def solve(prog, quantic=False):
"""Read the <prog>ram memory instructions and realize the necessary bit
maskings in the cells of the memory.
If <quantic>, treat the "X"s as floating ("quantic") bits, meaning that
the bits can assume every possible permutation. """
mask = "0" * 36
mem = {}
for ins in prog:
cmd = findall(r"^(\w+)", ins)[0]
if cmd == "mask":
mask = findall(MASK_REG, ins)[0]
elif cmd == "mem":
idx, val = regex([ins], (int, int), MEM_REG)[0]
if not quantic:
mem[idx] = masker(mask, val)
else:
idx |= int(sub(r"[^1]", "0", mask), 2)
# Intermediary mold.
mold = sub(r"[^X]", "~", mask)
# Count through every binary permutation of <n_bits>.
for i in range(2**(n_bits := mold.count("X"))):
count = bin(i)[2:].rjust(n_bits, "0")
# Fill the "X"s with each bit from the <count>.
filled = "".join(fill(mold, "X", count))
mem[masker(filled, idx)] = val
def __init__(self, drone: Drone, environment: Map):
sensor_energy = ternary(
environment.is_sensor(drone),
min(drone.battery, randint(0, MAX_SENSOR_COVERAGE)), 0)
self.path = [(drone.x, drone.y, sensor_energy)]
self.map = environment
self.spent_energy = fill(0, environment.m, environment.n)
self.spent_energy[drone.x][drone.y] = sensor_energy
self.battery_left = drone.battery - sensor_energy
def main(self):
self.net.eval()
with torch.no_grad():
output = self.net(self.data)
image = utils.fill(output, self.cls_num, cfg.COLOR).cpu().numpy()
image_array = image.squeeze().transpose(1, 2, 0)
image_separate = Image.fromarray(np.uint8(image_array))
figure = Image.new("RGB", (640, 240), (255, 255, 255))
figure.paste(self.image_origin, (0, 0))
figure.paste(image_separate, (self.image_origin.size[0], 0))
return figure.show()
def check_coverage(self):
marked = fill(False, self.map.n, self.map.m)
for i, j, energy in self.path:
if not energy:
continue
neighbours = [[i, j]] * 4
for _ in range(energy):
for direction_index, direction in enumerate(DIRECTIONS):
new_neighbour = deepcopy(neighbours[direction_index])
new_neighbour[0] += direction[0]
new_neighbour[1] += direction[1]
if self.map.is_tuple_in_bounds(
new_neighbour
) and not self.map.is_tuple_wall(new_neighbour):
marked[new_neighbour[0]][new_neighbour[1]] = True
neighbours[direction_index] = new_neighbour
return sum([row.count(True) for row in marked])
def choose_rhythm(self, dur_to_fill):
durs = fill(dur_to_fill * 4, min_note_dur=5)
durs = [d / 4.0 for d in durs]
notes = []
if durs[0] > 0:
notes.append({
'pitch': 'rest',
'duration': durs[0]
})
notes.append({
'pitch': None,
'duration': durs[1]
})
if durs[2] > 0:
notes.append({
'pitch': 'rest',
'duration': durs[2]
})
return notes
def choose_rhythm(self, dur_to_fill):
durs = fill(dur_to_fill * 2)
durs = [d / 2.0 for d in durs]
return durs
from lstm import LSTM
print("\nParameters:")
print(FLAGS)
# Data Preparation
# ==================================================
# Load data
print("Loading data...")
x, y, w2ix, ix2w = utils.load_data(max_sents=10)
seq_length = max(len(x) for x in x)
vocab_size = len(w2ix)
x = utils.fill(x, seq_length)
y = utils.fill(y, seq_length)
# Randomly shuffle data
np.random.seed(10)
shuffle_indices = np.random.permutation(np.arange(len(y)))
x_shuffled = x[shuffle_indices]
y_shuffled = y[shuffle_indices]
# Split train/test set
dev_sample_index = -1 * int(FLAGS.dev_sample_percentage * float(len(y)))
x_train, x_dev = x_shuffled[dev_sample_index:], x_shuffled[:dev_sample_index]
y_train, y_dev = y_shuffled[dev_sample_index:], y_shuffled[:dev_sample_index]
print("Vocabulary Size: {:d}".format(vocab_size))
print("Train/Dev split: {:d}/{:d}".format(len(y_train), len(y_dev)))