Scikit-learn 1.7 fixes

examples/regression/Ridge2FoldCVRegularization.py

@@ -203,12 +203,12 @@ def get_train_test_error(estimator):
     RidgeCV(
         alphas=alphas,
         cv=None,
-        store_cv_values=True,
+        store_cv_results=True,
         scoring=None,  # uses by default mean squared error
         fit_intercept=False,
     )
     .fit(X_train, y_train)
-    .cv_values_
+    .cv_results_
 )
 
 results["sklearn LOO CV Tikhonov"]["MSE validation"] = np.mean(

pyproject.toml

@@ -38,7 +38,7 @@ classifiers = [
     "Topic :: Scientific/Engineering",
 ]
 dependencies = [
-    "scikit-learn >= 1.6, < 1.7",
+    "scikit-learn >= 1.7, < 1.8",
     "scipy >= 1.15",  # explicit to adhere to scikit-learn dependencies
 ]
 dynamic = ["version"]

src/skmatter/linear_model/_ridge.py

@@ -303,7 +303,7 @@ def _2fold_loss_tikhonov(alpha):
             return ((Vt.T[:, :n_alpha] / s[:n_alpha]) @ (U.T[:n_alpha] @ y)).T
 
 
-class _IdentityRegressor:
+class _IdentityRegressor(BaseEstimator):
     """Fake regressor which will directly output the prediction."""
 
     def predict(self, y_predict):

src/skmatter/sample_selection/_base.py

@@ -6,6 +6,7 @@
 from scipy.interpolate import LinearNDInterpolator, interp1d
 from scipy.interpolate._interpnd import _ndim_coords_from_arrays
 from scipy.spatial import ConvexHull
+from sklearn.base import BaseEstimator
 from sklearn.utils.validation import check_array, check_is_fitted, check_X_y
 
 from .._selection import _CUR, _FPS, _PCovCUR, _PCovFPS
@@ -479,7 +480,7 @@ def _directional_distance(equations, points):
     return -orthogonal_distances / equations[:, :1].T
 
 
-class DirectionalConvexHull:
+class DirectionalConvexHull(BaseEstimator):
     """
     Performs Sample Selection by constructing a Directional Convex Hull and determining
     the distance to the hull as outlined in the reference [dch]_.