def __init__(self):
self.env_name = 'gaf-environment-v1.0'
self.reward_mult = 100000
self.cash = 100000
self.total_loss = 0.0
self.current_action = None
self.previous_action = None
self.clock = Clock()
self.dm = DataManager(self.clock)
self.pm = PositionManager(self.clock, self.dm, self.cash, 30)
self.benchmark = PositionManager(self.clock, self.dm, self.cash, 1)
self.symbols = self.dm.get_symbols()
self.action_space = gym.spaces.Discrete(3)
self.observation_space = gym.spaces.Box(low=-1,
high=1,
shape=(4, 30, 180))
self.symbols = None
self.final_cash_value = []
#self.print_intro()
self.avg_reward = 0
self.episodes_ran = 0
self.perm_symbols = [
self.dm.current_symbol,
]
self.returns = pd.DataFrame()
def __init__(self):
# Publish robot state (position, orientation, tool)
self.niryo_one_robot_state_publisher = NiryoRobotStatePublisher()
# Position Manager
positions_dir = rospy.get_param("~positions_dir")
self.pos_manager = PositionManager(positions_dir)
# trajectory_manager
trajectories_dir = rospy.get_param("~trajectories_dir")
self.traj_manager = TrajectoryManager(trajectories_dir)
# robot commander
self.robot_commander = RobotCommander(self.pos_manager, self.traj_manager)
self.robot_commander.start()
def __init__(self):
# Publish robot state (position, orientation, gripper)
self.niryo_one_robot_state_publisher = NiryoRobotStatePublisher()
# Position Manager
positions_dir = rospy.get_param("~positions_dir")
self.pos_manager = PositionManager(positions_dir)
# robot commander
self.robot_commander = RobotCommander(self.pos_manager)
# Load ROS action interface
self.robot_action_server = RobotActionServer(self.pos_manager)
self.robot_action_server.start()
def __init__(self):
#init RosLogger
self.debug_mode = rospy.get_param("~debug_mode")
self.gauss_ros_logger = RosLogger(debug_mode = self.debug_mode)
# Publish robot state (position, orientation, tool)
self.gauss_robot_state_publisher = GaussRobotStatePublisher()
self.simulator_mode = rospy.get_param("~simulator_mode")
# Position Manager
positions_dir = rospy.get_param("~positions_dir")
self.pos_manager = PositionManager(positions_dir, self.gauss_ros_logger)
#trajectory_manager
trajectories_dir = rospy.get_param("~trajectories_dir")
self.traj_manager = TrajectoryManager(trajectories_dir, self.gauss_ros_logger)
# robot commander
self.robot_commander = RobotCommander(self.simulator_mode, self.pos_manager, self.traj_manager, self.gauss_ros_logger)
self.robot_commander.start()
self.gauss_ros_logger.publish_log_status("INFO", "GaussCommanderNode start over")
def getExecutorMockWithPM(self):
#pm = PositionDBManager(db)
pm = PositionManager()
e = self.getExecutorMock()
e.setPositionManager(pm)
return e, pm
class StockEnv(gym.Env):
def __init__(self):
self.env_name = 'gaf-environment-v1.0'
self.reward_mult = 100000
self.cash = 100000
self.total_loss = 0.0
self.current_action = None
self.previous_action = None
self.clock = Clock()
self.dm = DataManager(self.clock)
self.pm = PositionManager(self.clock, self.dm, self.cash, 30)
self.benchmark = PositionManager(self.clock, self.dm, self.cash, 1)
self.symbols = self.dm.get_symbols()
self.action_space = gym.spaces.Discrete(3)
self.observation_space = gym.spaces.Box(low=-1,
high=1,
shape=(4, 30, 180))
self.symbols = None
self.final_cash_value = []
#self.print_intro()
self.avg_reward = 0
self.episodes_ran = 0
self.perm_symbols = [
self.dm.current_symbol,
]
self.returns = pd.DataFrame()
def get_frame(self):
return self.dm.get_frame()
def reset(self):
self.total_loss = 0.0
if self.episodes_ran > 0:
self.print_returns()
self.clock.reset()
self.cash = 100000
self.perm_symbols = []
self.dm.reset()
frame = self.dm.get_frame()
self.first_frame = frame
self.episodes_ran += 1
return frame
def position_to_close(self, index):
return self.pm.position_expired(index)
def close_positions(self):
return self.pm.close_positions()
def update_action_count(self, action):
if action == actions.Actions.BUY:
buy_count += 1
elif action == actions.Actions.SELL:
sell_count += 1
elif action == actions.Actions.HOLD:
hold_count += 1
def calculate_ma(self, reward, period=14):
return reward
def step(self, action):
self.update_action_count(action)
done = self.clock.done()
frame = self.dm.get_frame() if not done else self.first_frame
reward = self.pm.close_position() * self.reward_mult
info = {}
if not np.isfinite(reward):
reward = 0
if done == self.dm.INCR_FLAG:
self.final_cash_value.append(self.cash)
self.dm.increment_symbol()
self.perm_symbols.append(self.dm.current_symbol)
print(self.cash)
self.cash = 100000
done = False
self.pm.open_position(action, self.clock.index)
self.clock.tick()
elif done == False:
self.pm.open_position(action, self.clock.index)
self.update_cash(reward)
self.clock.tick()
else:
self.returns.append(pd.Series(reward), ignore_index=True)
self.update_cash(reward)
self.final_cash_value.append(self.cash)
if self.cash < 0.0:
print('Total Loss of Capital')
self.total_loss += 100000
self.cash = 100000
reward = -5
return frame, reward, done, info
def update_cash(self, reward):
reward /= self.reward_mult
self.old_cash = self.cash
self.cash = (1 + reward) * self.cash
def print_intro(self):
start = self.dm.get_date()
end = self.dm.get_ending_date()
delta = end - start
print('Preparing environment for:', fg('red'),
self.dm.get_symbol() + attr('reset'))
print('Starting date:', start, 'Ending Date:', end)
print("Time period:", delta.days, 'days')
print('Starting Balance:', fg('green'), self.get_cash(), attr('reset'))
def get_cash(self, value=None):
source = float(self.cash if value is None else value)
return '${value:,.2f}'.format(value=source)