fix: update type hint for OneOrMoreScores

py/src/braintrust/framework.py

@@ -216,7 +216,7 @@ class EvalScorerArgs(SerializableDataClass, Generic[Input, Output, Expected]):
     metadata: Metadata | None = None
 
 
-OneOrMoreScores = float | int | bool | None | ScoreLike | list[ScoreLike]
+OneOrMoreScores = float | int | bool | None | ScoreLike | Sequence[ScoreLike]
 OneOrMoreClassifications = None | Classification | Mapping[str, Any] | list[Classification | Mapping[str, Any]]
 
 

py/src/braintrust/type_tests/test_autoevals_scorers.py

@@ -2,7 +2,7 @@
 
 import pytest
 from autoevals import Levenshtein  # type: ignore[import-untyped]
-from braintrust.framework import Eval, EvalAsync, EvalCase, EvalScorer
+from braintrust.framework import Eval, EvalAsync, EvalCase, EvalScorer, Score
 
 
 def accepts_autoevals_scorer(
@@ -36,6 +36,25 @@ async def autoevals_task_async(input: str) -> str:
 ]
 
 
+def test_eval_accepts_autoevals_scorers_typed_sequence():
+    def scorer(input: str, output: str, expected: str | None = None) -> list[Score]:
+        return [Score(name="match", score=1.0)]
+
+    typed_scorer: EvalScorer[str, str, str] = scorer
+
+    result = Eval(
+        "test-autoevals-scorers",
+        data=autoevals_data,
+        task=autoevals_task,
+        scores=[typed_scorer],
+        no_send_logs=True,
+    )
+
+    score = result.results[0].scores["match"]
+    assert score is not None
+    assert score > 0
+
+
 def test_eval_accepts_autoevals_scorers_typed():
     result = Eval(
         "test-autoevals-scorers",