def test_sprites_color(self):
(x, _), (_, _), _, _ = load_cifar10(raw=True)
n = 500
x = x[:n]
sprite = create_sprite(x)
f_name = "test_cifar.jpg"
path = os.path.join(ART_DATA_PATH, f_name)
save_image(sprite, path)
self.assertTrue(os.path.isfile(path))
os.remove(path) # Remove data added
def test_sprites_gray(self):
# Get MNIST
(x, _), (_, _), _, _ = load_mnist(raw=True)
n = 100
x = x[:n]
sprite = create_sprite(x)
f_name = "test_sprite_mnist.png"
path = os.path.join(ART_DATA_PATH, f_name)
save_image(sprite, path)
self.assertTrue(os.path.isfile(path))
os.remove(path) # Remove data added
def test_save_image(self):
(x, _), (_, _), _, _ = load_mnist(raw=True)
f_name = "image1.png"
save_image(x[0], f_name)
path = os.path.join(ART_DATA_PATH, f_name)
self.assertTrue(os.path.isfile(path))
os.remove(path)
f_name = "image2.jpg"
save_image(x[1], f_name)
path = os.path.join(ART_DATA_PATH, f_name)
self.assertTrue(os.path.isfile(path))
os.remove(path)
folder = "images123456"
f_name_with_dir = os.path.join(folder, "image3.png")
save_image(x[3], f_name_with_dir)
path = os.path.join(ART_DATA_PATH, f_name_with_dir)
self.assertTrue(os.path.isfile(path))
os.remove(path)
os.rmdir(os.path.split(path)[0]) # Remove also test folder
folder = os.path.join("images123456", "inner")
f_name_with_dir = os.path.join(folder, "image4.png")
save_image(x[3], f_name_with_dir)
path_nested = os.path.join(ART_DATA_PATH, f_name_with_dir)
self.assertTrue(os.path.isfile(path_nested))
os.remove(path_nested)
os.rmdir(os.path.split(path_nested)[0]) # Remove inner test folder
os.rmdir(os.path.split(path)[0]) # Remove also test folder
def visualize_clusters(self,
x_raw: np.ndarray,
save: bool = True,
folder: str = ".",
**kwargs) -> List[List[np.ndarray]]:
"""
This function creates the sprite/mosaic visualization for clusters. When save=True,
it also stores a sprite (mosaic) per cluster in art.config.ART_DATA_PATH.
:param x_raw: Images used to train the classifier (before pre-processing).
:param save: Boolean specifying if image should be saved.
:param folder: Directory where the sprites will be saved inside art.config.ART_DATA_PATH folder.
:param kwargs: a dictionary of cluster-analysis-specific parameters.
:return: Array with sprite images sprites_by_class, where sprites_by_class[i][j] contains the
sprite of class i cluster j.
"""
self.set_params(**kwargs)
if not self.clusters_by_class:
self.cluster_activations()
x_raw_by_class = self._segment_by_class(x_raw, self.y_train)
x_raw_by_cluster: List[List[np.ndarray]] = [ # type: ignore
[[] for _ in range(self.nb_clusters)]
for _ in range(self.classifier.nb_classes) # type: ignore
]
# Get all data in x_raw in the right cluster
for n_class, cluster in enumerate(self.clusters_by_class):
for j, assigned_cluster in enumerate(cluster):
x_raw_by_cluster[n_class][assigned_cluster].append(
x_raw_by_class[n_class][j])
# Now create sprites:
sprites_by_class: List[List[np.ndarray]] = [ # type: ignore
[[] for _ in range(self.nb_clusters)]
for _ in range(self.classifier.nb_classes) # type: ignore
]
for i, class_i in enumerate(x_raw_by_cluster):
for j, images_cluster in enumerate(class_i):
title = "Class_" + str(i) + "_cluster_" + str(
j) + "_clusterSize_" + str(len(images_cluster))
f_name = title + ".png"
f_name = os.path.join(folder, f_name)
sprite = create_sprite(np.array(images_cluster))
if save:
save_image(sprite, f_name)
sprites_by_class[i][j] = sprite
return sprites_by_class
def visualize_clusters(self, x_raw, save=True, folder='.', **kwargs):
"""
This function creates the sprite/mosaic visualization for clusters. When save=True,
it also stores a sprite (mosaic) per cluster in DATA_PATH.
:param x_raw: Images used to train the classifier (before pre-processing)
:type x_raw: `np.darray`
:param save: Boolean specifying if image should be saved
:type save: `bool`
:param folder: Directory where the sprites will be saved inside DATA_PATH folder
:type folder: `str`
:param kwargs: a dictionary of cluster-analysis-specific parameters
:type kwargs: `dict`
:return: sprites_by_class: Array with sprite images sprites_by_class, where sprites_by_class[i][j] contains the
sprite of class i cluster j.
:rtype: sprites_by_class: `np.ndarray`
"""
self.set_params(**kwargs)
if not self.clusters_by_class:
self.cluster_activations()
x_raw_by_class = self._segment_by_class(x_raw, self.y_train)
x_raw_by_cluster = [[[] for x in range(self.nb_clusters)]
for y in range(self.classifier.nb_classes)]
# Get all data in x_raw in the right cluster
for n_class, cluster in enumerate(self.clusters_by_class):
for j, assigned_cluster in enumerate(cluster):
x_raw_by_cluster[n_class][assigned_cluster].append(
x_raw_by_class[n_class][j])
# Now create sprites:
sprites_by_class = [[[] for x in range(self.nb_clusters)]
for y in range(self.classifier.nb_classes)]
for i, class_i in enumerate(x_raw_by_cluster):
for j, images_cluster in enumerate(class_i):
title = 'Class_' + str(i) + '_cluster_' + str(
j) + '_clusterSize_' + str(len(images_cluster))
f_name = title + '.png'
f_name = os.path.join(folder, f_name)
sprite = create_sprite(images_cluster)
if save:
save_image(sprite, f_name)
sprites_by_class[i][j] = sprite
return sprites_by_class
