diff --git a/doc/api.rst b/doc/api.rst
index bcb3f2032ccef8e25f1087ad6f8582060a9e41b5..0114d9ff62b83dfc2178ade21426a7a2cb334dfe 100644
--- a/doc/api.rst
+++ b/doc/api.rst
@@ -57,8 +57,6 @@ Direct data-access through iterators.
ptbench.data.tbx11k_simplified_RS
ptbench.data.tbx11k_simplified_v2
ptbench.data.tbx11k_simplified_v2_RS
- ptbench.data.tbx11k_simplified_v3
- ptbench.data.tbx11k_simplified_v3_RS
.. _ptbench.api.models:
diff --git a/doc/config.rst b/doc/config.rst
index 4e402bca1d96513b36cc860e44d4b58c68214ffc..828bba07a4c3db58cdc4854c7ab1c4e407f21f17 100644
--- a/doc/config.rst
+++ b/doc/config.rst
@@ -80,9 +80,6 @@ if applicable. Use these datasets for training and evaluating your models.
ptbench.configs.datasets.tbx11k_simplified_v2.default
ptbench.configs.datasets.tbx11k_simplified_v2.rgb
ptbench.configs.datasets.tbx11k_simplified_v2_RS.default
- ptbench.configs.datasets.tbx11k_simplified_v3.default
- ptbench.configs.datasets.tbx11k_simplified_v3.rgb
- ptbench.configs.datasets.tbx11k_simplified_v3_RS.default
.. _ptbench.configs.datasets.folds:
@@ -133,9 +130,6 @@ datasets. Nine other folds are available for every configuration (from 1 to
ptbench.configs.datasets.tbx11k_simplified_v2.fold_0
ptbench.configs.datasets.tbx11k_simplified_v2.fold_0_rgb
ptbench.configs.datasets.tbx11k_simplified_v2_RS.fold_0
- ptbench.configs.datasets.tbx11k_simplified_v3.fold_0
- ptbench.configs.datasets.tbx11k_simplified_v3.fold_0_rgb
- ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_0
.. include:: links.rst
diff --git a/pyproject.toml b/pyproject.toml
index 609a0a08be71c96e5f338c59aeb4f8ad289ae8aa..4f6bf7f730c48448d1edd4e0382028de8c1fd978 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -256,41 +256,6 @@ tbx11k_simplified_v2_rs_f6 = "ptbench.configs.datasets.tbx11k_simplified_v2_RS.f
tbx11k_simplified_v2_rs_f7 = "ptbench.configs.datasets.tbx11k_simplified_v2_RS.fold_7"
tbx11k_simplified_v2_rs_f8 = "ptbench.configs.datasets.tbx11k_simplified_v2_RS.fold_8"
tbx11k_simplified_v2_rs_f9 = "ptbench.configs.datasets.tbx11k_simplified_v2_RS.fold_9"
-# TBX11K simplified dataset split 3 (and cross-validation folds)
-tbx11k_simplified_v3 = "ptbench.configs.datasets.tbx11k_simplified_v3.default"
-tbx11k_simplified_v3_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.rgb"
-tbx11k_simplified_v3_f0 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_0"
-tbx11k_simplified_v3_f1 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_1"
-tbx11k_simplified_v3_f2 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_2"
-tbx11k_simplified_v3_f3 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_3"
-tbx11k_simplified_v3_f4 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_4"
-tbx11k_simplified_v3_f5 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_5"
-tbx11k_simplified_v3_f6 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_6"
-tbx11k_simplified_v3_f7 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_7"
-tbx11k_simplified_v3_f8 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_8"
-tbx11k_simplified_v3_f9 = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_9"
-tbx11k_simplified_v3_f0_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_0_rgb"
-tbx11k_simplified_v3_f1_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_1_rgb"
-tbx11k_simplified_v3_f2_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_2_rgb"
-tbx11k_simplified_v3_f3_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_3_rgb"
-tbx11k_simplified_v3_f4_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_4_rgb"
-tbx11k_simplified_v3_f5_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_5_rgb"
-tbx11k_simplified_v3_f6_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_6_rgb"
-tbx11k_simplified_v3_f7_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_7_rgb"
-tbx11k_simplified_v3_f8_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_8_rgb"
-tbx11k_simplified_v3_f9_rgb = "ptbench.configs.datasets.tbx11k_simplified_v3.fold_9_rgb"
-# extended TBX11K simplified dataset split 3 (with radiological signs)
-tbx11k_simplified_v3_rs = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.default"
-tbx11k_simplified_v3_rs_f0 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_0"
-tbx11k_simplified_v3_rs_f1 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_1"
-tbx11k_simplified_v3_rs_f2 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_2"
-tbx11k_simplified_v3_rs_f3 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_3"
-tbx11k_simplified_v3_rs_f4 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_4"
-tbx11k_simplified_v3_rs_f5 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_5"
-tbx11k_simplified_v3_rs_f6 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_6"
-tbx11k_simplified_v3_rs_f7 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_7"
-tbx11k_simplified_v3_rs_f8 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_8"
-tbx11k_simplified_v3_rs_f9 = "ptbench.configs.datasets.tbx11k_simplified_v3_RS.fold_9"
# montgomery-shenzhen aggregated dataset
mc_ch = "ptbench.configs.datasets.mc_ch.default"
mc_ch_rgb = "ptbench.configs.datasets.mc_ch.rgb"
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/__init__.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/__init__.py
deleted file mode 100644
index bed38616924b7f736f5d33d12c7bdf7911dc590f..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/__init__.py
+++ /dev/null
@@ -1,25 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-
-def _maker(protocol, RGB=False):
- from torchvision import transforms
-
- from ....data.tbx11k_simplified_v3 import dataset as raw
- from ....data.transforms import ElasticDeformation
- 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)],
- [],
- [ElasticDeformation(p=0.8)],
- post_transforms,
- )
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/default.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/default.py
deleted file mode 100644
index 3805a14219167fbcc1cd0becfa28d592c422813d..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/default.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (default protocol)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("default")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_0.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_0.py
deleted file mode 100644
