def __init__(self):
rospy.init_node('constant_feature_publisher')
stream_name = rospy.get_param('~stream_name')
description = rospy.get_param('~description')
self.features = rospy.get_param('~features')
self.mode = rospy.get_param('~mode')
opt_args = {}
if self.mode == 'push':
opt_args['queue_size'] = rospy.get_param('~queue_size', 0)
self.publish_rate = rospy.Rate(rospy.get_param('~publish_rate'))
elif self.mode == 'pull':
opt_args['cache_time'] = float('inf')
else:
raise ValueError('Mode must be pull or push')
rate = rospy.get_param('~publish_rate', 1.0)
self.publish_rate = rospy.Rate(rate)
self.feat_tx = broadcast.Transmitter(stream_name=stream_name,
namespace='~',
feature_size=len(self.features),
description=description,
mode=self.mode,
**opt_args)
def __init__(self):
self.mode = rospy.get_param('dim_mode')
if self.mode == '2d':
dim = 3
elif self.mode == '3d':
dim = 6
else:
raise ValueError('Unknown mode %s' % self.mode)
topic = rospy.get_param('~stream_name')
self.tx = broadcast.Transmitter(stream_name=topic,
feature_size=dim,
description='Velocity components',
mode='push')
self.velsub = rospy.Subscriber('twist', TwistStamped,
self.twist_callback)
def __init__(self):
self.dim = rospy.get_param('~dim')
self.params = []
self._critique_server = rospy.Service('~get_critique', GetCritique,
self.critique_callback)
self.err_cutoff = rospy.get_param('~err_cutoff')
self.action_bias = np.random.uniform(-0.5, 0.5, self.dim)
rospy.loginfo('Action bias: %s', np.array_str(self.action_bias))
# Set up state generator and transmitter
self.state = None
self.state_tx = broadcast.Transmitter(
stream_name='next_bandit_state',
feature_size=self.dim,
description='Test bandit problem context/state',
mode='pull',
queue_size=10)
self.__update_state()
def __init__(self):
self.scale = rospy.get_param('~scale')
self.lap_k = rospy.get_param('~laplacian_k', 1)
self.bridge = CvBridge()
self.num_pyrs = rospy.get_param('~pyramid_levels', 0)
stream_name = rospy.get_param('~stream_name')
self.feat_tx = broadcast.Transmitter(stream_name=stream_name,
namespace='~',
feature_size=self.num_pyrs + 1,
description=['laplacian_std'],
mode='push',
queue_size=0)
buff_size = rospy.get_param('~buffer_size', 1)
self.image_sub = rospy.Subscriber('image', Image,
self.ImageCallback,
queue_size=buff_size)
def __init__(self):
dt = rospy.get_param('~cart_dt')
max_acc = rospy.get_param('~cart_max_acc')
max_steps = rospy.get_param('~cart_max_steps')
self.cart = CartSystem(dt=dt, max_acc=max_acc, max_steps=max_steps)
# Initialize cart parameters
self.parameters = []
for i in range(self.cart.pos_dim):
param = paraset.RuntimeParamGetter(param_type=float,
name='acc_%d' % i,
init_val=0,
description='cart acceleration')
self.parameters.append(param)
update_rate = rospy.get_param('~tick_rate')
self.timer = rospy.Timer(period=rospy.Duration(1.0 / update_rate),
callback=self.timer_callback)
self.last_acc = np.zeros(2)
self.rew_pub = rospy.Publisher('~reward', RewardStamped, queue_size=10)
self.break_pub = rospy.Publisher('~breaks',
EpisodeBreak,
queue_size=10)
# Set up state generator and transmitter
self.state_tx = broadcast.Transmitter(
stream_name='cart_state',
feature_size=self.cart.state_dim,
description='Test continuous problem state',
mode='push',
queue_size=10)
reward = self.cart.reset()
self.__publish_cart(time=rospy.Time.now(), reward=reward)
#!/usr/bin/env python
import rospy
import broadcast
if __name__ == '__main__':
rospy.init_node('test_transmitter')
push_tx = broadcast.Transmitter(stream_name='test_push_stream',
feature_size=1,
description='A test push stream',
mode='push',
queue_size=10)
pull_tx = broadcast.Transmitter(stream_name='test_pull_stream',
feature_size=1,
description='A test pull stream',
mode='pull',
cache_time=1.0)
counter = 0
rate = rospy.Rate(1.0)
while not rospy.is_shutdown():
now = rospy.Time.now()
data = [counter]
rospy.loginfo('Publishing %s at %s', str(data), str(now))
push_tx.publish(now, data)
pull_tx.publish(now, data)
counter += 1