fix: pep8 and flake check.

data/IXI_Guys/download_IXI_Guys.py

@@ -32,9 +32,9 @@ def resample_image(itk_image, out_spacing=[1.0, 1.0, 1.0], is_label=False):
     original_size = itk_image.GetSize()
 
     out_size = [
-        int(np.round(original_size[0] * (original_spacing[0]/out_spacing[0]))),
-        int(np.round(original_size[1] * (original_spacing[1]/out_spacing[1]))),
-        int(np.round(original_size[2] * (original_spacing[2]/out_spacing[2])))
+        int(np.round(original_size[0] * (original_spacing[0] / out_spacing[0]))),
+        int(np.round(original_size[1] * (original_spacing[1] / out_spacing[1]))),
+        int(np.round(original_size[2] * (original_spacing[2] / out_spacing[2])))
     ]
 
     resample = sitk.ResampleImageFilter()

data/IXI_HH/download_IXI_HH.py

@@ -32,9 +32,9 @@ def resample_image(itk_image, out_spacing=(1.0, 1.0, 1.0), is_label=False):
     original_spacing = itk_image.GetSpacing()
     original_size = itk_image.GetSize()
 
-    out_size = [int(np.round(original_size[0]*(original_spacing[0]/out_spacing[0]))),
-                int(np.round(original_size[1]*(original_spacing[1]/out_spacing[1]))),
-                int(np.round(original_size[2]*(original_spacing[2]/out_spacing[2])))]
+    out_size = [int(np.round(original_size[0] * (original_spacing[0] / out_spacing[0]))),
+                int(np.round(original_size[1] * (original_spacing[1] / out_spacing[1]))),
+                int(np.round(original_size[2] * (original_spacing[2] / out_spacing[2])))]
 
     resample = sitk.ResampleImageFilter()
     resample.SetOutputSpacing(out_spacing)

dltk/core/losses.py

@@ -35,11 +35,9 @@ def sparse_balanced_crossentropy(logits, labels):
         labels, minlength=num_classes, dtype=tf.float32))
 
     weights = (1. / (class_frequencies + tf.constant(1e-8)))
-    weights *= (tf.cast(tf.reduce_prod(tf.shape(labels)), tf.float32)
-                / tf.cast(num_classes, tf.float32))
+    weights *= (tf.cast(tf.reduce_prod(tf.shape(labels)), tf.float32) / tf.cast(num_classes, tf.float32))
 
-    new_shape = (([1, ] * len(labels.get_shape().as_list()))
-                 + [logits.get_shape().as_list()[-1]])
+    new_shape = (([1, ] * len(labels.get_shape().as_list())) + [logits.get_shape().as_list()[-1]])
 
     weights = tf.reshape(weights, new_shape)
 

dltk/io/abstract_reader.py

@@ -80,23 +80,17 @@ def clean_ex(ex, compare):
                     for k in list(ex.keys()):
                         if k not in list(compare.keys()):
                             del ex[k]
-                        elif isinstance(ex[k], dict) \
-                                and isinstance(compare[k], dict):
+                        elif isinstance(ex[k], dict) and isinstance(compare[k], dict):
                             clean_ex(ex[k], compare[k])
-                        elif (isinstance(ex[k], dict)
-                              and not isinstance(compare[k], dict)) \
-                                or (not isinstance(ex[k], dict)
-                                    and isinstance(compare[k], dict)):
+                        elif (isinstance(ex[k], dict) and not isinstance(compare[k], dict)) or \
+                             (not isinstance(ex[k], dict) and isinstance(compare[k], dict)):
                             raise ValueError('Entries between example and '
                                              'dtypes incompatible for key {}'
                                              ''.format(k))
-                        elif ((isinstance(ex[k], list)
-                               and not isinstance(compare[k], list))
-                              or (not isinstance(ex[k], list)
-                                  and isinstance(compare[k], list))
-                              or (isinstance(ex[k], list)
-                                  and isinstance(compare[k], list)
-                                  and not len(ex[k]) == len(compare[k]))):
+                        elif ((isinstance(ex[k], list) and not isinstance(compare[k], list)) or
+                              (not isinstance(ex[k], list) and isinstance(compare[k], list)) or
+                              (isinstance(ex[k], list) and isinstance(compare[k], list) and
+                               not len(ex[k]) == len(compare[k]))):
                             raise ValueError('Entries between example and '
                                              'dtypes incompatible for key {}'
                                              ''.format(k))
@@ -165,8 +159,7 @@ def serving_input_receiver_fn(self, placeholder_shapes):
         def f():
             inputs = {k: tf.placeholder(
                 shape=[None] + list(placeholder_shapes['features'][k]),
-                dtype=self.dtypes['features'][k])
-                      for k in list(self.dtypes['features'].keys())}
+                dtype=self.dtypes['features'][k]) for k in list(self.dtypes['features'].keys())}
 
