Merge pull request #38 from reghbali/reghbali_more_volume_features

Reghbali more volume features

.github/workflows/integration-test.yml

@@ -15,8 +15,14 @@ jobs:
         python-version: [ "3.9", "3.11" ]
 
     steps:
+      - name: Get branch
+        uses: xt0rted/pull-request-comment-branch@v2
+        id: comment-branch
+
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
+        with:
+          ref: ${{ steps.comment-branch.outputs.head_ref }}
 
       - name: Cache dependency binaries
         id: cache-bin

.github/workflows/tests.yml

@@ -16,7 +16,7 @@ jobs:
 
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
 
       - name: Cache dependency binaries
         id: cache-bin

pyalfe/tasks/quantification.py

@@ -101,11 +101,17 @@ def load_template_images(self, accession, target):
                     resampling_origin=Modality.T1,
                     sub_dir_name=roi_sub_dir,
                 )
-                print(template_image_to_target_file)
+
                 if os.path.exists(template_image_to_target_file):
                     target_images[roi_key], _ = self.load(template_image_to_target_file)
+                    self.logger.info(
+                        'loaded template: ' + template_image_to_target_file
+                    )
                     target_images[roi_key] = np.round(target_images[roi_key])
                 else:
+                    self.logger.info(
+                        'missing template: ' + template_image_to_target_file
+                    )
                     target_images[roi_key] = None
         return target_images
 
@@ -117,7 +123,7 @@ def get_radiomics(self, skullstripped_file, lesion_seg_file):
             return {}
 
     def get_brain_volume_stats(
-        self, brain_mask, tissue_seg, ventricles_seg, voxel_volume
+        self, brain_mask, tissue_seg, ventricles_seg, template_images, voxel_volume
     ):
         stats = {}
         brain_indices = np.where(brain_mask == 1)[0]
@@ -134,6 +140,15 @@ def get_brain_volume_stats(
                     len(np.where(tissue_seg == tissue)[0]) * voxel_volume
                 )
 
+        for template_key, template_image in template_images.items():
+            if 'regions' not in roi_dict[template_key]:
+                continue
+            regions = roi_dict[template_key]['regions']
+            for region_key, region_values in regions.items():
+                stats[f'volume_of_{region_key}'] = (
+                    np.sum(np.isin(template_image, region_values)) * voxel_volume
+                )
+
         return stats
 
     def get_lesion_stats(
@@ -219,6 +234,7 @@ def get_lesion_stats(
                 stats[
                     f'ninety_five_percentile_{modality_name.lower()}_signal'
                 ] = np.percentile(modality_image[lesion_indices], 95)
+
         if Modality.T1 in modality_images and Modality.T1Post in modality_images:
             t1_image = modality_images[Modality.T1]
             t1post_image = modality_images[Modality.T1Post]
@@ -279,6 +295,8 @@ def run(self, accession):
         else:
             tissue_seg, _ = self.load(tissue_seg_file)
 
+        template_images = self.load_template_images(accession, Modality.T1)
+
         ventricles_seg_file = self.pipeline_dir.get_output_image(
             accession=accession, modality=Modality.T1, image_type='VentriclesSeg'
         )

tests/integration/test_integration.py

@@ -33,7 +33,9 @@ def test_run(self):
             Modality.FLAIR,
             Modality.ADC,
         ]
-        pipeline_dir = DefaultALFEDataDir(output=self.output_dir, input=self.input_dir)
+        pipeline_dir = DefaultALFEDataDir(
+            output_dir=self.output_dir, input_dir=self.input_dir
+        )
 
