add random state to svd in _compute_pi function to make examples deterministic

Author: agoscinski

src/skmatter/_selection.py

@@ -649,14 +649,15 @@ def _compute_pi(self, X, y=None):
             :math:`\\pi` importance for the given samples or features
         """
 
+        svd_kwargs = dict(k=self.k, random_state=self.random_state)
         if self._axis == 0:
-            U, _, _ = scipy.sparse.linalg.svds(X, k=self.k, return_singular_vectors="u")
+            svd_kwargs["return_singular_vectors"] = "u"
+            U, _, _ = scipy.sparse.linalg.svds(X, **svd_kwargs)
             U = np.real(U)
             new_pi = (U[:, : self.k] ** 2.0).sum(axis=1)
         else:
-            _, _, Vt = scipy.sparse.linalg.svds(
-                X, k=self.k, return_singular_vectors="vh"
-            )
+            svd_kwargs["return_singular_vectors"] = "vh"
+            _, _, Vt = scipy.sparse.linalg.svds(X, **svd_kwargs)
             new_pi = (np.real(Vt) ** 2.0).sum(axis=0)
 
         return new_pi

src/skmatter/sample_selection/_base.py

@@ -342,7 +342,6 @@ class CUR(_CUR):
     ...         [-0.03, -0.53, 0.08],
     ...     ]
     ... )
-    >>> np.random.seed(0)  # there is a source of randomness in it
     >>> selector.fit(X)
     CUR(n_to_select=2)
     >>> np.round(selector.pi_, 2)  # importance scole