index bed7342dd6f28a83ca3c5a4e7984551a153344d7..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_0.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 0)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_0")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_0_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_0_rgb.py
deleted file mode 100644
index ed2f6d174d87521d3fe2ad12e211e0e32bab68e2..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_0_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 0, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_0", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_1.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_1.py
deleted file mode 100644
index 92cc9ce31f3f296a52d9ba51eae87d39740cb160..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_1.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 1)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_1")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_1_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_1_rgb.py
deleted file mode 100644
index 9380ec5b3910f463fea0cab48ea3cdb738647302..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_1_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 1, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_1", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_2.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_2.py
deleted file mode 100644
index 460c96df367fa982dad4dca111c41ad5118f11f1..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_2.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 2)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_2")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_2_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_2_rgb.py
deleted file mode 100644
index 69641e3acc24e1de79b09832eb7b8340879cfc4a..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_2_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 2, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_2", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_3.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_3.py
deleted file mode 100644
index c799c7cb54044be807bc3cc615658c39ca47667c..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_3.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 3)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_3")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_3_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_3_rgb.py
deleted file mode 100644
index 816eeca4c7eeae7ffc30bd58c6f527aec999b179..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_3_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 3, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_3", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_4.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_4.py
deleted file mode 100644
index acceda91f4caa5773b5f0b478bfc5c1e9e4dc1d8..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_4.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 4)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_4")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_4_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_4_rgb.py
deleted file mode 100644
index 07811e9c185f2c8165cc0e156dce5bb3292656c1..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_4_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 4, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_4", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_5.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_5.py
deleted file mode 100644
index 20b048d5be87e947db7622ed34c1aaa30f166084..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_5.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 5)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_5")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_5_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_5_rgb.py
deleted file mode 100644
index 6b851bdcdffb0029b2dd70fe74a4c9afffaff826..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_5_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 5, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_5", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_6.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_6.py
deleted file mode 100644
index e99afb4c0ab99ac8de31522181910e89b6f61b3e..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_6.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 6)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_6")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_6_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_6_rgb.py
deleted file mode 100644
index 01a94e08e348358983eb9bdd40f155818733555e..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_6_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 6, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_6", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_7.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_7.py
deleted file mode 100644
index 53e5af59d4bc1cdc145af9756c9a8f494f2519cb..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_7.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 7)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_7")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_7_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_7_rgb.py
deleted file mode 100644
index 546d63584886094cf6bd7a5267e0e131197faa3a..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_7_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 7, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_7", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_8.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_8.py
deleted file mode 100644
index d59887df9ce9849fca1483d8c3c4adb97bf7d0b5..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_8.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 8)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_8")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_8_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_8_rgb.py
deleted file mode 100644
index 243c1455a3d2a06e14b535b303b8a7bbfe70e8fa..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_8_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 8, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_8", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_9.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_9.py
deleted file mode 100644
