def codalab_run(run_id):
if run_id == 0:
## H shape map
env = environment.DiscreteSquareMapEnv(preset=5)
label_data = label(env, end=(2, 5))
desc_agent(env, label_data, mode="back")
stats = {
"turn": env.agent_turns,
"dist": env.agent_distance,
"notes": "Backward"
}
return stats
if run_id == 1:
## H shape map
env = environment.DiscreteSquareMapEnv(preset=5)
label_data = label(env, end=(1, 4))
desc_agent(env, label_data, mode="astar")
stats = {
"turn": env.agent_turns,
"dist": env.agent_distance,
"notes": "A*"
}
return stats
if run_id == 2:
## H shape map
env = environment.DiscreteSquareMapEnv(preset=101)
label_data = label(env, end=(5, 5))
desc_agent(env, label_data, mode="back")
stats = {
"turn": env.agent_turns,
"dist": env.agent_distance,
"notes": "Backward"
}
return stats
if run_id == 3:
## H shape map
env = environment.DiscreteSquareMapEnv(preset=101)
label_data = label(env, end=(5, 5))
desc_agent(env, label_data, mode="astar")
stats = {
"turn": env.agent_turns,
"dist": env.agent_distance,
"notes": "A*"
}
return stats
def codalab_run(run_id):
if run_id == 0:
## H shape map
test_env = environment.DiscreteSquareMapEnv(preset=5)
astar_test(test_env)
stats = {
"turn": test_env.agent_turns,
"dist": test_env.agent_distance,
"notes": "Pure Astar"
}
return stats
if run_id == 1:
## H shape map
test_env = environment.DiscreteSquareMapEnv(preset=5)
astar_test(test_env)
stats = {
"turn": test_env.agent_turns,
"dist": test_env.agent_distance,
"notes": "Pure Astar"
}
return stats
if run_id == 2:
## H shape map
test_env = environment.DiscreteSquareMapEnv(preset=101)
astar_test(test_env)
stats = {
"turn": test_env.agent_turns,
"dist": test_env.agent_distance,
"notes": "Pure Astar"
}
return stats
if run_id == 3:
## H shape map
test_env = environment.DiscreteSquareMapEnv(preset=101)
astar_test(test_env)
stats = {
"turn": test_env.agent_turns,
"dist": test_env.agent_distance,
"notes": "Pure Astar"
}
return stats
def codalab_run(run_id):
random.seed(0)
if run_id == 0:
## H shape map
env = environment.DiscreteSquareMapEnv(preset=5)
random_agent(env)
notes = ""
if env.num_unvisited_nodes() > 0:
notes = "agent trapped"
stats = {
"turn": env.agent_turns,
"dist": env.agent_distance,
"notes": notes
}
return stats
if run_id == 1:
## H shape map
env = environment.DiscreteSquareMapEnv(preset=101)
random_agent(env)
notes = ""
if env.num_unvisited_nodes() > 0:
notes = "agent trapped"
stats = {
"turn": env.agent_turns,
"dist": env.agent_distance,
"notes": notes
}
return stats
def main():
# env = DiscreteSquareMapEnv(map_dim=(6, 6), block_area=(((1, 2), (3, 3)), ((4, 4), (5, 5))))
env = environment.DiscreteSquareMapEnv(preset=5)
print(env.map.data)
label_data = label(env, end=(3, 3))
# print(label_data)
#
# desc_agent(env, label_data, mode="astar")
# # desc_agent(env, label_data)
env.plot_path(label_data=None)
def __init__(self):
self.env = environment.DiscreteSquareMapEnv()
def __init__(self, map_dim=(3, 3), block_area=None, start=(1, 1)):
self.env = environment.DiscreteSquareMapEnv(map_dim, block_area, start)
self.lmapsize = 3
# min_size = 5 # max_size = 25 # obs_corners_x = np.random.randint(0,width,num_obs) # obs_corners_y = np.random.randint(0,height,num_obs) # obs_lower_corners = np.vstack([obs_corners_x,obs_corners_y]).T # obs_sizes = np.random.randint(min_size,max_size,(num_obs,2)) # obs_upper_corners = obs_lower_corners + obs_sizes # obstacles = zip(obs_lower_corners,obs_upper_corners) # occupancy = DetOccupancyGrid2D(width, height, obstacles) # x_init = tuple(np.random.randint(0,width-2,2).tolist()) # x_goal = tuple(np.random.randint(0,height-2,2).tolist()) # while not (occupancy.is_free(x_init) and occupancy.is_free(x_goal)): # x_init = tuple(np.random.randint(0,width-2,2).tolist()) # x_goal = tuple(np.random.randint(0,height-2,2).tolist()) test_env = env.DiscreteSquareMapEnv(preset=5) # test_env = env.DiscreteSquareMapEnv(map_dim=(5, 5), block_area=((None)) # test_env = env.DiscreteSquareMapEnv(map_dim=(5, 5), block_area=(((1, 2), (3, 2)),)) width = test_env.map.height height = test_env.map.width obstacles = list(test_env.block_area) occupancy = DetOccupancyGrid2D(width, height, obstacles) action = test_env.available_actions()[0] iteration = 0 # action list act_list = [] # when agent visited all nearby cells, # record current location and new unvisitied location in this list. jump_node_list = [] # keep looping until all cells visieted while test_env.num_unvisited_nodes() != 0: