diff --git a/skluc/data/mldatasets/Dataset.py b/skluc/data/mldatasets/Dataset.py
index 5e7f7ca396219daf603da90833fef594a6c84387..5604a753f148dd010b0b7849feeefe591846d77c 100644
--- a/skluc/data/mldatasets/Dataset.py
+++ b/skluc/data/mldatasets/Dataset.py
@@ -1,7 +1,6 @@
import os
import numpy as np
-from sklearn.cross_validation import train_test_split
from sklearn.model_selection import StratifiedShuffleSplit
from sklearn.preprocessing import LabelBinarizer
diff --git a/skluc/data/mldatasets/ImageDataset.py b/skluc/data/mldatasets/ImageDataset.py
index a8180c095ca12cac8c987bd0bc743811d0bd6562..b8f6ee6d0a3f10cee9ba659785ec4b4514d5809d 100644
--- a/skluc/data/mldatasets/ImageDataset.py
+++ b/skluc/data/mldatasets/ImageDataset.py
@@ -46,7 +46,7 @@ class ImageDataset(Dataset):
# one need to apply it to the data
# then to save the transformation
logger.debug("Files {} don't exist or model md5 checksum doesn't match. Need to produce them".format(transform_filepaths))
- logger.info("Apply convolution of {} to dataset {}".format(transformer_name, self.s_name))
+ logger.info("Apply transformation of {} to dataset {}".format(transformer_name, self.s_name))
for kw in self.data_groups_private:
data, labels = getattr(self, kw)
transformed_data, transformed_labels = transformer.transform(data, labels)
diff --git a/skluc/data/transformation/ImageTransformer/RescaleTransformer.py b/skluc/data/transformation/RescaleTransformer.py
similarity index 70%
rename from skluc/data/transformation/ImageTransformer/RescaleTransformer.py
rename to skluc/data/transformation/RescaleTransformer.py
index f5f3ac64521e86610854caf562c97af853166ea0..f4f47909f01777e0f95b7fe3c8a8f3fb41dac468 100644
--- a/skluc/data/transformation/ImageTransformer/RescaleTransformer.py
+++ b/skluc/data/transformation/RescaleTransformer.py
@@ -1,4 +1,3 @@
-import os
import tensorflow as tf
import numpy as np
@@ -7,13 +6,13 @@ from skluc.utils import logger, Singleton
class RescaleTransformer(Transformer, metaclass=Singleton):
- def __init__(self, scaling_factor):
- self.rescale_factor = scaling_factor
- self.__name = os.path.join("resize", "{}".format(str(scaling_factor).replace(".", "-")))
+ def __init__(self, data_name, scaling_factor):
+ transformation_name = self.__class__.__name__ + "_" + "{}".format(str(scaling_factor).replace(".", "-"))
+
+ super().__init__(data_name=data_name,
+ transformation_name=transformation_name)
- @property
- def name(self):
- return self.__name
+ self.rescale_factor = scaling_factor
def transform(self, data, labels):
if len(data.shape) != 4:
@@ -25,10 +24,12 @@ class RescaleTransformer(Transformer, metaclass=Singleton):
sess = tf.InteractiveSession()
images_mat = data
- output_shape = np.multiply(images_mat.shape[1:-1], (self.rescale_factor, self.rescale_factor))
+ float_output_shape = np.multiply(images_mat.shape[1:-1], (self.rescale_factor, self.rescale_factor))
+ output_shape = float_output_shape.astype(np.int)
labels = labels
logger.debug("Expected output shape: {}".format((data.shape[0], *output_shape, data.shape[-1])))
new_images = tf.image.resize_images(images_mat, output_shape).eval()
logger.debug("Shape of data after rescaling: {}".format(new_images.shape))
sess.close()
+ tf.reset_default_graph()
return np.array(new_images), labels
\ No newline at end of file
diff --git a/skluc/data/transformation/ImageTransformer/ResizeTransformer.py b/skluc/data/transformation/ResizeTransformer.py
similarity index 67%
rename from skluc/data/transformation/ImageTransformer/ResizeTransformer.py
rename to skluc/data/transformation/ResizeTransformer.py
index 7be096fdbe3eefd26dedbde9fad1ba76fe19be77..02b9216231c6dfa48fbc4a9c2107139dc82a19d4 100644
--- a/skluc/data/transformation/ImageTransformer/ResizeTransformer.py
+++ b/skluc/data/transformation/ResizeTransformer.py
@@ -1,4 +1,3 @@
-import os
import tensorflow as tf
import numpy as np
@@ -7,15 +6,15 @@ from skluc.utils import logger, Singleton
class ResizeTransformer(Transformer, metaclass=Singleton):
- def __init__(self, output_shape):
+ def __init__(self, data_name, output_shape):
if len(output_shape) != 2:
raise AssertionError("Output shape should be 2D and it is {}D: {}".format(len(output_shape), output_shape))
- self.output_shape = output_shape
