Fixed filtering and card sizes in the feature importance page

callbacks.py

@@ -424,6 +424,22 @@ def update_sample_table(n_clicks, dataset_store, id_column, ts_column, label_col
         return (data_columns, filtered_df.to_dict('records'), False,
                 f'Sample from patient {subject_id} on timestamp {ts}')
 
+@app.callback(Output('feature_importance_dropdown', 'options'),
+              [Input('dataset_store', 'modified_timestamp')],
+              [State('dataset_store', 'data')])
+def load_feat_import_filter_options(dataset_mod, dataset_store):
+    # Reconvert the dataframe to Pandas
+    df = pd.DataFrame(dataset_store)
+    # Get the list of columns without the identifier and the feature importance columns
+    shap_column_names = [feature for feature in df.columns
+                         if feature.endswith('_shap')]
+    feature_names = [feature.split('_shap')[0] for feature in shap_column_names]
+    # Create the feature dropdown filter
+    options = list()
+    options.append(dict(label='All', value='All'))
+    [options.append(dict(label=feat, value=feat)) for feat in feature_names]
+    return options
+
 # Page headers callbacks
 @app.callback(Output('model_perf_header', 'children'),
               [Input('model_name_div', 'children')])
@@ -1664,7 +1680,9 @@ def create_fltd_feat_import_cards(df, data_filter=None):
     if data_filter is None:
         # Use the full dataframe
         cards_list.append(create_feat_import_card(df, card_title='Feature importance',
-                                                  xaxis_title='mean(|SHAP value|)'))
+                                                  max_display=15,
+                                                  xaxis_title='mean(|SHAP value|)',
+                                                  card_height=None))
     else:
         # Filter the data on the specified filter (categorical feature)
         categ_columns = [column for column in df.columns
@@ -1691,7 +1709,9 @@ def create_fltd_feat_import_cards(df, data_filter=None):
                 card_title = f'Feature importance on data with {data_filter} = {categ.split(data_filter)[1]}'
             # Add a feature importance card
             cards_list.append(create_feat_import_card(filtered_df, card_title=card_title,
-                                                      xaxis_title='mean(|SHAP value|)'))
+                                                      max_display=15,
+                                                      xaxis_title='mean(|SHAP value|)',
+                                                      card_height=None))
     return cards_list
 
 @app.callback(Output('feature_importance_cards', 'children'),
@@ -2134,11 +2154,11 @@ def apply_data_changes(new_data, dataset_store, id_column, ts_column, label_colu
         # Guarantee that the model is in evaluation mode, so as to deactivate dropout
         model.eval()
         # Recalculate the SHAP values
-        interpreter = ModelInterpreter(model, df, id_column_name=id_column, inst_column_name=ts_column,
+        interpreter = ModelInterpreter(model, id_column_name=id_column, inst_column_name=ts_column,
                                        label_column_name=label_column, fast_calc=True, 
                                        padding_value=padding_value, is_custom=is_custom,
                                        seq_len_dict=seq_len_dict,
-                                       feat_names=feature_names)
+                                       feat_names=feature_names+[label_column])
         _ = interpreter.interpret_model(test_data=features, test_labels=labels, 
                                         instance_importance=False, feature_importance='shap')
         # Join the updated SHAP values to the dataframe
@@ -2175,10 +2195,15 @@ def calc_exp_val(dataset_mod, model_mod, model_file_name, dataset_name):
     # Load the model
     model = du.deep_learning.load_checkpoint(filepath=f'{models_path}{subdata_path}{model_file_name}.pth', 
                                              ModelClass=model_class)
+    if 'mf1lstm' in model_file_name:
+        # Account for the `delta_ts` column, which isn't directly used as a feature
+        n_inputs = model.n_inputs+1
+    else:
+        n_inputs = model.n_inputs
     # Put the model in evaluation mode to deactivate dropout
     model.eval()
     # Create an all zeroes reference value
-    ref_value = torch.zeros((1, 1, model.n_inputs))
+    ref_value = torch.zeros((1, 1, n_inputs))
     # Calculate the expected value by getting the output from the reference value
     exp_val = model(ref_value).item()
     return exp_val

index.py

@@ -47,4 +47,4 @@
 ])
 
 if __name__ == '__main__':
-    app.run_server(debug=True)
+    app.run_server(debug=False)

layouts.py

@@ -826,15 +826,7 @@
             id='feature_importance_dropdown',
             # [TODO] Add options dynamically, according to the dataset's categorical features
             options=[
-                dict(label='All', value='All'),
-                dict(label='Var0', value='Var0'),
-                dict(label='Var4_a', value='Var4_a'),
-                dict(label='Var4_b', value='Var4_b'),
-                dict(label='Var4_c', value='Var4_c'),
-                # dict(label='Sex', value='Sex'),
-                # dict(label='Age', value='Age'),
-                # dict(label='Diagnostic', value='Diagnostic'),
-                # dict(label='Treatment', value='Treatment')
+                dict(label='All', value='All')
             ],
             placeholder='Choose how to filter the data',
             value='All',