status, pause, done
def append_episode_metrics(self):
self._number_of_trials_ += 1
eval_metric = dict()
eval_metric['completion_percentage'] = int(self._progress_)
eval_metric['metric_time'] = int(round(time.time() * 1000))
eval_metric['start_time'] = int(round(self._start_time_ * 1000))
eval_metric['elapsed_time_in_milliseconds'] = \
int(round((time.time() - self._start_time_) * 1000))
eval_metric['trial'] = int(self._number_of_trials_)
eval_metric['episode_status'] = EpisodeStatus.get_episode_status_label(
self._episode_status)
eval_metric['crash_count'] = self.reset_count_dict[
EpisodeStatus.CRASHED.value]
eval_metric['immobilized_count'] = self.reset_count_dict[
EpisodeStatus.IMMOBILIZED.value]
eval_metric['off_track_count'] = self.reset_count_dict[
EpisodeStatus.OFF_TRACK.value]
eval_metric['reversed_count'] = self.reset_count_dict[
EpisodeStatus.REVERSED.value]
eval_metric['reset_count'] = eval_metric['crash_count'] + \
eval_metric['immobilized_count'] + \
eval_metric['off_track_count'] + \
eval_metric['reversed_count']
self._best_lap_time = min(eval_metric['elapsed_time_in_milliseconds'],
self._best_lap_time)
self._total_evaluation_time += eval_metric[
'elapsed_time_in_milliseconds']
self._metrics_.append(eval_metric)
def _check_for_episode_termination(self, reset_rules_status,
agents_info_map):
'''Check for whether a episode should be terminated
Args:
reset_rules_status: dictionary of reset rules status with key as reset rule names and value as
reset rule bool status
agents_info_map: dictionary of agents info map with key as agent name and value as agent info
Returns:
tuple (string, bool, bool): episode status, pause flag, and done flag
'''
episode_status = EpisodeStatus.get_episode_status(reset_rules_status)
pause = False
done = False
# Note: check EPISODE_COMPLETE as the first item because agent might crash
# at the finish line.
if EpisodeStatus.EPISODE_COMPLETE.value in reset_rules_status and \
reset_rules_status[EpisodeStatus.EPISODE_COMPLETE.value]:
done = True
episode_status = EpisodeStatus.EPISODE_COMPLETE.value
elif EpisodeStatus.CRASHED.value in reset_rules_status and \
reset_rules_status[EpisodeStatus.CRASHED.value]:
# only check for crash when at RUN phase
if self._ctrl_status[
AgentCtrlStatus.AGENT_PHASE.value] == AgentPhase.RUN.value:
self._curr_crashed_object_name = agents_info_map[
self._agent_name_][AgentInfo.CRASHED_OBJECT_NAME.value]
# check crash with all other objects besides static obstacle
if 'obstacle' not in self._curr_crashed_object_name:
current_progress = agents_info_map[self._agent_name_][
AgentInfo.CURRENT_PROGRESS.value]
crashed_obj_info = agents_info_map[
self._curr_crashed_object_name]
crashed_obj_progress = crashed_obj_info[
AgentInfo.CURRENT_PROGRESS.value]
crashed_obj_start_ndist = crashed_obj_info[
AgentInfo.START_NDIST.value]
crashed_object_progress = get_normalized_progress(
crashed_obj_progress,
start_ndist=crashed_obj_start_ndist)
current_progress = get_normalized_progress(
current_progress,
start_ndist=self._data_dict_['start_ndist'])
if current_progress < crashed_object_progress:
done, pause = self._check_for_phase_change()
else:
episode_status = EpisodeStatus.IN_PROGRESS.value
else:
done, pause = self._check_for_phase_change()
else:
pause = True
elif any(reset_rules_status.values()):
done, pause = self._check_for_phase_change()
return episode_status, pause, done
def append_episode_metrics(self):
self._episode_ += 1 if not self._is_eval_ else 0
self._eval_trials_ += 1 if not self._is_eval_ else 0
training_metric = dict()
training_metric['reward_score'] = int(round(self._episode_reward_))
training_metric['metric_time'] = int(round(self._current_sim_time * 1000))
training_metric['start_time'] = int(round(self._start_time_ * 1000))
