👕 Change character per line limit to 100 & remove explicit dependencies from pyproject

pyproject.toml

@@ -25,18 +25,16 @@ keywords = [
 ]
 classifiers = [
     "License :: OSI Approved :: Apache Software License",
+    "Programming Language :: Python :: 3 :: Only",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
-    "Programming Language :: Python :: 3 :: Only",
 ]
 dependencies = [
     "timm",
     "lightning[pytorch-extra]>=2.0",
     "torchvision>=0.16",
     "einops",
-    "matplotlib",
-    "rich>=10.2.2",
     "seaborn",
 ]
 
@@ -82,7 +80,7 @@ repository = "https://github.com/ENSTA-U2IS-AI/torch-uncertainty.git"
 name = "torch_uncertainty"
 
 [tool.ruff]
-line-length = 80
+line-length = 100
 target-version = "py310"
 lint.extend-select = [
     "A",

tests/_dummies/dataset.py

@@ -65,9 +65,7 @@ def __init__(
         else:
             self.data = self.data.transpose((0, 2, 3, 1))  # convert to HWC
 
-        self.targets = torch.randint(
-            low=0, high=num_classes, size=(num_images,)
-        )
+        self.targets = torch.randint(low=0, high=num_classes, size=(num_images,))
         self.targets = torch.arange(start=0, end=num_classes).repeat(
             num_images // (num_classes) + 1
         )[:num_images]
@@ -123,10 +121,8 @@ def __init__(
         self.targets = []
 
         input_shape = (num_samples, in_features)
-        if out_features != 1:
-            output_shape = (num_samples, out_features)
-        else:
-            output_shape = (num_samples,)
+
+        output_shape = (num_samples, out_features) if out_features != 1 else (num_samples,)
 
         self.data = torch.rand(
             size=input_shape,

tests/_dummies/model.py

@@ -60,9 +60,7 @@ def __init__(
         self.in_channels = in_channels
         self.num_classes = num_classes
         self.image_size = image_size
-        self.conv = nn.Conv2d(
-            in_channels, num_classes, kernel_size=3, padding=1
-        )
+        self.conv = nn.Conv2d(in_channels, num_classes, kernel_size=3, padding=1)
         self.dropout = nn.Dropout(p=dropout_rate)
         self.last_layer = last_layer
 

tests/baselines/test_deep_ensembles.py

@@ -9,9 +9,7 @@ class TestDeepEnsembles:
     """Testing the Deep Ensembles baseline class."""
 
     def test_failure(self):
-        with pytest.raises(
-            ValueError, match="Models must not be an empty list."
-        ):
+        with pytest.raises(ValueError, match="Models must not be an empty list."):
             DeepEnsemblesBaseline(
                 log_path=".",
                 checkpoint_ids=[],

tests/datamodules/classification/test_cifar10.py

@@ -72,9 +72,7 @@ def test_cifar10_main(self):
                 auto_augment="rand-m9-n2-mstd0.5",
             )
 
-        with pytest.raises(
-            ValueError, match="CIFAR-H can only be used in testing."
-        ):
+        with pytest.raises(ValueError, match="CIFAR-H can only be used in testing."):
             dm = CIFAR10DataModule(
                 root="./data/",
                 batch_size=128,
@@ -100,25 +98,21 @@ def test_cifar10_main(self):
 
     def test_cifar10_cv(self):
         dm = CIFAR10DataModule(root="./data/", batch_size=128)
-        dm.dataset = (
-            lambda root, train, download, transform: DummyClassificationDataset(
-                root,
-                train=train,
-                download=download,
-                transform=transform,
-                num_images=20,
-            )
+        dm.dataset = lambda root, train, download, transform: DummyClassificationDataset(
+            root,
+            train=train,
+            download=download,
+            transform=transform,
+            num_images=20,
         )
         dm.make_cross_val_splits(2, 1)
 
         dm = CIFAR10DataModule(root="./data/", batch_size=128, val_split=0.1)
-        dm.dataset = (
-            lambda root, train, download, transform: DummyClassificationDataset(
-                root,
-                train=train,
-                download=download,
-                transform=transform,
-                num_images=20,
-            )
+        dm.dataset = lambda root, train, download, transform: DummyClassificationDataset(
+            root,
+            train=train,
+            download=download,
+            transform=transform,
+            num_images=20,
         )
         dm.make_cross_val_splits(2, 1)

tests/datamodules/classification/test_cifar100.py

@@ -58,35 +58,27 @@ def test_cifar100(self):
                 randaugment=True,
             )
 
