Refactor SemF1 aggregation logic and fix typo in comment

semf1.py

@@ -422,8 +422,8 @@ class SemF1(evaluate.Metric):
             recall_scores = [np.clip(r_scores, 0.0, 1.0).item() for (r_scores, _) in recall_scores]
 
             results.append(Scores(precision, recall_scores))
-        
-        # runn aggregation procedure
+
+        # run aggregation procedure
         if aggregate:
             mean_prec = np.mean(
                 [score.precision for score in results]
@@ -432,12 +432,9 @@ class SemF1(evaluate.Metric):
                 [np.array(score.recall) for score in results]
             ))
             aggregated_score = Scores(
-                float(mean_prec), 
+                float(mean_prec),
                 [float(mean_recall)]
             )
-            aggregated_score.f1 = float(np.mean(
-                [score.f1 for score in results]
-            ))
             results = aggregated_score
 
-        return results
+        return results

tests.py

@@ -708,5 +708,6 @@ class TestValidateInputFormat(unittest.TestCase):
 def run_tests():
     unittest.main(verbosity=2)
 
+
 if __name__ == '__main__':
     run_tests()