index 2f9a45f01f31a9f1c87d8b70aa9e5859cee83b04..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_9.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 9)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_9")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_9_rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_9_rgb.py
deleted file mode 100644
index ef3d504550a094970ad49d16b307e1d0ae0782db..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/fold_9_rgb.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 9, RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_9", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3/rgb.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3/rgb.py
deleted file mode 100644
index 1c274f79252bff0a06bc52edff8e1590b265fccd..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3/rgb.py
+++ /dev/null
@@ -1,18 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (default protocol, converted in
-RGB)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("default", RGB=True)
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/__init__.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/__init__.py
deleted file mode 100644
index 176a28448003e2107060b089d1e9b0418c49296a..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/__init__.py
+++ /dev/null
@@ -1,10 +0,0 @@
-# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-
-def _maker(protocol):
- from ....data.tbx11k_simplified_v3_RS import dataset as raw
- from .. import make_dataset as mk
-
- return mk([raw.subsets(protocol)])
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/default.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/default.py
deleted file mode 100644
index 7c62887d8a4eb2d66e9ad287cfa0fb20020c97b3..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/default.py
+++ /dev/null
@@ -1,18 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (default protocol) (extended with
-DensenetRS predictions)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("default")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_0.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_0.py
deleted file mode 100644
index bed7342dd6f28a83ca3c5a4e7984551a153344d7..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_0.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 0)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_0")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_1.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_1.py
deleted file mode 100644
index 92cc9ce31f3f296a52d9ba51eae87d39740cb160..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_1.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 1)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_1")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_2.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_2.py
deleted file mode 100644
index 460c96df367fa982dad4dca111c41ad5118f11f1..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_2.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 2)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_2")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_3.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_3.py
deleted file mode 100644
index c799c7cb54044be807bc3cc615658c39ca47667c..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_3.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 3)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_3")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_4.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_4.py
deleted file mode 100644
index acceda91f4caa5773b5f0b478bfc5c1e9e4dc1d8..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_4.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 4)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_4")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_5.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_5.py
deleted file mode 100644
index 20b048d5be87e947db7622ed34c1aaa30f166084..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_5.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 5)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_5")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_6.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_6.py
deleted file mode 100644
index e99afb4c0ab99ac8de31522181910e89b6f61b3e..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_6.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 6)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_6")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_7.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_7.py
deleted file mode 100644
index 53e5af59d4bc1cdc145af9756c9a8f494f2519cb..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_7.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 7)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_7")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_8.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_8.py
deleted file mode 100644
index d59887df9ce9849fca1483d8c3c4adb97bf7d0b5..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_8.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 8)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_8")
diff --git a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_9.py b/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_9.py
deleted file mode 100644
index 2f9a45f01f31a9f1c87d8b70aa9e5859cee83b04..0000000000000000000000000000000000000000
--- a/src/ptbench/configs/datasets/tbx11k_simplified_v3_RS/fold_9.py
+++ /dev/null
@@ -1,17 +0,0 @@
-# Copyright © 2022 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11k simplified dataset for TB detection (cross validation fold 9)
-
-* Split reference: first 62.6% of CXR for "train", 16% for "validation",
-* 21.4% for "test"
-* This split consists of the 4 labels "healthy", "latent TB", "sick & non-TB",
-* and "active TB"
-* This configuration resolution: 512 x 512 (default)
-* See :py:mod:`ptbench.data.tbx11k_v3` for dataset details
-"""
-
-from . import _maker
-
-dataset = _maker("fold_9")
diff --git a/src/ptbench/data/tbx11k_simplified_v3/__init__.py b/src/ptbench/data/tbx11k_simplified_v3/__init__.py
deleted file mode 100644
index 7cf676b00d4570c8b2f6b42724f55eeb0eab2fc4..0000000000000000000000000000000000000000
--- a/src/ptbench/data/tbx11k_simplified_v3/__init__.py
+++ /dev/null
@@ -1,90 +0,0 @@
-# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""TBX11K simplified dataset for computer-aided diagnosis.
