def getPath(self, start, end):
query = {}
query["sx"] = start[0]
query["sy"] = start[1]
query["ex"] = end[0]
query["ey"] = end[1]
cursor = mongo.db.paths.find(query)
if cursor.count() > 0:
path = cursor[0]["path"]
else:
came_from, cost_so_far = astar.a_star_search(self.grid, start, end)
path = [end]
e = end
try:
while( e != start ):
e = came_from[e]
path.append(e)
path.append(start)
except KeyError:
#Some paths may be impossible
path = [end, start]
finally:
path.reverse()
query["path"] = path
mongo.db.paths.insert(query)
return path
def new_level(player, player_loc, depth, dim):
''' Generate a new level.
Args:
player <Player>: The actor controlled by the player.
loc <Point>: The location of the player on the last level. (Assumption: dim is constant?)
depth <int>: The number of the floor, starting from 1.
dim <int>: The number of rows & columns to generate.
Returns:
map<Point, Cell>: A map from locations to the cells therein.
'''
for i in range(100):
# generate terrain
cells = gen_room(dim)
# place the player
assert player_loc in cells
cells[player_loc] = CL_FLOOR()
cells[player_loc].actor = player
# place the stairs out
stairs_loc = Point(int(dim/2), 1 if depth % 2 else dim - 2) # opposite the player, alternating
cells[stairs_loc] = CL_STAIR()
# place enemies
add_enemies(player, player_loc, cells, depth)
# sanity check
path = a_star_search(cells, player_loc, stairs_loc, lambda cell: not cell.solid)
if path:
return cells
# otherwise, generate a new level
# TODO: would be nice to log this somehow
raise RuntimeError("Couldn't generate a valid level!")
def add_enemies(player, player_loc, cells, depth):
''' Generate enemies and place them on the level.
Args:
player (Player): The character controlled by the player.
player_loc (Point): The location of the player.
cells (map<Point, Cell>): A map of cells' locations to the contents of hte cells.
depth (int): The depth of the level.
'''
assert player_loc != None
assert player_loc in cells
# spawn 1 mongoose per depth
for i in range(depth):
for i in range(10000): # eventually bail
# repeatedly choose random cells
loc, cell = choice(list(cells.items()))
# discard ones that already have terrain or other actors
if cell.is_full():
continue
# also discard those which can't get to the player, or which are too close
to_player = a_star_search(cells, loc, player_loc, lambda cell: not cell.solid)
if not to_player or len(to_player) < 4:
continue
# actually place an enemy
cell.actor = Ghost()
break
def move_toward_player(self, area):
''' Attempt to move toward the player.
Args:
area (Area): The area the ghost is in.
Returns:
bool: whether the ghost successfully moved.
'''
# TODO: deduplicate with mongoose code
player = area.get_player()
player_loc = area.find_actor(player)
assert player_loc != None
cur_loc = area.find_actor(self)
assert cur_loc != None
def passable_metric(cell):
''' Can the cell be pathed through? '''
if cell.actor and not cell.actor.is_mobile():
return False # don't path through fixed creatures
return True
# attempt to move toward the player
path = a_star_search(area.cells, cur_loc, player_loc, passable_metric)
# TODO: improve performance when the player is unreachable
if path:
return self.attempt_move(path[0] - cur_loc, area, None)
return False
def getPlan(self, request):
if self.map is None:
rospy.logwarn(
"Plan requested before map was made available! Please try again later!"
)
return GetPlanResponse()
w = self.map.info.width
h = self.map.info.height
d = self.map.data
res = self.map.info.resolution
o = self.map.info.origin.position
rospy.loginfo("Map metadata: {}".format(self.map.info))
rospy.loginfo("Origin of occupancy grid: ({},{})".format(o.x, o.y))
gx = int(float(request.goal.pose.position.x - o.x) / res)
gy = int(float(request.goal.pose.position.y - o.y) / res)
sx = int(float(request.start.pose.position.x - o.x) / res)
sy = int(float(request.start.pose.position.y - o.y) / res)
rospy.loginfo(
"Planning path from ({},{}) to ({},{}) across {} points with resolution {}"
.format(sx, sy, gx, gy, w * h, res))
grid = [[0 for i in range(w)] for j in range(h)]
num_ones = 0
num_zeros = 0
for r in range(h):
for c in range(w):
v = d[r * w + c]
grid[r][
c] = 1 if v > self.occupied_threshold * 100 or v < 0 else 0
if grid[r][c] == 1:
num_ones += 1
else:
num_zeros += 1
rospy.loginfo("Thresholding complete! {} ones, {} zeros.".format(
num_ones, num_zeros))
rospy.loginfo("value at goal: {}".format(grid[sy][sx]))
traj = a_star_search(grid, (sx, sy), (gx, gy))
if traj is None:
rospy.logwarn("No valid path found to goal! Returning")
return GetPlanResponse()
traj = [[pt[0] * res, pt[1] * res] for pt in traj]
rospy.loginfo("...Found path of length {} points".format(len(traj)))
response = GetPlanResponse()
response.plan = traj2path(traj)
return response
def debug_astar(area, scr):
player = [actor for actor in area.all_actors() if actor.is_player()][0]
mongoosen = [actor for actor in area.all_actors() if actor.cur_glyph() == 'o']
if not mongoosen:
return
mongoose = mongoosen[0]
mongoose_loc = area.find_actor(mongoose)
player_loc = area.find_actor(player)
from astar import a_star_search
path = a_star_search(area.cells, mongoose_loc, player_loc) or []
for p in path:
scr.addstr(p.y, p.x, '*'.encode('utf-8'))
def main():
"""main program
:param argv[1]: Start city
:param argv[2]: Dest city
:param argv[3]: Heuristique
:returns: List of the cities to go through
Correct syntax:
python app.py start_city dest_city heuristique
Concrete example :
python app.py Copenhagen Lisbon h0
List of available heuristiques:
h0 -> 0
h1 -> abs(city1.x-city2.x)
h2 -> abs(city1.y-city2.y)
h3 -> math.sqrt((c1.x-c2.x)**2+(c1.y-c2.y)**2)
h4 -> h1(c1,c2)+h2(c1,c2)
Notes: Cities' names are case insensitive
Dependencies:
positions.txt - Contains a list of all cities
connections.txt - Contains a list of all connections between cities
"""
import sys
if len(sys.argv) != 4:
print("Invalid use!")
