def import_bot_instance(
raw_game_data: Response,
raw_game_info: Response,
raw_observation: ResponseObservation,
) -> BotAI:
"""
import_bot_instance DocString
"""
bot = BotAI()
game_data = GameData(raw_game_data.data)
game_info = GameInfo(raw_game_info.game_info)
game_state = GameState(raw_observation)
# noinspection PyProtectedMember
bot._initialize_variables()
# noinspection PyProtectedMember
bot._prepare_start(
client=None, player_id=1, game_info=game_info, game_data=game_data
)
# noinspection PyProtectedMember
bot._prepare_first_step()
# noinspection PyProtectedMember
bot._prepare_step(state=game_state, proto_game_info=raw_game_info)
# noinspection PyProtectedMember
bot._find_expansion_locations()
return bot
def get_map_specific_bot(map_path: Path) -> BotAI:
assert map_path in MAPS
with lzma.open(str(map_path.absolute()), "rb") as f:
raw_game_data, raw_game_info, raw_observation = pickle.load(f)
# Build fresh bot object, and load the pickle'd data into the bot object
bot = BotAI()
game_data = GameData(raw_game_data.data)
game_info = GameInfo(raw_game_info.game_info)
game_state = GameState(raw_observation)
bot._initialize_variables()
bot._prepare_start(client=None, player_id=1, game_info=game_info, game_data=game_data)
bot._prepare_step(state=game_state, proto_game_info=raw_game_info)
return bot
def get_map_specific_bots() -> Iterable[BotAI]:
folder = os.path.dirname(__file__)
subfolder_name = "pickle_data"
pickle_folder_path = os.path.join(folder, subfolder_name)
files = os.listdir(pickle_folder_path)
for file in (f for f in files if f.endswith(".xz")):
with lzma.open(os.path.join(folder, subfolder_name, file), "rb") as f:
raw_game_data, raw_game_info, raw_observation = pickle.load(f)
# Build fresh bot object, and load the pickle'd data into the bot object
bot = BotAI()
game_data = GameData(raw_game_data.data)
game_info = GameInfo(raw_game_info.game_info)
game_state = GameState(raw_observation)
bot._initialize_variables()
bot._prepare_start(client=None, player_id=1, game_info=game_info, game_data=game_data)
bot._prepare_step(state=game_state, proto_game_info=raw_game_info)
yield bot
class Creeper:
"""Spread creep."""
def __init__(
self, raw_game_data: Any, raw_game_info: Any, raw_observation: Any
) -> None:
"""
Set up variables for use within Creeper.
Args:
raw_game_data (Any): self.game_data from main instance
raw_game_info (Any): self.game_info from main instance
raw_observation (Any): self.game_state from main instance
Returns:
None
"""
self.bot = BotAI()
game_data = GameData(raw_game_data.data)
game_info = GameInfo(raw_game_info.game_info)
game_state = GameState(raw_observation)
self.bot._initialize_variables()
self.bot._prepare_start(
client=None, player_id=1, game_info=game_info, game_data=game_data
)
self.bot._prepare_step(state=game_state, proto_game_info=raw_game_info)
self.pathing = PathManager(raw_game_data, raw_game_info, raw_observation)
self.pf = PathFind(self.pathing.map_grid)
def check_tumor_position(
self, possible_position: Point2, combined_grid: Any
) -> int:
"""
Calculate the number of tiles the tumor position would cover.
Arg:
possible_position (Point2): the point being checked.
combined_grid (Any): ndarray that's the pathfinding lib's path map and the
creep map added together
Returns:
int: the number of tiles that would be covered.
"""
x, y = floor(possible_position[0]), floor(possible_position[1])
future_tiles = 0
for i in range(-10, 11):
for j in range(-10, 11):
if 8 <= abs(j) <= 10:
if abs(i) <= 6 - 2 * (abs(j) - 8):
future_tiles += combined_grid[Point2((x + i, y + j))] % 2
elif abs(j) == 7:
if abs(i) <= 7:
future_tiles += combined_grid[Point2((x + i, y + j))] % 2
elif 3 <= abs(j) <= 6:
if abs(i) <= 9 - floor(abs(j) / 5):
future_tiles += combined_grid[Point2((x + i, y + j))] % 2
else:
if combined_grid[Point2((x + i, y + j))] == 1:
future_tiles += 1
return future_tiles
async def find_position(
self,
tumor: Unit,
tumor_positions: Set[Point2],
creep_grid: Any,
pathable_map: Any,
) -> Point2:
"""
Find the location to spread the tumor to.