-
-The TBX11K database has been established to foster research
-in computer-aided diagnosis of pulmonary diseases with a special
-focus on tuberculosis (aTB). The dataset was specifically
-designed to be used with CNNs. It contains 11,000 chest X-ray
-images, each of a unique patient. They were labeled by expert
-radiologists with 5 - 10+ years of experience. Possible labels
-are: "healthy", "active TB", "latent TB", and "sick & non-tb".
-The version of the dataset used in this benchmark is a simplified.
-
-* Reference: [TBX11K-SIMPLIFIED-2020]_
-* Original (released) resolution (height x width or width x height): 512 x 512
-* Split reference: none
-* Protocol ``default``:
-
- * Training samples: 62.6% of CXR (including labels)
- * Validation samples: 16% of CXR (including labels)
- * Test samples: 21.4% of CXR (including labels)
-"""
-
-import importlib.resources
-import os
-
-from ...utils.rc import load_rc
-from ..dataset import JSONDataset
-from ..loader import load_pil_baw, make_delayed, make_delayed_bbox
-
-_protocols = [
- importlib.resources.files(__name__).joinpath("default.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_0.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_1.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_2.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_3.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_4.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_5.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_6.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_7.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_8.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_9.json.bz2"),
-]
-
-_datadir = load_rc().get(
- "datadir.tbx11k_simplified", os.path.realpath(os.curdir)
-)
-
-
-def _raw_data_loader(sample):
- return dict(
- data=load_pil_baw(os.path.join(_datadir, sample["data"])), # type: ignore
- label=sample["label"],
- )
-
-
-def _raw_data_loader_bbox(sample):
- return dict(
- data=load_pil_baw(os.path.join(_datadir, sample["data"])), # type: ignore
- label=sample["label"],
- bboxes=sample["bboxes"],
- )
-
-
-def _loader(context, sample):
- # "context" is ignored in this case - database is homogeneous
- # we return delayed samples to avoid loading all images at once
- return make_delayed(sample, _raw_data_loader)
-
-
-def _loader_bbox(context, sample):
- # "context" is ignored in this case - database is homogeneous
- # we return delayed samples to avoid loading all images at once
- return make_delayed_bbox(sample, _raw_data_loader_bbox)
-
-
-dataset = JSONDataset(
- protocols=_protocols,
- fieldnames=("data", "label"),
- loader=_loader,
-)
-
-dataset_with_bboxes = JSONDataset(
- protocols=_protocols,
- fieldnames=("data", "label", "bboxes"),
- loader=_loader_bbox,
-)
-"""TBX11K simplified dataset object."""
