Added test for time points

src/pd_dwi/config/config.py

@@ -96,4 +96,11 @@ def validate_encoders_dataset(self):
         if not masks.issubset(self.dataset.masks):
             raise ValueError("Encoders contain masks that are not available in dataset")
 
+        time_points = self.pipeline.features_transformer.radiomics.encoders[0].time_points
+        for e in self.pipeline.features_transformer.radiomics.encoders[1:]:
+            time_points = time_points.union(e.time_points)
+
+        if not time_points.issubset(self.dataset.time_points):
+            raise ValueError("Encoders contain time points that are not available in dataset")
+
         return self

tests/data/invalid_encoders_bad_time_points.yaml

@@ -0,0 +1,57 @@
+dataset:
+  labels:
+    negative: Non-pCR
+    positive: pCR
+
+  time_points:
+    - T0
+    - T2
+
+  modalities:
+    - ADC 0100
+    - F
+
+  masks:
+    - DWI MASK
+
+pipeline:
+  features_transformer:
+    radiomics:
+      encoders: # encoder names format is {time point}_{image}
+        - image: ADC 0100
+          mask: DWI MASK
+          time_points:
+            - T0
+            - T1
+            - T2
+        - image: F
+          mask: DWI MASK
+          time_points:
+            - T0
+            - T1
+      engine:
+        imageType:
+          Original: { }
+        setting:
+          resampledPixelSpacing: [ 1.27,1.27,4.0 ]
+  feature_selection:
+    k: 100
+  classifier:
+    module: samples.XGBClassifier
+    parameters:
+      random_state: 42
+      use_label_encoder: False
+      validate_parameters: True
+      learning_rate: 0.01
+      n_estimators: 1000
+      max_depth: 20
+      min_child_weight: 10
+
+grid_search_cv:
+  verbose: 3
+  scoring: roc_auc
+  cv: 5
+  param_grid:
+    classifier:
+      scale_pos_weight:
+        - 2.22 # balanced

tests/test_config.py

@@ -20,7 +20,8 @@ def test_from_config_valid():
     'config_name,validation_msg',
     [
         ('invalid_no_modalities', 'Field required'),
-        ('invalid_encoders_bad_modality', 'Value error, Encoders contain modalities that are not available in dataset')
+        ('invalid_encoders_bad_modality', 'Value error, Encoders contain modalities that are not available in dataset'),
+        ('invalid_encoders_bad_time_points', 'Value error, Encoders contain time points that are not available in dataset')
     ]
 )
 def test_from_config_invalid(config_name, validation_msg):