Apply comment and Ruff

src/lighteval/metrics/harness_compatibility/drop.py

@@ -22,10 +22,10 @@
 
 import re
 import string
+from typing import List, Set, Tuple
 
 import numpy as np
 from scipy.optimize import linear_sum_assignment
-from typing import List, Set, Tuple
 
 
 def drop_metrics(predictions: list[str], formatted_doc, **kwargs):  # noqa: C901
@@ -41,7 +41,7 @@ def drop_metrics(predictions: list[str], formatted_doc, **kwargs):  # noqa: C901
     between prediction and the different answers is taken.
 
     For more information, please refer to the section 5 of the DROP paper (https://aclanthology.org/N19-1246/).
-    
+
     Todo: this code is really hard to follow, simplify when possible
     """
 
@@ -69,7 +69,9 @@ def _get_metrics(predicted: List[str], gold: List[str]):
         pred_normalized_spans, pred_bags = _answer_to_bags(predicted)
         gold_normalized_spans, gold_bags = _answer_to_bags(gold)
 
-        if set(pred_normalized_spans) == set(gold_normalized_spans) and len(gold_normalized_spans) == len(gold_normalized_spans):
+        if set(pred_normalized_spans) == set(gold_normalized_spans) and len(gold_normalized_spans) == len(
+            gold_normalized_spans
+        ):
             exact_match = 1.0
         else:
             exact_match = 0.0
@@ -161,7 +163,9 @@ def _normalize(answer: str):
     max_f1 = 0
     for gold_answer in formatted_doc.specific["golds_no_preprocessing"]:
         exact_match, f1_score = _get_metrics(predictions, gold_answer)
-        if gold_answer[0].strip():
+        if isinstance(gold_answer, list):
+            gold_answer = gold_answer[0]
+        if gold_answer.strip():
             max_em = max(max_em, exact_match)
             max_f1 = max(max_f1, f1_score)
     return {"qem": max_em, "f1": max_f1}