diff --git a/src/ptbench/data/tbx11k_simplified_v3/default.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/default.json.bz2
deleted file mode 100644
index 53b1c72281098faa727ce9b578210ae588e204c0..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/default.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_0.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_0.json.bz2
deleted file mode 100644
index 881297d6973db88dc2d6d1e19eb0c29f9062ed8a..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_0.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_1.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_1.json.bz2
deleted file mode 100644
index d85439be416a48142310799e6bf32b286224dd6a..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_1.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_2.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_2.json.bz2
deleted file mode 100644
index fb8c89af52076546d2a4da4831eb5f650c93cf18..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_2.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_3.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_3.json.bz2
deleted file mode 100644
index 0a8cc88fd74d7847785ea4a74cd7ced8c72fa89f..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_3.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_4.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_4.json.bz2
deleted file mode 100644
index 1ab03dfa21c6d8fe60a910343c4544f98e723634..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_4.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_5.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_5.json.bz2
deleted file mode 100644
index 17467ae2146adc756ebe7843e234aea18c5b9b14..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_5.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_6.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_6.json.bz2
deleted file mode 100644
index 2359d5e1ee6f14491715842f81a244acd083f7da..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_6.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_7.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_7.json.bz2
deleted file mode 100644
index 4cd04d5fd74a65704de41da44b5e638fa609a7fb..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_7.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_8.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_8.json.bz2
deleted file mode 100644
index fd8e62f18bf68ef535dc8e13f1291f9bdd5ad800..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_8.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3/fold_9.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3/fold_9.json.bz2
deleted file mode 100644
index 7018d30fae5a4a50f7aca57961274c3e0ef30987..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3/fold_9.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/__init__.py b/src/ptbench/data/tbx11k_simplified_v3_RS/__init__.py
deleted file mode 100644
index f2bcf8c2a53cd117f556c5854774d5ea81fb2f6c..0000000000000000000000000000000000000000
--- a/src/ptbench/data/tbx11k_simplified_v3_RS/__init__.py
+++ /dev/null
@@ -1,63 +0,0 @@
-# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""Extended TBX11K simplified dataset for computer-aided diagnosis (extended
-with DensenetRS predictions)
-
-The TBX11K database has been established to foster research
-in computer-aided diagnosis of pulmonary diseases with a special
-focus on tuberculosis (aTB). The dataset was specifically
-designed to be used with CNNs. It contains 11,000 chest X-ray
-images, each of a unique patient. They were labeled by expert
-radiologists with 5 - 10+ years of experience. Possible labels
-are: "healthy", "active TB", "latent TB", and "sick & non-tb".
-The version of the dataset used in this benchmark is a simplified.
-
-* Reference: [TBX11K-SIMPLIFIED-2020]_
-* Original (released) resolution (height x width or width x height): 512 x 512
-* Split reference: none
-* Protocol ``default``:
-
- * Training samples: 62.6% of CXR (including labels)
- * Validation samples: 16% of CXR (including labels)
- * Test samples: 21.4% of CXR (including labels)
-"""
-
-import importlib.resources
-
-from ..dataset import JSONDataset
-from ..loader import make_delayed
-
-_protocols = [
- importlib.resources.files(__name__).joinpath("default.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_0.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_1.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_2.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_3.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_4.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_5.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_6.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_7.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_8.json.bz2"),
- importlib.resources.files(__name__).joinpath("fold_9.json.bz2"),
-]
-
-
-def _raw_data_loader(sample):
- return dict(data=sample["data"], label=sample["label"])
-
-
-def _loader(context, sample):
- # "context" is ignored in this case - database is homogeneous
- # we returned delayed samples to avoid loading all images at once
- return make_delayed(sample, _raw_data_loader, key=sample["filename"])
-
-
-dataset = JSONDataset(
- protocols=_protocols,
- fieldnames=("filename", "label", "data"),
- loader=_loader,
-)
-
-"""Extended TBX11K simplified dataset object."""
