Update evaluation.py

evaluation.py

@@ -194,13 +194,31 @@ def calculate_metrics(question, q_dataset, response, docs, time_taken):
     for metric_name in ground_truth_metrics:
         ground_truth_value = ground_truth_metrics[metric_name]     
         print(f"RMSE for {metric_name}: {ground_truth_value}")
-        
+    rmse_values = []
+    ground_truth_values = []
+    for metric_name in predicted_metrics_rmse:
+        predicted_value = predicted_metrics_rmse[metric_name]
+        ground_truth_value = ground_truth_metrics.get(metric_name, None)
+
+        # Ensure both predicted and ground truth values are numeric
+        if isinstance(predicted_value, (int, float)) and isinstance(ground_truth_value, (int, float)):
+            rmse_values.append(predicted_value)
+            ground_truth_values.append(ground_truth_value)
+        else:
+            print(f"Skipping RMSE for {metric_name}: One or both values are non-numeric")
+
+    if rmse_values and ground_truth_values:
+        overall_rmse = compute_rmse(rmse_values, ground_truth_values)
+        print(f"Overall RMSE: {overall_rmse}")
+    else:
+        print("Invalid RMSE calculation due to non-numeric values.")
+    '''    
     if isinstance(predicted_metrics_rmse, (int, float)) and isinstance(ground_truth_metrics, (int, float)):
         rmse_value = compute_rmse(predicted_metrics_rmse.values(), ground_truth_metrics.values())
         predicted_metrics_rmse["rmse"] = rmse_value  # Adding RMSE to metrics
     else:
         predicted_metrics_rmse["rmse"] = "Invalid RMSE calculation"
-        
+     '''   
     return predicted_metrics
 
 ''' def retrieve_ground_truths(question, dataset):