         for modality in modalities:
             pipeline_dir.create_dir('input', accession, modality)
@@ -82,19 +84,26 @@ def test_run(self):
             segmentation_path = pipeline_dir.get_output_image(
                 accession,
                 modality,
-                image_type='CNNAbnormalMap_seg',
+                image_type='abnormal_seg',
                 sub_dir_name='abnormalmap',
             )
             self.assertTrue(
                 os.path.exists(segmentation_path),
                 msg=f'{segmentation_path} does not exist.',
             )
-            quantification_path = pipeline_dir.get_quantification_file(
+            summary_quantification_path = pipeline_dir.get_quantification_file(
                 accession, modality, 'SummaryLesionMeasures'
             )
+            individual_quantification_path = pipeline_dir.get_quantification_file(
+                accession, modality, 'IndividualLesionMeasures'
+            )
+
             self.assertTrue(
-                os.path.exists(quantification_path),
-                msg=f'{quantification_path} does not exist.',
+                os.path.exists(summary_quantification_path),
+                msg=f'{summary_quantification_path} does not exist.',
             )
-            quantification = pd.read_csv(quantification_path)
-            self.assertEqual(quantification.dropna().shape, (14, 2))
+            summary_quantification = pd.read_csv(summary_quantification_path)
+            self.assertEqual(summary_quantification.dropna().shape, (53, 2))
+
+            individual_quantification = pd.read_csv(individual_quantification_path)
+            self.assertEqual(individual_quantification.dropna().shape, (226, 51))

tests/unit/test_image_registration.py

@@ -93,7 +93,7 @@ def test_register_affine(self):
             np.testing.assert_array_less(
                 np.linalg.norm(get_nifti_data(output) - fixed_data)
                 / np.linalg.norm(get_nifti_data(output) - moving_data),
-                0.4,
+                0.45,
             )
 
             moving2_path = self.get_image_path('moving2.nii.gz')

tests/unit/test_tasks.py

@@ -584,6 +584,11 @@ def test_get_brain_volume_stats(self):
         brain_seg = np.array([0, 0, 1, 1, 1, 1, 1, 1, 1, 0, 0])
         tissue_seg = np.array([0, 0, 1, 2, 3, 4, 5, 6, 1, 0, 0])
         ventricles_seg = np.array([0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0])
+        template_images = {
+            'template': np.array([0.0, 1.0, 1.0, 1.0, 1.0, 1.0, 2.0, 1.0, 0.0]),
+            'lobes': np.array([0, 1, 2, 3, 4, 5, 6, 6, 2, 3]),
+            'CorpusCallosum': np.array([0, 1, 2, 3, 4, 5, 4, 3, 0, 0]),
+        }
 
         task = Quantification(
             pipeline_dir=self.pipeline_dir,
@@ -593,7 +598,7 @@ def test_get_brain_volume_stats(self):
         )
 
         volume_stats = task.get_brain_volume_stats(
-            brain_seg, tissue_seg, ventricles_seg, voxel_volume=2.0
+            brain_seg, tissue_seg, ventricles_seg, template_images, voxel_volume=2.0
         )
 
         self.assertEqual(14.0, volume_stats['total_brain_volume'])
@@ -605,3 +610,16 @@ def test_get_brain_volume_stats(self):
         self.assertEqual(2.0, volume_stats['volume_of_deep_gray_matter'])
         self.assertEqual(2.0, volume_stats['volume_of_brain_stem'])
         self.assertEqual(2.0, volume_stats['volume_of_cerebellum'])
+        self.assertEqual(2.0, volume_stats['volume_of_Frontal'])
+        self.assertEqual(4.0, volume_stats['volume_of_Parietal'])
+        self.assertEqual(4.0, volume_stats['volume_of_Occipital'])
+        self.assertEqual(2.0, volume_stats['volume_of_AnteriorTemporal'])
+        self.assertEqual(2.0, volume_stats['volume_of_MiddleTemporal'])
+        self.assertEqual(4.0, volume_stats['volume_of_PosteriorTemporal'])
+        self.assertEqual(8.0, volume_stats['volume_of_Parietal_Occipital'])
+        self.assertEqual(14.0, volume_stats['volume_of_CorpusCallosum'])
+        self.assertEqual(2.0, volume_stats['volume_of_CorpusCallosum_Rostrum'])
+        self.assertEqual(2.0, volume_stats['volume_of_CorpusCallosum_Genu'])
+        self.assertEqual(4.0, volume_stats['volume_of_CorpusCallosum_Body'])
+        self.assertEqual(4.0, volume_stats['volume_of_CorpusCallosum_Isthmus'])
+        self.assertEqual(2.0, volume_stats['volume_of_CorpusCallosum_Splenium'])