TST Enhance ROC Curve Display Tests for Improved Clarity and Maintainability (scikit-learn#31266)

Author: NEREUScode

sklearn/metrics/_plot/tests/test_roc_curve_display.py

@@ -5,7 +5,7 @@
 
 from sklearn import clone
 from sklearn.compose import make_column_transformer
-from sklearn.datasets import load_breast_cancer, load_iris
+from sklearn.datasets import load_breast_cancer, make_classification
 from sklearn.exceptions import NotFittedError
 from sklearn.linear_model import LogisticRegression
 from sklearn.metrics import RocCurveDisplay, auc, roc_curve
@@ -16,20 +16,19 @@
 
 
 @pytest.fixture(scope="module")
-def data():
-    X, y = load_iris(return_X_y=True)
-    # Avoid introducing test dependencies by mistake.
-    X.flags.writeable = False
-    y.flags.writeable = False
+def data_binary():
+    X, y = make_classification(
+        n_samples=200,
+        n_features=20,
+        n_informative=5,
+        n_redundant=2,
+        flip_y=0.1,
+        class_sep=0.8,
+        random_state=42,
+    )
     return X, y
 
 
-@pytest.fixture(scope="module")
-def data_binary(data):
-    X, y = data
-    return X[y < 2], y[y < 2]
-
-
 @pytest.mark.parametrize("response_method", ["predict_proba", "decision_function"])
 @pytest.mark.parametrize("with_sample_weight", [True, False])
 @pytest.mark.parametrize("drop_intermediate", [True, False])