- self.__name = os.path.join("resize", "{}x{}".format(output_shape[0], output_shape[1]))
- @property
- def name(self):
- return self.__name
+ transformation_name = self.__class__.__name__ + "_" + "{}x{}".format(output_shape[0], output_shape[1])
+
+ super().__init__(data_name=data_name,
+ transformation_name=transformation_name)
+ self.output_shape = output_shape
def transform(self, data, labels):
if len(data.shape) != 4:
@@ -29,10 +28,10 @@ class ResizeTransformer(Transformer, metaclass=Singleton):
sess = tf.InteractiveSession()
images_mat = data
labels = labels
- lst_new_image = []
- for image_mat in images_mat:
- new_image = tf.image.resize_images(image_mat, self.output_shape).eval()
- lst_new_image.append(new_image)
- logger.debug("Shape data after resize: {}".format(np.array(lst_new_image).shape))
+
+ new_images = tf.image.resize_images(images_mat, self.output_shape).eval()
+
+ logger.debug("Shape data after resize: {}".format(new_images.shape))
sess.close()
- return np.array(lst_new_image), labels
+ tf.reset_default_graph()
+ return np.array(new_images), labels
diff --git a/skluc/test/test_transformation/TestRescaleTransformer.py b/skluc/test/test_transformation/TestRescaleTransformer.py
new file mode 100644
index 0000000000000000000000000000000000000000..8b01f4825c8c5fbaf0cba84ee250fdb2fd06a5c8
--- /dev/null
+++ b/skluc/test/test_transformation/TestRescaleTransformer.py
@@ -0,0 +1,47 @@
+import unittest
+
+from skluc.data.mldatasets import MnistDataset, Cifar10Dataset, Cifar100FineDataset, SVHNDataset
+from skluc.data.transformation.RescaleTransformer import RescaleTransformer
+from skluc.utils import logger
+
+
+class TestResizeTransformer(unittest.TestCase):
+ def setUp(self):
+ self.dict_datasets = {
+ "mnist": MnistDataset,
+ "cifar10": Cifar10Dataset,
+ "cifar100": Cifar100FineDataset,
+ "svhn": SVHNDataset
+ }
+ self.lst_scales = [
+ 0.5,
+ 0.7,
+ 1,
+ 2
+ ]
+
+ def test_transform(self):
+ valsize = 10000
+ for data_name in self.dict_datasets:
+ logger.info("Testing dataset {}".format(data_name))
+ for scale in self.lst_scales:
+ logger.info("Testing size {}".format(str(scale)))
+ dataset = self.dict_datasets[data_name]
+ d = dataset(validation_size=valsize)
+ d.load()
+ d.flatten()
+ d.to_image()
+ trans = RescaleTransformer(data_name=data_name, scaling_factor=scale)
+ d.apply_transformer(transformer=trans)
+ del trans
+
+ def test_init(self):
+ for data_name in self.dict_datasets:
+ for scale in self.lst_scales:
+ logger.info("Testing size {}".format(str(scale)))
+ trans = RescaleTransformer(data_name=data_name, scaling_factor=scale)
+ del trans
+
+
+if __name__ == '__main__':
+ unittest.main()
diff --git a/skluc/test/test_transformation/TestResizeTransformer.py b/skluc/test/test_transformation/TestResizeTransformer.py
new file mode 100644
index 0000000000000000000000000000000000000000..60ea6c000217edd2c4f26deec912feecdf4c95ad
--- /dev/null
+++ b/skluc/test/test_transformation/TestResizeTransformer.py
@@ -0,0 +1,47 @@
+import unittest
+
+from skluc.data.mldatasets import MnistDataset, Cifar10Dataset, Cifar100FineDataset, SVHNDataset
+from skluc.data.transformation.ResizeTransformer import ResizeTransformer
+from skluc.utils import logger
+
+
+class TestResizeTransformer(unittest.TestCase):
+ def setUp(self):
+ self.dict_datasets = {
+ "mnist": MnistDataset,
+ "cifar10": Cifar10Dataset,
+ "cifar100": Cifar100FineDataset,
+ "svhn": SVHNDataset
+ }
+ self.lst_sizes = [
+ (28, 32),
+ (32, 32),
+ (28, 28),
+ (32, 28)
+ ]
+
+ def test_transform(self):
+ valsize = 10000
+ for data_name in self.dict_datasets:
+ logger.info("Testing dataset {}".format(data_name))
+ for size in self.lst_sizes:
+ logger.info("Testing size {}".format(str(size)))
+ dataset = self.dict_datasets[data_name]
+ d = dataset(validation_size=valsize)
+ d.load()
+ d.flatten()
+ d.to_image()
+ trans = ResizeTransformer(data_name=data_name, output_shape=size)
+ d.apply_transformer(transformer=trans)
+ del trans
+
+ def test_init(self):
+ for data_name in self.dict_datasets:
+ for size in self.lst_sizes:
+ logger.info("Testing size {}".format(str(size)))
+ trans = ResizeTransformer(data_name=data_name, output_shape=size)
+ del trans
+
+
+if __name__ == '__main__':
+ unittest.main()