refactor code style in DCA

DCA/adapter.py

@@ -30,12 +30,7 @@ class DcaAdapter(AdapterBase):
     Adapter to work with DCA algorithm
     """
 
-    def __init__(
-            self,
-            factor: int,
-            model: Path,
-            device: str
-    ):
+    def __init__(self, factor: int, model: Path, device: str):
         """
         Args:
             factor (int): factor to divide image into patches
@@ -57,37 +52,39 @@ def process(self, image: np.ndarray):
     def _transform_image(self, image: np.ndarray):
         mean = np.mean(image, axis=(0, 1))
         std = np.std(image, axis=(0, 1))
-        transformer = Compose([
-            Normalize(mean=mean,
-                      std=std,
-                      max_pixel_value=1, always_apply=True),
-            ever.preprocess.albu.ToTensor()
-        ], is_check_shapes=False)
+        transformer = Compose(
+            [
+                Normalize(mean=mean, std=std, max_pixel_value=1, always_apply=True),
+                ever.preprocess.albu.ToTensor(),
+            ],
+            is_check_shapes=False,
+        )
         blob = transformer(image=image)
         image = blob["image"].to(self._device)
         image = image[None, :]
         return image
 
     def _build_model(self):
-        model = Deeplabv2(dict(
-            backbone=dict(
-                resnet_type="resnet50",
-                output_stride=16,
-                pretrained=True,
-            ),
-            multi_layer=True,
-            cascade=False,
-            use_ppm=True,
-            ppm=dict(
+        model = Deeplabv2(
+            dict(
+                backbone=dict(
+                    resnet_type="resnet50",
+                    output_stride=16,
+                    pretrained=True,
+                ),
+                multi_layer=True,
+                cascade=False,
+                use_ppm=True,
+                ppm=dict(
+                    num_classes=7,
+                    use_aux=False,
+                    fc_dim=2048,
+                ),
+                inchannels=2048,
                 num_classes=7,
-                use_aux=False,
-                fc_dim=2048,
-            ),
-            inchannels=2048,
-            num_classes=7
-        )).to(self._device)
-        model_state_dict = torch.load(self._model,
-                                      map_location=self._device)
+            )
+        ).to(self._device)
+        model_state_dict = torch.load(self._model, map_location=self._device)
         model.load_state_dict(model_state_dict, strict=True)
         model.eval()
         return model
@@ -99,11 +96,20 @@ def _process(self, model, image):
                 model=model,
                 image=image,
                 num_classes=7,
-                tile_size=(image.shape[2] // self._factor, image.shape[3] // self._factor),
+                tile_size=(
+                    image.shape[2] // self._factor,
+                    image.shape[3] // self._factor,
+                ),
                 tta=True,
-                device=self._device
+                device=self._device,
+            )
+            cls = (
+                cls.argmax(dim=1)
+                .to(self._device)
+                .numpy()
+                .reshape(shape)
+                .astype(np.uint8)
             )
-            cls = cls.argmax(dim=1).to(self._device).numpy().reshape(shape).astype(np.uint8)
             return cls
 
     def _postprocess_predictions(self, raw_predictions):

DCA/run.py

@@ -44,6 +44,6 @@
         image = imread(file_path)
         logging.info(f"Image shape: {image.shape}")
         mask = adapter.process(image)
-        output_path = output_dir / Path(file_path.stem).with_suffix('.npy')
+        output_path = output_dir / Path(file_path.stem).with_suffix(".npy")
         logging.info(f" Saving to {output_path}")
-        np.save(output_path, mask)
+        np.save(output_path, mask)

DCA/utils.py

@@ -18,15 +18,40 @@
 
 def get_args():
     parser = ArgumentParser()
-    parser.add_argument("-f", "--factor", type=int, default=2,
-                        help="Factor shows how images must be scaled to create patches, for factor = n there will be "
-                             "n^2 patches (default: 2)")
-    parser.add_argument("-m", "--model", type=Path, default=Path("/", "DCA", "weights", "Urban.pth"),
-                        help="Pretrained model path (default: weights/Urban.pth)")
-    parser.add_argument("-d", "--device", type=str, default="cuda",
-                        help="Which device to run network on (default: GPU)")
-    parser.add_argument("-i", "--input", type=Path, default=Path("/", "DCA", "input"),
-                        help="Images input folder")
-    parser.add_argument("-o", "--output", type=Path, default=Path("/", "DCA", "output"),
-                        help="Masks output folder")
+    parser.add_argument(
+        "-f",
+        "--factor",
+        type=int,
+        default=2,
+        help="Factor shows how images must be scaled to create patches, for factor = n there will be "
+        "n^2 patches (default: 2)",
+    )
+    parser.add_argument(
+        "-m",
+        "--model",
+        type=Path,
+        default=Path("/", "DCA", "weights", "Urban.pth"),
+        help="Pretrained model path (default: weights/Urban.pth)",
+    )
+    parser.add_argument(
+        "-d",
+        "--device",
+        type=str,
+        default="cuda",
+        help="Which device to run network on (default: GPU)",
+    )
+    parser.add_argument(
+        "-i",
+        "--input",
+        type=Path,
+        default=Path("/", "DCA", "input"),
+        help="Images input folder",
+    )
+    parser.add_argument(
+        "-o",
+        "--output",
+        type=Path,
+        default=Path("/", "DCA", "output"),
+        help="Masks output folder",
+    )
     return parser.parse_args()