Moved indian configs to data

pyproject.toml

@@ -140,7 +140,7 @@ shenzhen_f7_rgb = "ptbench.data.shenzhen.fold_7_rgb"
 shenzhen_f8_rgb = "ptbench.data.shenzhen.fold_8_rgb"
 shenzhen_f9_rgb = "ptbench.data.shenzhen.fold_9_rgb"
 # extended shenzhen dataset (with radiological signs)
-shenzhen_rs = "ptbench.configs.datasets.shenzhen_RS.default"
+shenzhen_rs = "ptbench.data.shenzhen_RS.default"
 shenzhen_rs_f0 = "ptbench.configs.datasets.shenzhen_RS.fold_0"
 shenzhen_rs_f1 = "ptbench.configs.datasets.shenzhen_RS.fold_1"
 shenzhen_rs_f2 = "ptbench.configs.datasets.shenzhen_RS.fold_2"
@@ -153,27 +153,27 @@ shenzhen_rs_f8 = "ptbench.configs.datasets.shenzhen_RS.fold_8"
 shenzhen_rs_f9 = "ptbench.configs.datasets.shenzhen_RS.fold_9"
 # indian dataset (and cross-validation folds)
 indian = "ptbench.data.indian.default"
-indian_rgb = "ptbench.configs.datasets.indian.rgb"
-indian_f0 = "ptbench.configs.datasets.indian.fold_0"
-indian_f1 = "ptbench.configs.datasets.indian.fold_1"
-indian_f2 = "ptbench.configs.datasets.indian.fold_2"
-indian_f3 = "ptbench.configs.datasets.indian.fold_3"
-indian_f4 = "ptbench.configs.datasets.indian.fold_4"
-indian_f5 = "ptbench.configs.datasets.indian.fold_5"
-indian_f6 = "ptbench.configs.datasets.indian.fold_6"
-indian_f7 = "ptbench.configs.datasets.indian.fold_7"
-indian_f8 = "ptbench.configs.datasets.indian.fold_8"
-indian_f9 = "ptbench.configs.datasets.indian.fold_9"
-indian_f0_rgb = "ptbench.configs.datasets.indian.fold_0_rgb"
-indian_f1_rgb = "ptbench.configs.datasets.indian.fold_1_rgb"
-indian_f2_rgb = "ptbench.configs.datasets.indian.fold_2_rgb"
-indian_f3_rgb = "ptbench.configs.datasets.indian.fold_3_rgb"
-indian_f4_rgb = "ptbench.configs.datasets.indian.fold_4_rgb"
-indian_f5_rgb = "ptbench.configs.datasets.indian.fold_5_rgb"
-indian_f6_rgb = "ptbench.configs.datasets.indian.fold_6_rgb"
-indian_f7_rgb = "ptbench.configs.datasets.indian.fold_7_rgb"
-indian_f8_rgb = "ptbench.configs.datasets.indian.fold_8_rgb"
-indian_f9_rgb = "ptbench.configs.datasets.indian.fold_9_rgb"
+indian_rgb = "ptbench.data.indian.rgb"
+indian_f0 = "ptbench.data.indian.fold_0"
+indian_f1 = "ptbench.data.indian.fold_1"
+indian_f2 = "ptbench.data.indian.fold_2"
+indian_f3 = "ptbench.data.indian.fold_3"
+indian_f4 = "ptbench.data.indian.fold_4"
+indian_f5 = "ptbench.data.indian.fold_5"
+indian_f6 = "ptbench.data.indian.fold_6"
+indian_f7 = "ptbench.data.indian.fold_7"
+indian_f8 = "ptbench.data.indian.fold_8"
+indian_f9 = "ptbench.data.indian.fold_9"
+indian_f0_rgb = "ptbench.data.indian.fold_0_rgb"
+indian_f1_rgb = "ptbench.data.indian.fold_1_rgb"
+indian_f2_rgb = "ptbench.data.indian.fold_2_rgb"
+indian_f3_rgb = "ptbench.data.indian.fold_3_rgb"
+indian_f4_rgb = "ptbench.data.indian.fold_4_rgb"
+indian_f5_rgb = "ptbench.data.indian.fold_5_rgb"
+indian_f6_rgb = "ptbench.data.indian.fold_6_rgb"
+indian_f7_rgb = "ptbench.data.indian.fold_7_rgb"
+indian_f8_rgb = "ptbench.data.indian.fold_8_rgb"
+indian_f9_rgb = "ptbench.data.indian.fold_9_rgb"
 # extended indian dataset (with radiological signs)
 indian_rs = "ptbench.configs.datasets.indian_RS.default"
 indian_rs_f0 = "ptbench.configs.datasets.indian_RS.fold_0"
@@ -187,28 +187,28 @@ indian_rs_f7 = "ptbench.configs.datasets.indian_RS.fold_7"
 indian_rs_f8 = "ptbench.configs.datasets.indian_RS.fold_8"
 indian_rs_f9 = "ptbench.configs.datasets.indian_RS.fold_9"
 # TBX11K simplified dataset split 1 (and cross-validation folds)
-tbx11k_simplified = "ptbench.configs.datasets.tbx11k_simplified.default"