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/default.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/default.json.bz2
deleted file mode 100644
index cc25eb3513cb826148474b6099c17e01c0337db9..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/default.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_0.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_0.json.bz2
deleted file mode 100644
index e3ed3611aba3587a53794658da76249462f11057..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_0.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_1.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_1.json.bz2
deleted file mode 100644
index 66980d95ee4551812530722d3ae6d734f874c061..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_1.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_2.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_2.json.bz2
deleted file mode 100644
index 423b892e1f47432b7c4e514815abc791be1613e0..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_2.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_3.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_3.json.bz2
deleted file mode 100644
index f9ab24097cbcd4d81e3e329e0bb51f6876b1df87..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_3.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_4.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_4.json.bz2
deleted file mode 100644
index bca9274bbd055bf821243c5739c08041b3013c25..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_4.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_5.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_5.json.bz2
deleted file mode 100644
index c1ad523698e9f9d2fdc2d61028558a0196f23216..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_5.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_6.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_6.json.bz2
deleted file mode 100644
index 518c12aa21e138241f5aa5a5d1829f2103a5a789..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_6.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_7.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_7.json.bz2
deleted file mode 100644
index f9d53add2a73dfb115581763061d8f3675d5767d..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_7.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_8.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_8.json.bz2
deleted file mode 100644
index 96667c33ee29d13ee056b4e0164a092b5d1b1149..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_8.json.bz2 and /dev/null differ
diff --git a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_9.json.bz2 b/src/ptbench/data/tbx11k_simplified_v3_RS/fold_9.json.bz2
deleted file mode 100644
index 7d78901b8ec9e3efdf1d57bc9ca92d9d591de562..0000000000000000000000000000000000000000
Binary files a/src/ptbench/data/tbx11k_simplified_v3_RS/fold_9.json.bz2 and /dev/null differ
diff --git a/tests/test_11k_v3.py b/tests/test_11k_v3.py
deleted file mode 100644
index 848b48757d6dc745f83f03d994732ea4b11ca2b0..0000000000000000000000000000000000000000
--- a/tests/test_11k_v3.py
+++ /dev/null
@@ -1,270 +0,0 @@
-# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""Tests for TBX11K simplified dataset split 3."""
-
-import pytest
-
-
-def test_protocol_consistency():
- from ptbench.data.tbx11k_simplified_v3 import dataset
-
- # Default protocol
- subset = dataset.subsets("default")
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 5241
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1335
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 1793
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- # Cross-validation fold 0-8
- for f in range(9):
- subset = dataset.subsets("fold_" + str(f))
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 6003
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1529
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 837
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- # Cross-validation fold 9
- subset = dataset.subsets("fold_9")
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 6003
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1530
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 836
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
-
-def test_protocol_consistency_bbox():
- from ptbench.data.tbx11k_simplified_v3 import dataset_with_bboxes
-
- # Default protocol
- subset = dataset_with_bboxes.subsets("default")
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 5241
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1335
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 1793
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- # Check bounding boxes
- for s in subset["train"]:
- assert s.bboxes == "none" or s.bboxes[0].startswith("{'xmin':")
-
- # Cross-validation fold 0-8
- for f in range(9):
- subset = dataset_with_bboxes.subsets("fold_" + str(f))
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 6003