training_metric['elapsed_time_in_milliseconds'] = \
int(round((self._current_sim_time - self._start_time_) * 1000))
training_metric['episode'] = int(self._episode_)
training_metric['trial'] = int(self._eval_trials_)
training_metric['phase'] = 'evaluation' if self._is_eval_ else 'training'
training_metric['completion_percentage'] = int(self._progress_)
training_metric['episode_status'] = EpisodeStatus.get_episode_status_label(self._episode_status)
self._metrics_.append(training_metric)
def upload_episode_metrics(self):
self._number_of_trials_ += 1
eval_metric = dict()
eval_metric['completion_percentage'] = int(self._progress_)
eval_metric['metric_time'] = int(round(time.time() * 1000))
eval_metric['start_time'] = int(round(self._start_time_ * 1000))
eval_metric['elapsed_time_in_milliseconds'] = \
int(round((time.time() - self._start_time_) * 1000))
eval_metric['trial'] = int(self._number_of_trials_)
eval_metric['episode_status'] = EpisodeStatus.get_episode_status_label(
self._episode_status)
self._metrics_.append(eval_metric)
write_metrics_to_s3(self._s3_dict_[MetricsS3Keys.METRICS_BUCKET.value],
self._s3_dict_[MetricsS3Keys.METRICS_KEY.value],
self._s3_dict_[MetricsS3Keys.REGION.value],
self._metrics_)
self._simtrace_data_.upload_to_s3(self._number_of_trials_)
def append_episode_metrics(self):
self._episode_ += 1 if not self._is_eval_ else 0
self._eval_trials_ += 1 if not self._is_eval_ else 0
training_metric = dict()
training_metric["reward_score"] = int(round(self._episode_reward_))
training_metric["metric_time"] = int(
round(self._current_sim_time * 1000))
training_metric["start_time"] = int(round(self._start_time_ * 1000))
training_metric["elapsed_time_in_milliseconds"] = int(
round((self._current_sim_time - self._start_time_) * 1000))
training_metric["episode"] = int(self._episode_)
training_metric["trial"] = int(self._eval_trials_)
training_metric[
"phase"] = "evaluation" if self._is_eval_ else "training"
training_metric["completion_percentage"] = int(self._progress_)
training_metric[
"episode_status"] = EpisodeStatus.get_episode_status_label(
self._episode_status)
self._metrics_.append(training_metric)
def append_episode_metrics(self, is_complete=True):
if not is_complete and self._number_of_trials_ != 0:
# Note: for virtual event, if the racer did not even finish one lap
# for the duration of the event, we display DNF.
# However, our friends at the game want the DNF ranks as well
# so we append the incomplete metrics for ppl who didn't finish
# first lap
LOGGER.info(
"Appending episode metrics for incomplete lap skipped, laps completed %s",
self._number_of_trials_,
)
return
eval_metric = dict()
eval_metric["completion_percentage"] = int(self._progress_)
eval_metric["metric_time"] = int(round(self._current_sim_time * 1000))
eval_metric["start_time"] = int(round(self._start_time_ * 1000))
eval_metric["elapsed_time_in_milliseconds"] = int(
round((self._current_sim_time - self._start_time_) * 1000))
eval_metric["episode_status"] = EpisodeStatus.get_episode_status_label(
self._episode_status)
eval_metric["crash_count"] = self.reset_count_dict[
EpisodeStatus.CRASHED.value]
eval_metric["immobilized_count"] = self.reset_count_dict[
EpisodeStatus.IMMOBILIZED.value]
eval_metric["off_track_count"] = self.reset_count_dict[
EpisodeStatus.OFF_TRACK.value]
eval_metric["reversed_count"] = self.reset_count_dict[
EpisodeStatus.REVERSED.value]
eval_metric["reset_count"] = (eval_metric["crash_count"] +
eval_metric["immobilized_count"] +
eval_metric["off_track_count"] +
eval_metric["reversed_count"])
if is_complete:
self._number_of_trials_ += 1
self._best_lap_time = min(
eval_metric["elapsed_time_in_milliseconds"],
self._best_lap_time)
self._total_evaluation_time += eval_metric[
