[ENH] add conditional test skips to estimator specific tests (#607)

adds conditional test skips to estimator specific tests to manage test
time, see #603.

Uses `scikit-base` to check for changes in estimators.

Author: fkiraly

.github/workflows/testing.yml

@@ -28,6 +28,10 @@ jobs:
     steps:
       - uses: actions/checkout@v6
 
+      - run: git remote set-branches origin 'main'
+
+      - run: git fetch --depth 1
+
       - name: Install uv
         uses: astral-sh/setup-uv@v7
         with:

pyproject.toml

@@ -56,6 +56,7 @@ dependencies = [
     "einops>=0.6.0,<0.8.0",
     "fsspec>=2022.5.0,<2024.4.0; python_version == '3.8'",
     "rich",
+    "scikit-base",
 ]
 
 

tests/test_autoint.py

@@ -1,13 +1,18 @@
 #!/usr/bin/env python
 """Tests for `pytorch_tabular` package."""
 
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 
 from pytorch_tabular import TabularModel
 from pytorch_tabular.config import DataConfig, OptimizerConfig, TrainerConfig
 from pytorch_tabular.models import AutoIntConfig
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.autoint"),
+    reason="run test only if autoint module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols", [["AveRooms", "AveBedrms", "Population", "AveOccup", "Latitude", "Longitude"]]
@@ -78,6 +83,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.autoint"),
+    reason="run test only if autoint module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",

tests/test_danet.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python
 """Tests for `pytorch_tabular` package."""
 
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 
 from pytorch_tabular import TabularModel
@@ -10,6 +11,10 @@
 # from pytorch_tabular.categorical_encoders import CategoricalEmbeddingTransformer
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.danet"),
+    reason="run test only if danet module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols",
@@ -80,6 +85,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.danet"),
+    reason="run test only if danet module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",

tests/test_ft_transformer.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python
 """Tests for `pytorch_tabular` package."""
 
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 
 from pytorch_tabular import TabularModel
@@ -9,6 +10,10 @@
 from pytorch_tabular.models import FTTransformerConfig
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.ft_transformer"),
+    reason="run test only if ft_transformer module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols",
@@ -86,6 +91,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.ft_transformer"),
+    reason="run test only if ft_transformer module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",

tests/test_gandalf.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python
 """Tests for `pytorch_tabular` package."""
 
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 
 from pytorch_tabular import TabularModel
@@ -10,6 +11,10 @@
 # from pytorch_tabular.categorical_encoders import CategoricalEmbeddingTransformer
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.gandalf"),
+    reason="run test only if gandalf module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols",
@@ -80,6 +85,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.gandalf"),
+    reason="run test only if gandalf module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",

tests/test_gate.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python
 """Tests for `pytorch_tabular` package."""
 
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 
 from pytorch_tabular import TabularModel
@@ -10,6 +11,10 @@
 # from pytorch_tabular.categorical_encoders import CategoricalEmbeddingTransformer
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.gate"),
+    reason="run test only if gate module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols",
@@ -85,6 +90,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.gate"),
+    reason="run test only if gate module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",

tests/test_model_stacking.py

@@ -1,4 +1,5 @@
 import numpy as np
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 import torch
 from sklearn.preprocessing import PowerTransformer
@@ -50,6 +51,10 @@ def get_model_configs(task):
     return [model_config(task) for model_config in all_model_configs]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.stacking"),
+    reason="run test only if stacking module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols",
@@ -163,6 +168,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.stacking"),
+    reason="run test only if stacking module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",

tests/test_tabnet.py

@@ -1,13 +1,18 @@
 #!/usr/bin/env python
 """Tests for `pytorch_tabular` package."""
 
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 
 from pytorch_tabular import TabularModel
 from pytorch_tabular.config import DataConfig, OptimizerConfig, TrainerConfig
 from pytorch_tabular.models import TabNetModelConfig
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.tabnet"),
+    reason="run test only if tabnet module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols",
@@ -78,6 +83,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.tabnet"),
+    reason="run test only if tabnet module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",

tests/test_tabtransformer.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python
 """Tests for `pytorch_tabular` package."""
 
+from skbase.utils.git_diff import _is_module_changed
 import pytest
 
 from pytorch_tabular import TabularModel
@@ -9,6 +10,10 @@
 from pytorch_tabular.models import TabTransformerConfig
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.tab_transformer"),
+    reason="run test only if tab_transformer module is changed",
+)
 @pytest.mark.parametrize("multi_target", [True, False])
 @pytest.mark.parametrize(
     "continuous_cols",
@@ -84,6 +89,10 @@ def test_regression(
     assert pred_df.shape[0] == test.shape[0]
 
 
+@pytest.mark.skipif(
+    not _is_module_changed("pytorch_tabular.models.tab_transformer"),
+    reason="run test only if tab_transformer module is changed",
+)
 @pytest.mark.parametrize("multi_target", [False, True])
 @pytest.mark.parametrize(
     "continuous_cols",