print(main.__doc__)
exit(-1)
cities = load_config()
start_city = check_city(cities, sys.argv[1])
dest_city = check_city(cities, sys.argv[2])
heuristique = check_heuristique(MapHeuristic, sys.argv[3])
(cities, total_length,
nb_iteration) = a_star_search(start_city, dest_city, heuristique)
print('RESULT :')
print(' -> '.join([city.name for city in cities]))
print(f"Total Length : {total_length}")
print(f"Nb iterations : {nb_iteration}")
def walkToRedCell(self, grid, pos):
gridFind = astar.GridWithWeights(len(grid), len(grid[0]))
new = []
for i in range(0, len(grid)):
for j in range(0, len(grid[0])):
if grid[i][j].flag == 1:
new.append((i, j))
gridFind.weights = {loc: 5 for loc in new}
gridFind.walls = new
end = (self.ideal[0], self.ideal[1])
start = (pos[0], pos[1])
came, cost = astar.a_star_search(gridFind, start, end)
path = astar.reconstruct_path(came, start, end)
self.path = path
# print(path)
# astar.draw_grid(gridFind, width=2, path=astar.reconstruct_path(came, start=start, goal=end))
return path
def main():
# Cross Configuration
root = board.translate_input_to_array(['--000--',
'--0X0--',
'00XXX00',
'000X000',
'000X000',
'--000--',
'--000--'])
# Plus Configuration
# root = board.translate_input_to_array(['--000--',
# '--0X0--',
# '000X000',
# '0XXXXX0',
# '000X000',
# '--0X0--',
# '--000--'])
# Fireplace Configuration
# root = board.translate_input_to_array(['--XXX--',
# '--XXX--',
# '00XXX00',
# '00X0X00',
# '0000000',
# '--000--',
# '--000--'])
# Up Configuration
# root = board.translate_input_to_array(['--0X0--',
# '--XXX--',
# '0XXXXX0',
# '000X000',
# '000X000',
# '--XXX--',
# '--XXX--'])
# Pyramid Configuration
# root = board.translate_input_to_array(['--000--',
# '--0X0--',
# '00XXX00',
# '0XXXXX0',
# 'XXXXXXX',
# '--000--',
# '--000--'])
# Diamond Configuration
# root = board.translate_input_to_array(['--0X0--',
# '--XXX--',
# '0XXXXX0',
# 'XXX0XXX',
# '0XXXXX0',
# '--XXX--',
# '--0X0--'])
# Goal Configuration
goal = board.translate_input_to_array(['--000--',
'--000--',
'0000000',
'000X000',
'0000000',
'--000--',
'--000--'])
hp = hpy()
before = hp.heap()
start_time = time.time()
count, parent = astar.a_star_search(root, goal)
end_time = time.time()
after = hp.heap()
answer = astar.backtrack(parent, root, goal)
leftover = after - before
print "A* Search: Total Manhattan distance heuristic"
print "Total number of nodes expanded: %d" % count
print "Answer depth: %d" % len([move for move, _ in answer])
print "Moves: %s" % [move for move, _ in answer]
print "Time: %d seconds" % (end_time - start_time)
# print leftover
print "---\n"
before = hp.heap()
start_time = time.time()
count, parent = astar.a_star_search(root, goal, simple=True)
end_time = time.time()
after = hp.heap()
answer = astar.backtrack(parent, root, goal)
leftover = after - before
print "A* Search: Number of pegs heuristic"
print "Total number of nodes expanded: %d" % count
print "Answer depth: %d" % len([move for move, _ in answer])
print "Moves: %s" % [move for move, _ in answer]
print "Time: %d seconds" % (end_time - start_time)
# print leftover
print "---\n"
before = hp.heap()
start_time = time.time()
depth, count = pruned_ids.iterative_deepening_search(root, goal)
end_time = time.time()
after = hp.heap()
leftover = after - before
moves = pruned_ids.solution_moves[::-1]
print "Pruned Iterative Deepening Search:"
print "Total number of nodes expanded: %d" % count
print "Answer depth: %d" % depth
print "Moves: %s" % moves
print "Time: %d seconds" % (end_time - start_time)
# print leftover
print "---\n"
before = hp.heap()
start_time = time.time()
depth, count = ids.iterative_deepening_search(root, goal)
end_time = time.time()
after = hp.heap()
leftover = after - before
moves = ids.solution_moves[::-1]
print "Iterative Deepening Search:"
print "Total number of nodes expanded: %d" % count
print "Answer depth: %d" % depth
print "Moves: %s" % moves
print "Time: %d seconds" % (end_time - start_time)
# print leftover
print "---\n"