"elapsed_time_in_milliseconds"]
eval_metric["trial"] = int(self._number_of_trials_)
self._metrics_.append(eval_metric)
def judge_action(self, action):
try:
# Get the position of the agent
pos_dict = self._track_data_.get_agent_pos(
self._agent_name_, self._agent_link_name_list_,
const.RELATIVE_POSITION_OF_FRONT_OF_CAR)
model_point = pos_dict[AgentPos.POINT.value]
# Compute the next index
current_ndist = self._track_data_.get_norm_dist(model_point)
prev_index, next_index = self._track_data_.find_prev_next_waypoints(
current_ndist, normalized=True, reverse_dir=self._reverse_dir_)
# Set the basic reward and training metrics
set_reward_and_metrics(self._reward_params_, self._step_metrics_,
pos_dict, self._track_data_, next_index,
prev_index, action, self._json_actions_)
# Convert current progress to be [0,100] starting at the initial waypoint
if self._reverse_dir_:
self._reward_params_[const.RewardParam.LEFT_CENT.value[0]] = \
not self._reward_params_[const.RewardParam.LEFT_CENT.value[0]]
current_progress = self._data_dict_[
'start_ndist'] - current_ndist
else:
current_progress = current_ndist - self._data_dict_[
'start_ndist']
current_progress = compute_current_prog(
current_progress, self._data_dict_['prev_progress'])
self._data_dict_['steps'] += 1
# Add the agen specific metrics
self._step_metrics_[StepMetrics.STEPS.value] = \
self._reward_params_[const.RewardParam.STEPS.value[0]] = self._data_dict_['steps']
self._reward_params_[
const.RewardParam.REVERSE.value[0]] = self._reverse_dir_
self._step_metrics_[StepMetrics.PROG.value] = \
self._reward_params_[const.RewardParam.PROG.value[0]] = current_progress
except Exception as ex:
raise GenericRolloutException(
'Cannot find position: {}'.format(ex))
# This code should be replaced with the contact sensor code
is_crashed = False
model_heading = self._reward_params_[
const.RewardParam.HEADING.value[0]]
obstacle_reward_params = self._track_data_.get_object_reward_params(
model_point, model_heading, current_progress, self._reverse_dir_)
if obstacle_reward_params:
self._reward_params_.update(obstacle_reward_params)
is_crashed = self._track_data_.is_racecar_collided(
pos_dict[AgentPos.LINK_POINTS.value])
prev_pnt_dist = min(
model_point.distance(self._prev_waypoints_['prev_point']),
model_point.distance(self._prev_waypoints_['prev_point_2']))
is_off_track = not any(
self._track_data_.points_on_track(
pos_dict[AgentPos.LINK_POINTS.value]))
is_immobilized = (prev_pnt_dist <= 0.0001
and self._data_dict_['steps'] %
const.NUM_STEPS_TO_CHECK_STUCK == 0)
is_lap_complete = current_progress >= 100.0
self._reward_params_[const.RewardParam.CRASHED.value[0]] = is_crashed
self._reward_params_[
const.RewardParam.OFFTRACK.value[0]] = is_off_track
done = is_crashed or is_immobilized or is_off_track or is_lap_complete
episode_status = EpisodeStatus.get_episode_status(
is_crashed=is_crashed,
is_immobilized=is_immobilized,
is_off_track=is_off_track,
is_lap_complete=is_lap_complete)
try:
reward = float(self._reward_(copy.deepcopy(self._reward_params_)))
except Exception as ex:
raise RewardFunctionError(
'Reward function exception {}'.format(ex))
if math.isnan(reward) or math.isinf(reward):
raise RewardFunctionError('{} returned as reward'.format(reward))
self._prev_waypoints_['prev_point_2'] = self._prev_waypoints_[
'prev_point']
self._prev_waypoints_['prev_point'] = model_point
self._data_dict_['prev_progress'] = current_progress
#Get the last of the step metrics
self._step_metrics_[StepMetrics.REWARD.value] = reward
self._step_metrics_[StepMetrics.DONE.value] = done
self._step_metrics_[StepMetrics.TIME.value] = time.time()
self._step_metrics_[
StepMetrics.EPISODE_STATUS.value] = episode_status.value
return reward, done, self._step_metrics_