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1529
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 837
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- # Check bounding boxes
- for s in subset["train"]:
- assert s.bboxes == "none" or s.bboxes[0].startswith("{'xmin':")
-
- # Cross-validation fold 9
- subset = dataset_with_bboxes.subsets("fold_9")
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 6003
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1530
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 836
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- # Check bounding boxes
- for s in subset["train"]:
- assert s.bboxes == "none" or s.bboxes[0].startswith("{'xmin':")
-
-
-@pytest.mark.skip_if_rc_var_not_set("datadir.tbx11k_simplified_v3")
-def test_loading():
- from ptbench.data.tbx11k_simplified_v3 import dataset
-
- def _check_sample(s):
- data = s.data
- assert isinstance(data, dict)
- assert len(data) == 2
-
- assert "data" in data
- assert data["data"].size == (512, 512)
-
- assert data["data"].mode == "L" # Check colors
-
- assert "label" in data
- assert data["label"] in [0, 1, 2, 3] # Check labels
-
- limit = 30 # use this to limit testing to first images only, else None
-
- subset = dataset.subsets("default")
- for s in subset["train"][:limit]:
- _check_sample(s)
-
-
-@pytest.mark.skip_if_rc_var_not_set("datadir.tbx11k_simplified_v3")
-def test_loading_bbox():
- from ptbench.data.tbx11k_simplified_v3 import dataset_with_bboxes
-
- def _check_sample(s):
- data = s.data
- assert isinstance(data, dict)
- assert len(data) == 3
-
- assert "data" in data
- assert data["data"].size == (512, 512)
-
- assert data["data"].mode == "L" # Check colors
-
- assert "label" in data
- assert data["label"] in [0, 1, 2, 3] # Check labels
-
- assert "bboxes" in data
- assert data["bboxes"] == "none" or data["bboxes"][0].startswith(
- "{'xmin':"
- )
-
- limit = 30 # use this to limit testing to first images only, else None
-
- subset = dataset_with_bboxes.subsets("default")
- for s in subset["train"][:limit]:
- _check_sample(s)
-
-
-@pytest.mark.skip_if_rc_var_not_set("datadir.tbx11k_simplified_v3")
-def test_check():
- from ptbench.data.tbx11k_simplified_v3 import dataset
-
- assert dataset.check() == 0
-
-
-@pytest.mark.skip_if_rc_var_not_set("datadir.tbx11k_simplified_v3")
-def test_check_bbox():
- from ptbench.data.tbx11k_simplified_v3 import dataset_with_bboxes
-
- assert dataset_with_bboxes.check() == 0
diff --git a/tests/test_11k_v3_RS.py b/tests/test_11k_v3_RS.py
deleted file mode 100644
index 76a83499d6147a44df5e6deb2c1db0dde49554d6..0000000000000000000000000000000000000000
--- a/tests/test_11k_v3_RS.py
+++ /dev/null
@@ -1,117 +0,0 @@
-# SPDX-FileCopyrightText: Copyright © 2023 Idiap Research Institute <contact@idiap.ch>
-#
-# SPDX-License-Identifier: GPL-3.0-or-later
-
-"""Tests for Extended TBX11K simplified dataset split 3."""
-
-import pytest
-
-
-def test_protocol_consistency():
- from ptbench.data.tbx11k_simplified_v3_RS import dataset
-
- # Default protocol
- subset = dataset.subsets("default")
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 5241
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1335
-
- assert "test" in subset
- assert len(subset["test"]) == 1793
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- # Cross-validation fold 0-8
- for f in range(9):
- subset = dataset.subsets("fold_" + str(f))
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 6003
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1529
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 837
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- # Cross-validation fold 9
- subset = dataset.subsets("fold_9")
- assert len(subset) == 3
-
- assert "train" in subset
- assert len(subset["train"]) == 6003
- for s in subset["train"]:
- assert s.key.startswith("images/")
-
- assert "validation" in subset
- assert len(subset["validation"]) == 1530
- for s in subset["validation"]:
- assert s.key.startswith("images/")
-
- assert "test" in subset
- assert len(subset["test"]) == 836
- for s in subset["test"]:
- assert s.key.startswith("images/")
-
- # Check labels
- for s in subset["train"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["validation"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
- for s in subset["test"]:
- assert s.label in [0.0, 1.0, 2.0, 3.0]
-
-
-@pytest.mark.skip_if_rc_var_not_set("datadir.tbx11k_simplified")
-def test_loading():
- from ptbench.data.tbx11k_simplified_v3_RS import dataset
-
- def _check_sample(s):
- data = s.data
- assert isinstance(data, dict)
- assert len(data) == 2
-
- assert "data" in data
- assert len(data["data"]) == 14 # Check radiological signs
-
- assert "label" in data
- assert data["label"] in [0, 1, 2, 3] # Check labels
-
- limit = 30 # use this to limit testing to first images only, else None
-
- subset = dataset.subsets("default")
- for s in subset["train"][:limit]:
- _check_sample(s)