-        dm = CIFAR100DataModule(
-            root="./data/", batch_size=128, randaugment=True
-        )
+        dm = CIFAR100DataModule(root="./data/", batch_size=128, randaugment=True)
 
-        dm = CIFAR100DataModule(
-            root="./data/", batch_size=128, auto_augment="rand-m9-n2-mstd0.5"
-        )
+        dm = CIFAR100DataModule(root="./data/", batch_size=128, auto_augment="rand-m9-n2-mstd0.5")
 
     def test_cifar100_cv(self):
         dm = CIFAR100DataModule(root="./data/", batch_size=128)
-        dm.dataset = (
-            lambda root, train, download, transform: DummyClassificationDataset(
-                root,
-                train=train,
-                download=download,
-                transform=transform,
-                num_images=20,
-            )
+        dm.dataset = lambda root, train, download, transform: DummyClassificationDataset(
+            root,
+            train=train,
+            download=download,
+            transform=transform,
+            num_images=20,
         )
         dm.make_cross_val_splits(2, 1)
 
         dm = CIFAR100DataModule(root="./data/", batch_size=128, val_split=0.1)
-        dm.dataset = (
-            lambda root, train, download, transform: DummyClassificationDataset(
-                root,
-                train=train,
-                download=download,
-                transform=transform,
-                num_images=20,
-            )
+        dm.dataset = lambda root, train, download, transform: DummyClassificationDataset(
+            root,
+            train=train,
+            download=download,
+            transform=transform,
+            num_images=20,
         )
         dm.make_cross_val_splits(2, 1)

tests/datamodules/classification/test_imagenet.py

@@ -18,9 +18,7 @@ def test_imagenet(self):
         dm.prepare_data()
         dm.setup()
 
-        path = (
-            Path(__file__).parent.resolve() / "../../assets/dummy_indices.yaml"
-        )
+        path = Path(__file__).parent.resolve() / "../../assets/dummy_indices.yaml"
         dm = ImageNetDataModule(root="./data/", batch_size=128, val_split=path)
         dm.dataset = DummyClassificationDataset
         dm.ood_dataset = DummyClassificationDataset
@@ -55,22 +53,16 @@ def test_imagenet(self):
             dm.setup("other")
 
         for test_alt in ["r", "o", "a"]:
-            dm = ImageNetDataModule(
-                root="./data/", batch_size=128, test_alt=test_alt
-            )
+            dm = ImageNetDataModule(root="./data/", batch_size=128, test_alt=test_alt)
 
         with pytest.raises(ValueError):
             dm.setup()
 
         with pytest.raises(ValueError):
-            dm = ImageNetDataModule(
-                root="./data/", batch_size=128, test_alt="x"
-            )
+            dm = ImageNetDataModule(root="./data/", batch_size=128, test_alt="x")
 
         for ood_ds in ["inaturalist", "imagenet-o", "textures", "openimage-o"]:
-            dm = ImageNetDataModule(
-                root="./data/", batch_size=128, ood_ds=ood_ds
-            )
+            dm = ImageNetDataModule(root="./data/", batch_size=128, ood_ds=ood_ds)
             if ood_ds == "inaturalist":
                 dm.eval_ood = True
                 dm.dataset = DummyClassificationDataset
@@ -80,9 +72,7 @@ def test_imagenet(self):
                 dm.test_dataloader()
 
         with pytest.raises(ValueError):
-            dm = ImageNetDataModule(
-                root="./data/", batch_size=128, ood_ds="other"
-            )
+            dm = ImageNetDataModule(root="./data/", batch_size=128, ood_ds="other")
 
         for procedure in ["ViT", "A3"]:
             dm = ImageNetDataModule(
@@ -93,9 +83,7 @@ def test_imagenet(self):
             )
 