-tbx11k_simplified_rgb = "ptbench.configs.datasets.tbx11k_simplified.rgb"
-tbx11k_simplified_f0 = "ptbench.configs.datasets.tbx11k_simplified.fold_0"
-tbx11k_simplified_f1 = "ptbench.configs.datasets.tbx11k_simplified.fold_1"
-tbx11k_simplified_f2 = "ptbench.configs.datasets.tbx11k_simplified.fold_2"
-tbx11k_simplified_f3 = "ptbench.configs.datasets.tbx11k_simplified.fold_3"
-tbx11k_simplified_f4 = "ptbench.configs.datasets.tbx11k_simplified.fold_4"
-tbx11k_simplified_f5 = "ptbench.configs.datasets.tbx11k_simplified.fold_5"
-tbx11k_simplified_f6 = "ptbench.configs.datasets.tbx11k_simplified.fold_6"
-tbx11k_simplified_f7 = "ptbench.configs.datasets.tbx11k_simplified.fold_7"
-tbx11k_simplified_f8 = "ptbench.configs.datasets.tbx11k_simplified.fold_8"
-tbx11k_simplified_f9 = "ptbench.configs.datasets.tbx11k_simplified.fold_9"
-tbx11k_simplified_f0_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_0_rgb"
-tbx11k_simplified_f1_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_1_rgb"
-tbx11k_simplified_f2_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_2_rgb"
-tbx11k_simplified_f3_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_3_rgb"
-tbx11k_simplified_f4_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_4_rgb"
-tbx11k_simplified_f5_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_5_rgb"
-tbx11k_simplified_f6_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_6_rgb"
-tbx11k_simplified_f7_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_7_rgb"
-tbx11k_simplified_f8_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_8_rgb"
-tbx11k_simplified_f9_rgb = "ptbench.configs.datasets.tbx11k_simplified.fold_9_rgb"
+tbx11k_simplified = "ptbench.data.tbx11k_simplified.default"
+tbx11k_simplified_rgb = "ptbench.data.tbx11k_simplified.rgb"
+tbx11k_simplified_f0 = "ptbench.data.tbx11k_simplified.fold_0"
+tbx11k_simplified_f1 = "ptbench.data.tbx11k_simplified.fold_1"
+tbx11k_simplified_f2 = "ptbench.data.tbx11k_simplified.fold_2"
+tbx11k_simplified_f3 = "ptbench.data.tbx11k_simplified.fold_3"
+tbx11k_simplified_f4 = "ptbench.data.tbx11k_simplified.fold_4"
+tbx11k_simplified_f5 = "ptbench.data.tbx11k_simplified.fold_5"
+tbx11k_simplified_f6 = "ptbench.data.tbx11k_simplified.fold_6"
+tbx11k_simplified_f7 = "ptbench.data.tbx11k_simplified.fold_7"
+tbx11k_simplified_f8 = "ptbench.data.tbx11k_simplified.fold_8"
+tbx11k_simplified_f9 = "ptbench.data.tbx11k_simplified.fold_9"
+tbx11k_simplified_f0_rgb = "ptbench.data.tbx11k_simplified.fold_0_rgb"
+tbx11k_simplified_f1_rgb = "ptbench.data.tbx11k_simplified.fold_1_rgb"
+tbx11k_simplified_f2_rgb = "ptbench.data.tbx11k_simplified.fold_2_rgb"
+tbx11k_simplified_f3_rgb = "ptbench.data.tbx11k_simplified.fold_3_rgb"
+tbx11k_simplified_f4_rgb = "ptbench.data.tbx11k_simplified.fold_4_rgb"
+tbx11k_simplified_f5_rgb = "ptbench.data.tbx11k_simplified.fold_5_rgb"
+tbx11k_simplified_f6_rgb = "ptbench.data.tbx11k_simplified.fold_6_rgb"
+tbx11k_simplified_f7_rgb = "ptbench.data.tbx11k_simplified.fold_7_rgb"
+tbx11k_simplified_f8_rgb = "ptbench.data.tbx11k_simplified.fold_8_rgb"
+tbx11k_simplified_f9_rgb = "ptbench.data.tbx11k_simplified.fold_9_rgb"
 # extended TBX11K simplified dataset split 1 (with radiological signs)
 tbx11k_simplified_rs = "ptbench.configs.datasets.tbx11k_simplified_RS.default"
 tbx11k_simplified_rs_f0 = "ptbench.configs.datasets.tbx11k_simplified_RS.fold_0"

