UPD: update tests control risks

mapie/tests/test_control_risk.py

@@ -66,9 +66,6 @@
 
 random_state = 42
 prng = np.random.RandomState(random_state)
-y_1 = prng.random(51)
-y_2 = prng.random((51, 5))
-y_3 = prng.randint(0, 2, 51)
 
 
 def test_compute_recall_equal() -> None:
@@ -86,14 +83,14 @@ def test_compute_precision() -> None:
 def test_recall_with_zero_sum_is_equal_nan() -> None:
  """Test compute_recall with nan values"""
  y_toy = np.zeros((4, 3))
- y_preds_proba = np.random.rand(4, 3, 1)
+ y_preds_proba = prng.rand(4, 3, 1)
  recall = compute_risk_recall(lambdas, y_preds_proba, y_toy)
  np.testing.assert_array_equal(recall, np.empty_like(recall))
 
 
 def test_precision_with_zero_sum_is_equal_ones() -> None:
  """Test compute_precision with nan values"""
- y_toy = np.random.rand(4, 3)
+ y_toy = prng.rand(4, 3)
  y_preds_proba = np.zeros((4, 3, 1))
  precision = compute_risk_precision(lambdas, y_preds_proba, y_toy)
  np.testing.assert_array_equal(precision, np.ones_like(precision))

mapie/tests/test_multi_label_classification.py

@@ -457,7 +457,6 @@ def test_array_output_model(strategy: str, alpha: Any, delta: Any, bound: Any):
  metric_control=args["metric_control"],
  random_state=random_state
  )
-
  mapie_clf.fit(X_toy, y_toy)
  mapie_clf.predict(
  X_toy,
@@ -705,8 +704,6 @@ def test_error_no_fit() -> None:
 def test_error_estimator_none_partial() -> None:
  """Test error estimator none partial"""
  mapie_clf = MapieMultiLabelClassifier(random_state=random_state)
-
-
  with pytest.raises(
  ValueError,
  match=r".*Invalid estimator with partial_fit*"
@@ -739,8 +736,7 @@ def test_toy_dataset_predictions(strategy: str) -> None:
  clf,
  method=args["method"],
  metric_control=args["metric_control"],
- random_state=random_state
-
+ random_state=random_state
  )
  mapie_clf.fit(X_toy, y_toy)
  _, y_ps = mapie_clf.predict(