def setUp(self):
reset_plugins()
# Generic image collection with images of different shapes.
self.images = ImageCollection(self.pattern)
# Image collection with images having shapes that match.
self.images_matched = ImageCollection(self.pattern_matched)
# Same images as a collection of frames
self.frames_matched = MultiImage(self.pattern_matched)
def setUp(self):
# This multipage TIF file was created with imagemagick:
# convert im1.tif im2.tif -adjoin multipage.tif
use_plugin('pil')
paths = [os.path.join(data_dir, 'multipage_rgb.tif'),
os.path.join(data_dir, 'no_time_for_that_tiny.gif')]
self.imgs = [MultiImage(paths[0]),
MultiImage(paths[0], conserve_memory=False),
MultiImage(paths[1]),
MultiImage(paths[1], conserve_memory=False),
ImageCollection(paths[0]),
ImageCollection(paths[1], conserve_memory=False),
ImageCollection(os.pathsep.join(paths))]
def setUp(self):
# This multipage TIF file was created with imagemagick:
# convert im1.tif im2.tif -adjoin multipage.tif
paths = [
os.path.join(data_dir, 'multipage.tif'),
os.path.join(data_dir, 'no_time_for_that.gif')
]
self.imgs = [
MultiImage(paths[0]),
MultiImage(paths[0], conserve_memory=False),
MultiImage(paths[1]),
MultiImage(paths[1], conserve_memory=False),
ImageCollection(paths[0]),
ImageCollection(paths[1], conserve_memory=False),
ImageCollection('%s:%s' % (paths[0], paths[1]))
]
def test_custom_load_func(self):
def load_fn(x):
return x
ic = ImageCollection(os.pathsep.join(self.pattern), load_func=load_fn)
assert_equal(ic[0], self.pattern[0])
def load_from_dir(directory, extension):
"""
load_from_dir: Takes in the directory path and extension of images to be read, and reads them in.
Args:
(String) directroy: path to directory
(String) extension: extension of the files to be read
Returns:
(Numpy Array)data: The normal data w the 3 channels
(Numpy Array) data_gray: the data converted to gray scale format
"""
data = np.array(
ImageCollection('{a}/*.{b}'.format(a=directory, b=extension),
load_func=imread))
data_gray = np.array(
ImageCollection('{a}/*.{b}'.format(a=directory, b=extension),
load_func=imread_grayscale))
return (data, data_gray)
def test_custom_load(self):
load_pattern = [(1, 'one'), (2, 'two')]
def load_fn(x):
return x
ic = ImageCollection(load_pattern, load_func=load_fn)
assert_equal(ic[1], (2, 'two'))
def imgs():
use_plugin('pil')
paths = [
os.path.join(data_dir, 'multipage_rgb.tif'),
os.path.join(data_dir, 'no_time_for_that_tiny.gif')
]
imgs = [
MultiImage(paths[0]),
MultiImage(paths[0], conserve_memory=False),
MultiImage(paths[1]),
MultiImage(paths[1], conserve_memory=False),
ImageCollection(paths[0]),
ImageCollection(paths[1], conserve_memory=False),
ImageCollection(os.pathsep.join(paths))
]
yield imgs
reset_plugins()
def test_custom_load_func_sequence(self):
filename = fetch('data/no_time_for_that_tiny.gif')
def reader(frameno):
vid = imageio.get_reader(filename)
return vid.get_data(frameno)
ic = ImageCollection(range(24), load_func=reader)
# the length of ic should be that of the given load_pattern sequence
assert len(ic) == 24
# GIF file has frames of size 25x14 with 4 channels (RGBA)
assert ic[0].shape == (25, 14, 4)
def test_custom_load_func_w_kwarg(self):
load_pattern = os.path.join(data_dir, 'no_time_for_that_tiny.gif')
def load_fn(f, step):
vid = imageio.get_reader(f)
seq = [v for v in vid.iter_data()]
return seq[::step]
ic = ImageCollection(load_pattern, load_func=load_fn, step=3)
# Each file should map to one image (array).
assert len(ic) == 1
# GIF file has 24 frames, so 24 / 3 equals 8.
assert len(ic[0]) == 8
def test_imagecollection_input():
"""Test function for ImageCollection. The new behavior (implemented
in 0.16) allows the `pattern` argument to accept a list of strings
as the input.
Notes
-----
If correct, `images` will receive three images.
"""
pattern = [
os.path.join(data_dir, pic) for pic in
['palette_gray.png', 'chessboard_GRAY_U16.tif', 'rocket.jpg']
]
images = ImageCollection(pattern)
assert len(images) == 3
def test_imagecollection_input():
"""Test function for ImageCollection. The new behavior (implemented
in 0.16) allows the `pattern` argument to accept a list of strings
as the input.
Notes
-----
If correct, `images` will receive three images.
"""
# Ensure that these images are part of the legacy datasets
# this means they will always be available in the user's install
# regarless of the availability of pooch
pattern = [
os.path.join(data_dir, pic)
for pic in ['coffee.png', 'chessboard_GRAY.png', 'rocket.jpg']
]
images = ImageCollection(pattern)
assert len(images) == 3
def get_sample_of_class(classs, labels, input_path, num=-1):
"""
Returns images sampled from files by class
"""
existing_files = pd.DataFrame(
[path.splitext(f)[0] for f in os.listdir(input_path)],
columns=[labels.columns[0]])
class_labels = pd.DataFrame(labels[labels['level'] == classs]['image'],
columns=[labels.columns[0]])
class_labels = class_labels.merge(existing_files,
on=class_labels.columns[0])
if num == -1:
num = class_labels.count()
files = [
path.join(raw_train, c + ".jpeg")
for c in class_labels[:num][class_labels.columns[0]]
]
return ImageCollection(files, conserve_memory=True)
def setUp(self):
# Generic image collection with images of different shapes.
self.images = ImageCollection(self.pattern)
# Image collection with images having shapes that match.
self.images_matched = ImageCollection(self.pattern_matched)
def get_images(path):
pattern = os.path.abspath(path) + os.sep + '*.jpg'
return ImageCollection(pattern, load_func=imread_convert)
