Update evaluation.py

evaluation.py

@@ -102,6 +102,13 @@ def retrieve_ground_truths(question, dataset):
     for split_name, instances in dataset.items():
         for instance in instances:
             if instance['question'] == question:
+                instance_response = instance['response']
+                ground_truth_metrics = {
+                    "context_relevance": instance['relevance_score'],
+                    "context_utilization": instance['utilization_score'],
+                    "completeness": instance['completeness_score'],
+                    "adherence": instance['adherence_score']
+                }
                 return instance['response']  # Return the ground truth response immediately
     return None  # Return None if no match is found
 
@@ -143,6 +150,8 @@ def calculate_metrics(question, q_dataset, response, docs, time_taken):
         "response_time": time_taken,
         "ground truth completeness": ground_truth_completeness
     }
+
+
     store_rmse(question, predicted_metrics, ground_truth_metrics)
      # Now, make sure the values passed to RMSE calculation are numeric
     predicted_completeness = predicted_metrics['completeness']