def images_iter(directory, mode):
mode = 'dev' if mode == 'validation' else mode
directory = os.path.join(directory, mode, 'images')
for root, dirs, files in os.walk(directory):
if root == directory:
assert not files
else:
assert not dirs
for filename in files:
assert filename[:len(mode) + 1] == mode + '-'
identifier = filename[len(mode) + 1: -6]
assert identifier[-2:] in ('-0', '-1', '-2', '-3')
with open(os.path.join(root, filename), 'rb') as filehandle:
image = np.array(object=Image.open(fp=filehandle))
world1 = World.from_image(image[:, :100, :])
world2 = World.from_image(image[:, 150:250, :])
world3 = World.from_image(image[:, 300:, :])
yield identifier, (world1, world2, world3)
def images_iter(directory, mode, parts=None):
split = 'val' if mode == 'validation' else mode
if parts is not None:
split += parts[mode]
directory = os.path.join(directory, 'images', split)
for root, dirs, files in os.walk(directory):
assert root == directory
assert not dirs
for n in range(len(files)):
filename = 'CLEVR_{}_{:0>6}.png'.format(split, n)
files.remove(filename)
with open(os.path.join(root, filename), 'rb') as filehandle:
image = Image.open(fp=filehandle)
world = World.from_image(image)
yield world
def deserialize_value(value_name, value_type, read_file, word2id=None):
if value_type == 'int':
value = read_file(value_name + '.txt')
value = [int(x) for x in value.split()]
return value
elif value_type == 'float':
value = read_file(value_name + '.txt')
value = [float(x) for x in value.split()]
return value
elif value_type == 'vector':
value = read_file(value_name + '.txt')
value = [[float(x) for x in vector.split(',')]
for vector in value.split()]
return value
elif value_type == 'text':
assert word2id
value = read_file(value_name + '.txt')
value = [[word2id[word] for word in text.split(' ')]
for text in value.split('\n')[:-1]]
return value
elif value_type == 'world':
value = []
n = 0
while True:
image_bytes = read_file('{}-{}.bmp'.format(value_name, n),
binary=True)
if image_bytes is None:
break
image_bytes = BytesIO(image_bytes)
image = Image.open(image_bytes)
value.append(World.from_image(image))
n += 1
return value
elif value_type == 'model':
value = read_file(value_name + '.json')
value = json.loads(value)
return value
def deserialize_value(value_name,
value_type,
read_file,
num_concat_worlds=0,
word2id=None):
value_type, alts = alternatives_type(value_type=value_type)
if value_type == 'int':
value = read_file(value_name + '.txt')
if alts:
value = [[int(x) for x in xs.split(';')]
for xs in value.split('\n')[:-1]]
else:
value = [int(x) for x in value.split('\n')[:-1]]
return value
elif value_type == 'float':
value = read_file(value_name + '.txt')
if alts:
value = [[float(x) for x in xs.split(';')]
for xs in value.split('\n')[:-1]]
else:
value = [float(x) for x in value.split('\n')[:-1]]
return value
elif value_type == 'vector(int)':
value = read_file(value_name + '.txt')
if alts:
value = [[[int(x) for x in vector.split(',')]
for vector in vectors.split(';')]
for vectors in value.split('\n')[:-1]]
else:
value = [[int(x) for x in vector.split(',')]
for vector in value.split('\n')[:-1]]
return value
elif value_type == 'vector(float)':
value = read_file(value_name + '.txt')
if alts:
value = [[[float(x) for x in vector.split(',')]
for vector in vectors.split(';')]
for vectors in value.split('\n')[:-1]]
else:
value = [[float(x) for x in vector.split(',')]
for vector in value.split('\n')[:-1]]
return value
elif value_type == 'text':
assert word2id
value = read_file(value_name + '.txt')
if alts:
value = [[[word2id[word] for word in text.split(' ')]
for text in texts.split('\n')]
for texts in value.split('\n\n')[:-1]]
else:
value = [[word2id[word] for word in text.split(' ')]
for text in value.split('\n')[:-1]]
return value
elif value_type == 'world':
if num_concat_worlds:
size = ceil(sqrt(num_concat_worlds))
image_bytes = read_file(value_name + '.bmp', binary=True)
assert image_bytes is not None
image_bytes = BytesIO(image_bytes)
image = Image.open(image_bytes)
worlds = World.from_image(image)
height = worlds.shape[0] // ceil(num_concat_worlds / size)
assert worlds.shape[0] % ceil(num_concat_worlds / size) == 0
width = worlds.shape[1] // size
assert worlds.shape[1] % size == 0
value = []
for y in range(ceil(num_concat_worlds / size)):
for x in range(size if y < num_concat_worlds //
size else num_concat_worlds % size):
value.append(worlds[y * height:(y + 1) * height,
x * width:(x + 1) * width, :])
else:
value = []
n = 0
while True:
image_bytes = read_file('{}-{}.bmp'.format(value_name, n),
binary=True)
if image_bytes is None:
break
image_bytes = BytesIO(image_bytes)
image = Image.open(image_bytes)
value.append(World.from_image(image))
n += 1
return value
elif value_type == 'model':
value = read_file(value_name + '.json')
value = json.loads(value)
return value
