change to 0.24.0

requirements.txt

@@ -1,3 +1,4 @@
 numpy
 scipy
-scikit-learn>=0.19.0
+scikit-learn>=0.24.0
+joblib

spmimage/decomposition/ksvd.py

@@ -2,7 +2,7 @@
 
 import numpy as np
 from sklearn.base import BaseEstimator
-from sklearn.decomposition.dict_learning import SparseCodingMixin, sparse_encode
+from sklearn.decomposition._dict_learning import _BaseSparseCoding, sparse_encode
 from sklearn.utils import check_array
 from sklearn.utils.validation import check_is_fitted
 
@@ -116,7 +116,7 @@ def _ksvd(Y: np.ndarray, n_components: int, n_nonzero_coefs: int, max_iter: int,
     return W, H, errors, k + 1
 
 
-class KSVD(BaseEstimator, SparseCodingMixin):
+class KSVD(BaseEstimator, _BaseSparseCoding):
     """ K-SVD
     Finds a dictionary that can be used to represent data using a sparse code.
     Solves the optimization problem:

spmimage/linear_model/admm.py

@@ -5,8 +5,8 @@
 
 from sklearn.utils import check_array, check_X_y
 from sklearn.base import RegressorMixin
-from sklearn.linear_model.base import LinearModel
-from sklearn.linear_model.coordinate_descent import _alpha_grid
+from sklearn.linear_model._base import LinearModel
+from sklearn.linear_model._coordinate_descent import _alpha_grid
 from joblib import Parallel, delayed
 
 logger = getLogger(__name__)