def test_put_probmap_data_when_no_labels_available(self):
img_path = os.path.abspath(
os.path.join(base_path, '../test_data/tiffconnector_1/im/*'))
c = io_connector(img_path, '', savepath=self.tmpdir)
d = Dataset(c)
size = (1, 3, 4)
batch_size = 2
p = PredictionBatch(d, batch_size, size)
data = np.ones((2, 2, 1, 3, 4))
p[0].put_probmap_data(data)
data = np.ones((2, 2, 1, 3, 4))
p[1].put_probmap_data(data)
data = np.ones((2, 2, 1, 3, 4))
p[2].put_probmap_data(data)
val = [
'40width26height3slices_rgb_class_1.tif',
'40width26height3slices_rgb_class_2.tif'
]
self.assertEqual(sorted(os.listdir(self.tmpdir)), val)
def load_training_data(self, image_path, label_path, batch_size=None):
'''
Connect to a training dataset.
Parameters
----------
image_path : string
Path to folder with tiff images.
label_path : string
Path to folder with label tiff images or path to ilastik project
file (.ilp file).
'''
_connector = io_connector(image_path, label_path)
self.lbl_map = _connector.labelvalue_mapping[0]
self.dataset = Dataset(_connector)
if batch_size is None:
self.batch_size = len(self.dataset.label_values())
else:
self.batch_size = batch_size
msg = '\n\nImport taining dataset:\n{}\n'.format(
self.dataset.pixel_connector.__repr__())
sys.stdout.write(msg)
def load_prediction_data(self, image_path, save_path):
'''
Connect to a prediction dataset.
Parameters
----------
image_path : string
Path to folder with tiff images to predict.
save_path : string
Path to folder for saving prediction images.
'''
self.dataset = Dataset(
io_connector(image_path,
'/tmp/this_should_not_exist',
savepath=save_path))
def test_load_label_counts_from_ilastik(self):
img_path = os.path.join(base_path, '../test_data/ilastik')
lbl_path = os.path.join(base_path,
'../test_data/ilastik/ilastik-1.2.ilp')
c = io_connector(img_path, lbl_path)
d = Dataset(c)
actual_counts = c.label_count_for_image(0)
print(actual_counts)
label_counts = d.load_label_counts()
print(label_counts)
assert_array_equal(label_counts[1], np.array([1]))
assert_array_equal(label_counts[2], np.array([1]))
assert_array_equal(label_counts[3], np.array([1]))
def test_pixel_dimensions(self):
img_path = os.path.abspath(
os.path.join(base_path, '../test_data/tiffconnector_1/im/*'))
c = io_connector(img_path, '', savepath=self.tmpdir)
d = Dataset(c)
size = (1, 5, 4)
batch_size = 2
p = PredictionBatch(d, batch_size, size)[0]
print(p.pixels().shape)
self.assertEqual((2, 3, 1, 5, 4), p.pixels().shape)
p.set_pixel_dimension_order('bzxyc')
self.assertEqual((2, 1, 5, 4, 3), p.pixels().shape)
def load_training_data(self, image_path, label_path):
'''
Connect to a training dataset.
Parameters
----------
image_path : string
Path to folder with tiff images.
label_path : string
Path to folder with label tiff images or path to ilastik project
file (.ilp file).
'''
print(image_path)
print(label_path)
self.dataset = Dataset(io_connector(image_path, label_path))
def load_prediction_data(self, image_path, save_path, batch_size=None):
'''
Connect to a prediction dataset.
Parameters
----------
image_path : string
Path to folder with tiff images to predict.
save_path : string
Path to folder for saving prediction images.
'''
self.dataset = Dataset(
io_connector(image_path,
'/tmp/this_should_not_exist',
savepath=save_path))
if batch_size is None:
self.batch_size = 1
else:
self.batch_size = batch_size
msg = '\n\nImport dataset for prediction:\n{}\n'.format(
self.dataset.pixel_connector.__repr__())
sys.stdout.write(msg)