         with pytest.raises(ValueError):
-            dm = ImageNetDataModule(
-                root="./data/", batch_size=128, procedure="A2"
-            )
+            dm = ImageNetDataModule(root="./data/", batch_size=128, procedure="A2")
 
         with pytest.raises(FileNotFoundError):
             dm._verify_splits(split="test")

tests/datamodules/classification/test_tiny_imagenet.py

@@ -28,9 +28,7 @@ def test_tiny_imagenet(self):
         )
 
         with pytest.raises(ValueError):
-            TinyImageNetDataModule(
-                root="./data/", batch_size=128, ood_ds="other"
-            )
+            TinyImageNetDataModule(root="./data/", batch_size=128, ood_ds="other")
 
         dm.dataset = DummyClassificationDataset
         dm.ood_dataset = DummyClassificationDataset
@@ -52,9 +50,7 @@ def test_tiny_imagenet(self):
         dm.setup("test")
         dm.test_dataloader()
 
-        dm = TinyImageNetDataModule(
-            root="./data/", batch_size=128, ood_ds="svhn"
-        )
+        dm = TinyImageNetDataModule(root="./data/", batch_size=128, ood_ds="svhn")
         dm.dataset = DummyClassificationDataset
         dm.ood_dataset = DummyClassificationDataset
         dm.shift_dataset = DummyClassificationDataset
@@ -70,9 +66,7 @@ def test_tiny_imagenet_cv(self):
         )
         dm.make_cross_val_splits(2, 1)
 
-        dm = TinyImageNetDataModule(
-            root="./data/", batch_size=128, val_split=0.1
-        )
+        dm = TinyImageNetDataModule(root="./data/", batch_size=128, val_split=0.1)
         dm.dataset = lambda root, split, transform: DummyClassificationDataset(
             root, split=split, transform=transform, num_images=20
         )

tests/datamodules/segmentation/test_camvid.py

@@ -9,12 +9,8 @@ class TestCamVidDataModule:
     """Testing the CamVidDataModule datamodule."""
 
     def test_camvid_main(self):
-        dm = CamVidDataModule(
-            root="./data/", batch_size=128, group_classes=False
-        )
-        dm = CamVidDataModule(
-            root="./data/", batch_size=128, basic_augment=False
-        )
+        dm = CamVidDataModule(root="./data/", batch_size=128, group_classes=False)
+        dm = CamVidDataModule(root="./data/", batch_size=128, basic_augment=False)
 
         assert dm.dataset == CamVid
 

tests/datamodules/segmentation/test_cityscapes.py

@@ -10,9 +10,7 @@ class TestCityscapesDataModule:
 
     def test_camvid_main(self):
         dm = CityscapesDataModule(root="./data/", batch_size=128)
-        dm = CityscapesDataModule(
-            root="./data/", batch_size=128, basic_augment=False
-        )
+        dm = CityscapesDataModule(root="./data/", batch_size=128, basic_augment=False)
 
         assert dm.dataset == Cityscapes
 

tests/datamodules/test_abstract_datamodule.py

@@ -31,9 +31,7 @@ def test_cv_main(self):
         dm.train = ds
         dm.val = ds
         dm.test = ds
-        cv_dm = CrossValDataModule(
-            "root", [0], [1], dm, 128, 0.0, 4, True, True
-        )
+        cv_dm = CrossValDataModule("root", [0], [1], dm, 128, 0.0, 4, True, True)
 
         cv_dm.setup()
         cv_dm.setup("test")
@@ -54,9 +52,7 @@ def test_errors(self):
         dm.train = ds
         dm.val = ds
         dm.test = ds
-        cv_dm = CrossValDataModule(
-            "root", [0], [1], dm, 128, 0.0, 4, True, True
-        )
+        cv_dm = CrossValDataModule("root", [0], [1], dm, 128, 0.0, 4, True, True)
         with pytest.raises(NotImplementedError):
             cv_dm.setup()
             cv_dm._get_train_data()

tests/datamodules/test_depth.py

@@ -13,9 +13,7 @@ class TestMUADDataModule:
     """Testing the MUADDataModule datamodule."""
 