Args:
tumor (Unit): the creep tumor ready to be spread
tumor_positions (Set[Point2]): list of existing tumor locations
creep_grid (ndarray): the creep grid from the main bot
pathable_map (ndarray): the pathfinding lib's pathable map
Returns:
Point2: where to spread the tumor.
"""
path_creep_map = creep_grid + pathable_map
tposx, tposy = tumor.position.x, tumor.position.y
max_tiles = 0
location = None
for i in range(-10, 11):
for j in range(-10, 11):
if 81 <= i ** 2 + j ** 2 <= 105:
pos = Point2((tposx + i, tposy + j))
if pos not in tumor_positions:
if path_creep_map[floor(pos[0])][floor(pos[1])] != 2:
continue
tiles = self.check_tumor_position((pos), path_creep_map)
if tiles > max_tiles:
max_tiles = tiles
location = pos
if max_tiles < 75:
floored_unit_pos = Point2((floor(tposx), floor(tposy)))
floored_e_base = Point2(
(
floor(self.bot.enemy_start_locations[0].position.x),
floor(self.bot.enemy_start_locations[0].position.y),
)
)
path_to_e_base = self.pf.find_path(floored_unit_pos, floored_e_base)[0]
if path_to_e_base:
for k in range(9, 5, -1):
pos = path_to_e_base[k]
if path_creep_map[pos[0]][pos[1]] == 2:
location = Point2(pos)
break
if location:
return location
else:
return tumor.position
class PathManager:
"""Manage unit pathing."""
def __init__(
self, raw_game_data: Any, raw_game_info: Any, raw_observation: Any
) -> None:
"""
Set up variables for use within PathMangager.
Args:
raw_game_data (Any): self.game_data from main instance
raw_game_info (Any): self.game_info from main instance
raw_observation (Any): self.game_state from main instance
Returns:
None
"""
self.bot = BotAI()
game_data = GameData(raw_game_data.data)
game_info = GameInfo(raw_game_info.game_info)
game_state = GameState(raw_observation)
self.bot._initialize_variables()
self.bot._prepare_start(
client=None, player_id=1, game_info=game_info, game_data=game_data
)
self.bot._prepare_step(state=game_state, proto_game_info=raw_game_info)
map_name = game_info.map_name
self.pathing_dict: Dict[int, PathDict] = {}
map_width = game_info.map_size.width
map_height = game_info.map_size.height
raw_grid = np.zeros((map_width, map_height))
for x in range(map_width):
for y in range(map_height):
pos = Point2((x, y))
raw_grid[x][y] = game_info.pathing_grid[pos]
self.map_grid = np.rot90(raw_grid.astype(int))
np.save(f"map_grids/{map_name}_grid", self.map_grid)
self.pf = PathFind(self.map_grid)
def add_to_path_dict(self, unit: Unit, destination: Point2) -> None:
"""
Add unit's path to the path storage dictionary.
Args:
unit (Unit): the unit for pathing
destination (Point2): where the unit is going
Returns:
None
"""
self.pathing_dict[unit.tag] = {"path": [], "step": 0}
raw_unit_pos = unit.position
floored_unit_pos = Point2((floor(raw_unit_pos.x), floor(raw_unit_pos.y)))
floored_dest = Point2((floor(destination[0]), floor(destination[1])))
self.pathing_dict[unit.tag]["path"] = self.pf.find_path(
floored_unit_pos, floored_dest
)[0]
def follow_path(self, unit: Unit, default: Point2) -> Point2:
"""
Follow the path set or set a new one if none exists.
Args:
unit (Unit): the unit moving
Returns:
Point2: the location to attack
"""
if (
len(
self.bot.structures.filter(
lambda unit: unit.type_id
in {UnitTypeId.NYDUSNETWORK, UnitTypeId.NYDUSCANAL}
)
)
< 2
):
if unit.tag not in self.pathing_dict:
self.add_to_path_dict(unit, tuple(default))
advance_factor = int(unit.movement_speed) + 2
self.pathing_dict[unit.tag]["step"] += advance_factor
curr_step = self.pathing_dict[unit.tag]["step"]
if curr_step >= len(self.pathing_dict[unit.tag]["path"]):
curr_step = len(self.pathing_dict[unit.tag]["path"]) - 1
if curr_step < 0:
return default
a_move_to = Point2(self.pathing_dict[unit.tag]["path"][curr_step])
if curr_step == len(self.pathing_dict[unit.tag]["path"]) - 1:
del self.pathing_dict[unit.tag]
return a_move_to