[~] tests refaco

Author: Kirscher

.github/workflows/python-package.yml

@@ -16,7 +16,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: ["3.11", "3.12"]
+        python-version: ["3.12"]
 
     steps:
     - uses: actions/checkout@v4
@@ -37,4 +37,4 @@ jobs:
         flake8 . --count --exit-zero --max-complexity=10 --max-line-length=127 --statistics
     - name: Test with pytest
       run: |
-        pytest tests/test_create_totalseg_subset.py
+        pytest -q

README.md

@@ -21,3 +21,11 @@ PSAT addresses pediatric segmentation challenges by combining:
 - [Notebooks](notebooks/notebooks.md)
 - [Resources](resources/resources.md)
 - [Scripts](scripts/scripts.md)
+
+## Running Tests
+
+Install dependencies listed in `requirements.txt` and run:
+
+```bash
+pytest -q
+```

requirements.txt

@@ -4,3 +4,4 @@ numpy
 matplotlib
 nnunetv2
 scipy
+nibabel

scripts/create_dataset_subset.py

@@ -0,0 +1,44 @@
+import os
+import sys
+import types
+import numpy as np
+import nibabel as nib
+
+# Ensure repo root is on path
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
+
+# Dummy modules to satisfy imports in compute_metrics
+sys.modules.setdefault('p_tqdm', types.SimpleNamespace(p_map=lambda *a, **k: []))
+sys.modules.setdefault('surface_distance', types.SimpleNamespace(
+    compute_surface_distances=lambda *a, **k: None,
+    compute_robust_hausdorff=lambda *a, **k: 0,
+))
+
+from scripts.compute_metrics import dice_score, calculate_confidence_interval, reorient_to_ras
+
+
+def test_dice_score_perfect_match():
+    y_true = np.array([[1, 0], [0, 1]])
+    y_pred = np.array([[1, 0], [0, 1]])
+    assert np.isclose(dice_score(y_true, y_pred), 1)
+
+
+def test_dice_score_partial():
+    y_true = np.array([[1, 1], [0, 0]])
+    y_pred = np.array([[1, 0], [1, 0]])
+    assert np.isclose(dice_score(y_true, y_pred), 0.5)
+
+
+def test_confidence_interval():
+    data = [1, 1, 1, 1]
+    mean, lower, upper = calculate_confidence_interval(data, confidence=0.95)
+    assert mean == 1
+    assert lower <= mean <= upper
+
+
+def test_reorient_to_ras_identity():
+    arr = np.zeros((2, 2, 2))
+    affine = np.eye(4)
+    img = nib.Nifti1Image(arr, affine)
+    ras_img = reorient_to_ras(img)
+    assert ras_img.get_fdata().shape == img.get_fdata().shape

tests/test_compute_metrics.py

@@ -1,29 +1,60 @@
 # FILE: tests/test_create_totalseg_subset.py
-import unittest
-import pandas as pd
 import os
-from scripts.create_totalseg_subset import plot_and_save_distribution
+import sys
+import pandas as pd
+import pytest
+
+# Ensure package import works when tests are run directly
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
+
+from scripts.create_totalseg_subset import plot_and_save_distribution, create_directory_structure, copy_selected_files
 
-class TestCreateTotalSegSubset(unittest.TestCase):
-    def setUp(self):
+class TestCreateTotalSegSubset:
+    @pytest.fixture(autouse=True)
+    def setup_and_teardown(self, tmp_path):
         # Create a sample dataframe
         self.data = pd.DataFrame({
             'age': [25, 35, 45, 55, 65, 75],
             'gender': ['M', 'F', 'M', 'F', 'M', 'F']
         })
-        self.filename = 'test_distribution.png'
+        self.filename = tmp_path / 'test_distribution.png'
+        yield
+        # Remove the file after test if it exists
+        if self.filename.exists():
+            self.filename.unlink()
 
     def test_plot_and_save_distribution(self):
         # Call the function to plot and save the distribution
-        plot_and_save_distribution(self.data, "Test Title", self.filename)
-        
+        plot_and_save_distribution(self.data, "Test Title", str(self.filename))
         # Check if the file is created
-        self.assertTrue(os.path.exists(self.filename))
+        assert self.filename.exists()
 
-    def tearDown(self):
-        # Remove the file after test
-        if os.path.exists(self.filename):
-            os.remove(self.filename)
+    def test_create_directory_structure(self, tmp_path):
+        subdirs = ["imagesTr", "labelsTr"]
+        create_directory_structure(tmp_path, subdirs)
+        for subdir in subdirs:
+            assert (tmp_path / subdir).exists()
+            assert (tmp_path / subdir).is_dir()
 
-if __name__ == '__main__':
-    unittest.main()
+    def test_copy_selected_files(self, tmp_path):
+        # Setup original and new base directories
+        original_base = tmp_path / "original"
+        new_base = tmp_path / "new"
+        subdirs = ["imagesTr", "labelsTr"]
+        image_ids = ["img001", "img002"]
+        create_directory_structure(original_base, subdirs)
+        create_directory_structure(new_base, subdirs)
+        # Create files in original_base
+        for subdir in subdirs:
+            for img_id in image_ids:
+                file_path = original_base / subdir / f"{img_id}_something.nii.gz"
+                file_path.write_text("dummy data")
+            # Add a file that should not be copied
+            (original_base / subdir / "otherfile.nii.gz").write_text("not copied")
+        # Copy selected files
+        copy_selected_files(original_base, new_base, subdirs, image_ids)
+        # Check that only the correct files are copied
+        for subdir in subdirs:
+            for img_id in image_ids:
+                assert (new_base / subdir / f"{img_id}_something.nii.gz").exists()
+            assert not (new_base / subdir / "otherfile.nii.gz").exists()

tests/test_create_totalseg_subset.py

@@ -0,0 +1,28 @@
+import os
+import sys
+sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))
+
+from scripts.get_results import escape_latex, is_baseline, compute_statistical_tests, determine_best_scores
+import numpy as np
+
+
+def test_escape_latex():
+    assert escape_latex('foo_bar') == 'foo\\_bar'
+
+
+def test_is_baseline():
+    assert is_baseline('Something_TotalSegmentator')
+    assert not is_baseline('OtherModel')
+
+
+def test_compute_statistical_tests_and_best_scores():
+    results = {
+        'Baseline_TotalSegmentator': {'ROI': {'dataset': {'dice': [0.8, 0.9]}}},
+        'NewModel': {'ROI': {'dataset': {'dice': [0.95, 0.96]}}},
+    }
+    compute_statistical_tests(results)
+    assert 'pvalue' in results['NewModel']['ROI']['dataset']
+    assert 'max_baseline' in results['NewModel']['ROI']['dataset']
+
+    best = determine_best_scores(results, ['dataset'], ['Baseline_TotalSegmentator', 'NewModel'], ['ROI'])
+    assert best['ROI']['dataset'] == 'NewModel'