     def test_muad_main(self):
-        dm = MUADDataModule(
-            root="./data/", min_depth=0, max_depth=100, batch_size=128
-        )
+        dm = MUADDataModule(root="./data/", min_depth=0, max_depth=100, batch_size=128)
 
         assert dm.dataset == MUAD
 

tests/datamodules/test_uci_regression.py

@@ -8,9 +8,7 @@ class TestUCIRegressionDataModule:
     """Testing the UCIRegressionDataModule datamodule class."""
 
     def test_uci_regression(self):
-        dm = UCIRegressionDataModule(
-            dataset_name="kin8nm", root="./data/", batch_size=128
-        )
+        dm = UCIRegressionDataModule(dataset_name="kin8nm", root="./data/", batch_size=128)
 
         dm.dataset = partial(DummyRegressionDataset, num_samples=64)
         dm.prepare_data()

tests/layers/test_bayesian.py

@@ -80,9 +80,7 @@ def test_conv1(self, feat_input_odd: torch.Tensor) -> None:
         assert out.shape == torch.Size([2, 10])
 
     def test_conv1_even(self, feat_input_even: torch.Tensor) -> None:
-        layer = BayesConv1d(
-            8, 2, kernel_size=1, sigma_init=0, padding_mode="reflect"
-        )
+        layer = BayesConv1d(8, 2, kernel_size=1, sigma_init=0, padding_mode="reflect")
         print(layer)
         out = layer(feat_input_even)
         assert out.shape == torch.Size([2, 10])
@@ -94,9 +92,7 @@ def test_conv1_even(self, feat_input_even: torch.Tensor) -> None:
 
     def test_error(self):
         with pytest.raises(ValueError):
-            BayesConv1d(
-                8, 2, kernel_size=1, sigma_init=0, padding_mode="random"
-            )
+            BayesConv1d(8, 2, kernel_size=1, sigma_init=0, padding_mode="random")
 
 
 class TestBayesConv2d:
@@ -115,9 +111,7 @@ def test_conv2(self, img_input_odd: torch.Tensor) -> None:
         layer.sample()
 
     def test_conv2_even(self, img_input_even: torch.Tensor) -> None:
-        layer = BayesConv2d(
-            10, 2, kernel_size=1, sigma_init=0, padding_mode="reflect"
-        )
+        layer = BayesConv2d(10, 2, kernel_size=1, sigma_init=0, padding_mode="reflect")
         print(layer)
         out = layer(img_input_even)
         assert out.shape == torch.Size([8, 2, 3, 3])
@@ -141,9 +135,7 @@ def test_conv3(self, cube_input_odd: torch.Tensor) -> None:
         assert out.shape == torch.Size([1, 2, 3, 3, 3])
 
     def test_conv3_even(self, cube_input_even: torch.Tensor) -> None:
-        layer = BayesConv3d(
-            10, 2, kernel_size=1, sigma_init=0, padding_mode="reflect"
-        )
+        layer = BayesConv3d(10, 2, kernel_size=1, sigma_init=0, padding_mode="reflect")
         print(layer)
         out = layer(cube_input_even)
         assert out.shape == torch.Size([2, 2, 3, 3, 3])
@@ -192,17 +184,13 @@ def test_conv2(self, img_input_odd: torch.Tensor) -> None:
         out = layer(img_input_odd.repeat(4, 1, 1, 1))
         assert out.shape == torch.Size([5 * 4, 2, 3, 3])
 
-        layer = LPBNNConv2d(
-            10, 2, kernel_size=1, num_estimators=4, bias=False, gamma=False
-        )
+        layer = LPBNNConv2d(10, 2, kernel_size=1, num_estimators=4, bias=False, gamma=False)
         layer = layer.eval()
         out = layer(img_input_odd.repeat(4, 1, 1, 1))
         assert out.shape == torch.Size([5 * 4, 2, 3, 3])
 
     def test_conv2_even(self, img_input_even: torch.Tensor) -> None:
-        layer = LPBNNConv2d(
-            10, 2, kernel_size=1, num_estimators=4, padding_mode="reflect"
-        )
+        layer = LPBNNConv2d(10, 2, kernel_size=1, num_estimators=4, padding_mode="reflect")
         print(layer)
         out = layer(img_input_even.repeat(4, 1, 1, 1))
         assert out.shape == torch.Size([8 * 4, 2, 3, 3])