             return tf.estimator.export.ServingInputReceiver(inputs, inputs)
         return f

dltk/io/augmentation.py

@@ -196,9 +196,9 @@ def extract_class_balanced_example_array(image,
 
         for i in range(len(r_idx)):
             # Extract class-balanced examples from the original image
-            slicer = [slice(r_idx[i][dim]
-                            - ex_rad[dim][0], r_idx[i][dim]
-                            + ex_rad[dim][1]) for dim in range(rank)]
+            slicer = [slice(r_idx[i][dim] -
+                            ex_rad[dim][0], r_idx[i][dim] +
+                            ex_rad[dim][1]) for dim in range(rank)]
 
             ex_image = image[slicer][np.newaxis, :]
 
@@ -252,15 +252,15 @@ def extract_random_example_array(image_list,
 
     assert all([i_s >= e_s for i_s, e_s in zip(image_list[0].shape, example_size)]), \
         'Image must be bigger than example shape'
-    assert (image_list[0].ndim - 1 == len(example_size)
-            or image_list[0].ndim == len(example_size)), \
+    assert (image_list[0].ndim - 1 == len(example_size) or
+            image_list[0].ndim == len(example_size)), \
         'Example size doesnt fit image size'
 
     for i in image_list:
         if len(image_list) > 1:
-            assert (i.ndim - 1 == image_list[0].ndim
-                    or i.ndim == image_list[0].ndim
-                    or i.ndim + 1 == image_list[0].ndim), \
+            assert (i.ndim - 1 == image_list[0].ndim or
+                    i.ndim == image_list[0].ndim or
+                    i.ndim + 1 == image_list[0].ndim), \
                 'Example size doesnt fit image size'
 
             assert all([i0_s == i_s for i0_s, i_s in zip(image_list[0].shape, i.shape)]), \

dltk/networks/gan/dcgan.py

@@ -74,8 +74,7 @@ def dcgan_generator_3d(inputs,
             tf.logging.info('Generator at res_scale after up {} tensor '
                             'shape: {}'.format(res_scale, x.get_shape()))
 
-            x = tf.layers.batch_normalization(
-               x, training=mode == tf.estimator.ModeKeys.TRAIN)
+            x = tf.layers.batch_normalization(x, training=mode == tf.estimator.ModeKeys.TRAIN)
 
             x = leaky_relu(x, 0.2)
             tf.logging.info('Generator at res_scale {} tensor shape: '

dltk/networks/segmentation/deepmedic.py

@@ -151,8 +151,8 @@ def _residual_connection(x, prev_x):
         # add prev_x to first channels of x
 
         to_pad = [[0, 0]] * (len(x.get_shape().as_list()) - 1)
-        to_pad += [[0, x.get_shape().as_list()[-1]
-                    - prev_x.get_shape().as_list()[-1]]]
+        to_pad += [[0, x.get_shape().as_list()[-1] -
+                    prev_x.get_shape().as_list()[-1]]]
         prev_x = tf.pad(prev_x, to_pad)
 
         return x + prev_x

dltk/utils.py

@@ -115,8 +115,7 @@ def sliding_window_segmentation_inference(session,
 
     out_dummies = [np.zeros(
         [inp_shape[0], ] + inp_bshape + [op.get_shape().as_list()[-1]]
-        if len(op.get_shape().as_list()) == len(inp_shape) else [])
-                   for op in ops_list]
+        if len(op.get_shape().as_list()) == len(inp_shape) else []) for op in ops_list]
 
     out_dummy_counter = [np.zeros_like(o) for o in out_dummies]
 
@@ -162,8 +161,7 @@ def sliding_window_segmentation_inference(session,
             out_slicers.append(out_slicer)
             if len(slicers) == batch_size or done:
                 slices_dict = {k: np.concatenate(
-                    [v[slicer] for slicer in slicers], 0)
-                               for k, v in padded_dict.items()}
+                    [v[slicer] for slicer in slicers], 0) for k, v in padded_dict.items()}
 
                 all_op_parts = session.run(ops_list, feed_dict=slices_dict)
 

docs/source/conf.py

@@ -126,8 +126,7 @@
 # so a file named "default.css" will overwrite the builtin "default.css".
 html_static_path = ['_static']
 
-html_context = {'css_files': [
-            '_static/theme_overrides.css']}
+html_context = {'css_files': ['_static/theme_overrides.css']}
 
 # -- Options for HTMLHelp output ------------------------------------------