src/ptbench/configs/datasets/indian/__init__.py

-# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-
-def _maker(protocol, resize_size=512, cc_size=512, RGB=False):
-    from torchvision import transforms
-
-    from ....data.indian import dataset as raw
-    from ....data.transforms import ElasticDeformation, RemoveBlackBorders
-    from .. import make_dataset as mk
-
-    post_transforms = []
-    if RGB:
-        post_transforms = [
-            transforms.Lambda(lambda x: x.convert("RGB")),
-            transforms.ToTensor(),
-        ]
-
-    return mk(
-        [raw.subsets(protocol)],
-        [
-            RemoveBlackBorders(),
-            transforms.Resize(resize_size),
-            transforms.CenterCrop(cc_size),
-        ],
-        [ElasticDeformation(p=0.8)],
-        post_transforms,
-    )

src/ptbench/data/indian/__init__.py

@@ -17,6 +17,7 @@ import importlib.resources
 import os
 
 from ...utils.rc import load_rc
+from .. import make_dataset
 from ..dataset import JSONDataset
 from ..loader import load_pil_baw, make_delayed
 
@@ -50,9 +51,33 @@ def _loader(context, sample):
     return make_delayed(sample, _raw_data_loader)
 
 
-dataset = JSONDataset(
+json_dataset = JSONDataset(
     protocols=_protocols,
     fieldnames=("data", "label"),
     loader=_loader,
 )
 """Indian dataset object."""
+
+
+def _maker(protocol, resize_size=512, cc_size=512, RGB=False):
+    from torchvision import transforms
+
+    from ..transforms import ElasticDeformation, RemoveBlackBorders
+
+    post_transforms = []
+    if RGB:
+        post_transforms = [
+            transforms.Lambda(lambda x: x.convert("RGB")),
+            transforms.ToTensor(),
+        ]
+
+    return make_dataset(
+        [json_dataset.subsets(protocol)],
+        [
+            RemoveBlackBorders(),
+            transforms.Resize(resize_size),
+            transforms.CenterCrop(cc_size),
+        ],
+        [ElasticDeformation(p=0.8)],
+        post_transforms,
+    )

src/ptbench/configs/datasets/indian/default.py → src/ptbench/data/indian/default.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("default")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class DefaultModule(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("default")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = DefaultModule

src/ptbench/configs/datasets/indian/fold_0.py → src/ptbench/data/indian/fold_0.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_0")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_0")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_0_rgb.py → src/ptbench/data/indian/fold_0_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_0", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_0", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_1.py → src/ptbench/data/indian/fold_1.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_1")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_1")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_1_rgb.py → src/ptbench/data/indian/fold_1_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_1", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_1", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_2.py → src/ptbench/data/indian/fold_2.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_2")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_2")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_2_rgb.py → src/ptbench/data/indian/fold_2_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_2", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_2", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_3.py → src/ptbench/data/indian/fold_3.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_3")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_3")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_3_rgb.py → src/ptbench/data/indian/fold_3_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_3", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_3", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_4.py → src/ptbench/data/indian/fold_4.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_4")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_4")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_4_rgb.py → src/ptbench/data/indian/fold_4_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_4", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_4", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_5.py → src/ptbench/data/indian/fold_5.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_5")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_5")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_5_rgb.py → src/ptbench/data/indian/fold_5_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_5", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_5", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_6.py → src/ptbench/data/indian/fold_6.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_6")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_6")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_6_rgb.py → src/ptbench/data/indian/fold_6_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_6", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_6", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_7.py → src/ptbench/data/indian/fold_7.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_7")
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_7")
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module

src/ptbench/configs/datasets/indian/fold_7_rgb.py → src/ptbench/data/indian/fold_7_rgb.py

@@ -9,6 +9,38 @@
 * See :py:mod:`ptbench.data.indian` for dataset details
 """
 
+from clapper.logging import setup
+
+from .. import return_subsets
+from ..base_datamodule import BaseDataModule
 from . import _maker
 
-dataset = _maker("fold_7", RGB=True)
+logger = setup(__name__.split(".")[0], format="%(levelname)s: %(message)s")
+
+
+class Fold0Module(BaseDataModule):
+    def __init__(
+        self,
+        train_batch_size=1,
+        predict_batch_size=1,
+        drop_incomplete_batch=False,
+        multiproc_kwargs=None,
+    ):
+        super().__init__(
+            train_batch_size=train_batch_size,
+            predict_batch_size=predict_batch_size,
+            drop_incomplete_batch=drop_incomplete_batch,
+            multiproc_kwargs=multiproc_kwargs,
+        )
+
+    def setup(self, stage: str):
+        self.dataset = _maker("fold_7", RGB=True)
+        (
+            self.train_dataset,
+            self.validation_dataset,
+            self.extra_validation_datasets,
+            self.predict_dataset,
+        ) = return_subsets(self.dataset)
+
+
+datamodule = Fold0Module