mollify ruff

src/inspect_ai/_eval/task/results.py

@@ -18,7 +18,7 @@
 )
 from inspect_ai.log._log import EvalSampleReductions
 from inspect_ai.scorer import Metric, Score, Scorer
-from inspect_ai.scorer._metric import SampleScore, ReducedScore
+from inspect_ai.scorer._metric import ReducedScore, SampleScore
 from inspect_ai.scorer._reducer import ScoreReducer, mean_score, reducer_log_name
 from inspect_ai.scorer._scorer import (
     SCORER_METRICS,

src/inspect_ai/log/_log.py

@@ -28,8 +28,8 @@
 from inspect_ai.util._store import Store
 from inspect_ai.util._store_model import SMT
 
-from ._transcript import Event
 from ..scorer._metric import ReducedScore
+from ._transcript import Event
 
 logger = getLogger(__name__)
 

src/inspect_ai/scorer/_metrics/std.py

@@ -5,7 +5,6 @@
 
 from .._metric import (
     Metric,
-    Score,
     ReducedScore,
     ValueToFloat,
     metric,
@@ -50,7 +49,9 @@ def metric(scores: list[ReducedScore]) -> float:
 
 @metric
 def stderr(to_float: ValueToFloat = value_to_float()) -> Metric:
-    """Clustered standard error of the mean, where each ``ReducedScore``'s children form a cluster.
+    """Clustered standard error of the mean.
+
+    Where each ``ReducedScore``'s children form a cluster.
     If ``epochs=1`` such that each ``ReducedScore`` has only one child, clustered standard errors
     reduce to heteroskedasticity-robust (White) standard errors.
 

src/inspect_ai/scorer/_reducer/reducer.py

@@ -4,7 +4,13 @@
 
 import numpy as np
 
-from inspect_ai.scorer._metric import Score, Value, ValueToFloat, value_to_float, ReducedScore
+from inspect_ai.scorer._metric import (
+    ReducedScore,
+    Score,
+    Value,
+    ValueToFloat,
+    value_to_float,
+)
 
 from .registry import REDUCER_NAME, score_reducer
 from .types import ScoreReducer

src/inspect_ai/scorer/_reducer/types.py

@@ -1,6 +1,6 @@
 from typing import Protocol, runtime_checkable
 
-from .._metric import Score, ReducedScore
+from .._metric import ReducedScore, Score
 
 
 @runtime_checkable

tests/scorer/test_metric.py

@@ -17,7 +17,7 @@
     scorer,
     std,
 )
-from inspect_ai.scorer._metric import metric_create, ReducedScore
+from inspect_ai.scorer._metric import ReducedScore, metric_create
 from inspect_ai.scorer._target import Target
 from inspect_ai.solver._task_state import TaskState
 

tests/scorer/test_metric_stderr.py

@@ -1,9 +1,8 @@
 import pytest
 
-from inspect_ai.scorer._metric import Score, ReducedScore
+from inspect_ai.scorer._metric import ReducedScore, Score
 from inspect_ai.scorer._metrics.std import stderr
 
-
 """
 Comparisons to ``statsmodels`` are done using the following code:
 ```python
@@ -27,9 +26,7 @@ def cluster_se(data: pd.DataFrame) -> float:
 
 
 def test_stderr_single_cluster():
-    """
-    Backward compatibility: previous implementation of stderr returned 0 for a single reduced score.
-    """
+    """Backward compatibility: previous implementation of stderr returned 0 for a single reduced score."""
     scores = [
         ReducedScore(value=2.5, children=[
             Score(value=1.0),
@@ -47,8 +44,8 @@ def test_stderr_single_cluster():
 
 
 def test_stderr_singleton_clusters():
-    """
-    Test clustered SE with three clusters of size 1 each.
+    """Test clustered SE with three clusters of size 1 each.
+
     This should reduce to the heteroskedasticity-robust standard error.
 
     Statsmodels verification:
@@ -112,8 +109,8 @@ def test_stderr_identical_within_varied_between():
 
 
 def test_stderr_1():
-    """
-    Statsmodels verification:
+    """Statsmodels verification.
+
     ```python
     data = pd.DataFrame({
         "y": [1, 1, 0, 0, 1, 0],
@@ -137,8 +134,8 @@ def test_stderr_1():
 
 
 def test_stderr_2():
-    """
-    Statsmodels verification:
+    """Statsmodels verification.
+
     ```python
     data = pd.DataFrame({
         "y": [9.0, 4.0, 11.0, 6.0, 13.0, 8.0],