print target class labels too in feature imporance tables

BIMSBbioinfo · May 31, 2024 · d81fe29 · d81fe29
1 parent 53dee25
commit d81fe29
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 3 deletions.
diff --git a/flexynesis/__main__.py b/flexynesis/__main__.py
@@ -246,7 +246,7 @@ class AvailableModels(NamedTuple):
 
     # evaluate predictions;  (if any supervised learning happened)
     if any([args.target_variables, args.surv_event_var, args.batch_variables]):
-        print("[INFO] Computing model evaluation metrics sdfadf")
+        print("[INFO] Computing model evaluation metrics")
         metrics_df = flexynesis.evaluate_wrapper(model.predict(test_dataset), test_dataset, 
                                                  surv_event_var=model.surv_event_var, 
                                                  surv_time_var=model.surv_time_var)

diff --git a/flexynesis/models/direct_pred.py b/flexynesis/models/direct_pred.py
@@ -317,8 +317,10 @@ def compute_feature_importance(self, dataset, target_var, steps=5, batch_size =
                 features = dataset.features[layers[j]]
                 # Ensure tensors are already on CPU before converting to numpy
                 importances = imp[i][j][0].detach().numpy()
+                target_class_label = dataset.label_mappings[target_var].get(i) if target_var in dataset.label_mappings else ''
                 df_list.append(pd.DataFrame({'target_variable': target_var, 
                                              'target_class': i, 
+                                             'target_class_label': target_class_label,
                                              'layer': layers[j], 
                                              'name': features, 
                                              'importance': importances}))    

diff --git a/flexynesis/models/supervised_vae.py b/flexynesis/models/supervised_vae.py
@@ -468,9 +468,12 @@ def compute_feature_importance(self, dataset, target_var, steps = 5, batch_size
             for j in range(len(layers)):
                 features = self.dataset.features[layers[j]]
                 importances = imp[i][j][0].detach().numpy()
+                target_class_label = dataset.label_mappings[target_var].get(i) if target_var in dataset.label_mappings else ''
                 df_list.append(pd.DataFrame({'target_variable': target_var, 
-                                             'target_class': i, 'layer': layers[j], 
-                                             'name': features, 'importance': importances}))    
+                                             'target_class': i, 
+                                             'target_class_label': target_class_label,
+                                             'layer': layers[j], 
+                                             'name': features, 'importance': importances}))  
         df_imp = pd.concat(df_list, ignore_index = True)
 
         # save scores in model

diff --git a/flexynesis/models/triplet_encoder.py b/flexynesis/models/triplet_encoder.py
@@ -390,8 +390,10 @@ def compute_feature_importance(self, dataset, target_var, steps = 5, batch_size
                 features = dataset.features[layers[j]]
                 # Ensure tensors are already on CPU before converting to numpy
                 importances = imp[i][j][0].detach().numpy() # 0 => extract importances only for the anchor
+                target_class_label = dataset.label_mappings[target_var].get(i) if target_var in dataset.label_mappings else ''
                 df_list.append(pd.DataFrame({'target_variable': target_var, 
                                              'target_class': i, 
+                                             'target_class_label': target_class_label,
                                              'layer': layers[j], 
                                              'name': features, 
                                              'importance': importances}))