FIX Make DistanceMetrics support readonly buffers attributes (#21694)

doc/whats_new/v1.0.rst

@@ -28,6 +28,15 @@ Changelog
   and :class:`decomposition.MiniBatchSparsePCA` to be convex and match the referenced
   article. :pr:`19210` by :user:`Jérémie du Boisberranger <jeremiedbb>`.
 
+:mod:`sklearn.metrics`
+......................
+
+- |Fix| All :class:`sklearn.metrics.DistanceMetric` subclasses now correctly support
+  read-only buffer attributes.
+  This fixes a regression introduced in 1.0.0 with respect to 0.24.2.
+  :pr:`21694` by :user:`Julien Jerphanion <jjerphan>`.
+
+
 :mod:`sklearn.preprocessing`
 ............................
 

sklearn/metrics/_dist_metrics.pyx

@@ -29,7 +29,7 @@ cdef DTYPE_t INF = np.inf
 
 from ..utils._typedefs cimport DTYPE_t, ITYPE_t, DITYPE_t, DTYPECODE
 from ..utils._typedefs import DTYPE, ITYPE
-
+from ..utils._readonly_array_wrapper import ReadonlyArrayWrapper
 
 ######################################################################
 # newObj function
@@ -214,8 +214,8 @@ cdef class DistanceMetric:
         set state for pickling
         """
         self.p = state[0]
-        self.vec = state[1]
-        self.mat = state[2]
+        self.vec = ReadonlyArrayWrapper(state[1])
+        self.mat = ReadonlyArrayWrapper(state[2])
         if self.__class__.__name__ == "PyFuncDistance":
             self.func = state[3]
             self.kwargs = state[4]
@@ -444,7 +444,7 @@ cdef class SEuclideanDistance(DistanceMetric):
        D(x, y) = \sqrt{ \sum_i \frac{ (x_i - y_i) ^ 2}{V_i} }
     """
     def __init__(self, V):
-        self.vec = np.asarray(V, dtype=DTYPE)
+        self.vec = ReadonlyArrayWrapper(np.asarray(V, dtype=DTYPE))
         self.size = self.vec.shape[0]
         self.p = 2
 
@@ -605,7 +605,7 @@ cdef class WMinkowskiDistance(DistanceMetric):
             raise ValueError("WMinkowskiDistance requires finite p. "
                              "For p=inf, use ChebyshevDistance.")
         self.p = p
-        self.vec = np.asarray(w, dtype=DTYPE)
+        self.vec = ReadonlyArrayWrapper(np.asarray(w, dtype=DTYPE))
         self.size = self.vec.shape[0]
 
     def _validate_data(self, X):
@@ -665,7 +665,7 @@ cdef class MahalanobisDistance(DistanceMetric):
         if VI.ndim != 2 or VI.shape[0] != VI.shape[1]:
             raise ValueError("V/VI must be square")
 
-        self.mat = np.asarray(VI, dtype=float, order='C')
+        self.mat = ReadonlyArrayWrapper(np.asarray(VI, dtype=float, order='C'))
 
         self.size = self.mat.shape[0]
 

sklearn/metrics/tests/test_dist_metrics.py

@@ -158,11 +158,16 @@ def check_pdist_bool(metric, D_true):
     assert_array_almost_equal(D12, D_true)
 
 
+@pytest.mark.parametrize("use_read_only_kwargs", [True, False])
 @pytest.mark.parametrize("metric", METRICS_DEFAULT_PARAMS)
-def test_pickle(metric):
+def test_pickle(use_read_only_kwargs, metric):
     argdict = METRICS_DEFAULT_PARAMS[metric]
     keys = argdict.keys()
     for vals in itertools.product(*argdict.values()):
+        if use_read_only_kwargs:
+            for val in vals:
+                if isinstance(val, np.ndarray):
+                    val.setflags(write=False)
         kwargs = dict(zip(keys, vals))
         check_pickle(metric, kwargs)
 
@@ -242,3 +247,20 @@ def custom_metric(x, y):
     pyfunc = DistanceMetric.get_metric("pyfunc", func=custom_metric)
     eucl = DistanceMetric.get_metric("euclidean")
     assert_array_almost_equal(pyfunc.pairwise(X), eucl.pairwise(X) ** 2)
+
+
+def test_readonly_kwargs():
+    # Non-regression test for:
+    # https://github.com/scikit-learn/scikit-learn/issues/21685
+
+    rng = check_random_state(0)
+
+    weights = rng.rand(100)
+    VI = rng.rand(10, 10)
+    weights.setflags(write=False)
+    VI.setflags(write=False)
+
+    # Those distances metrics have to support readonly buffers.
+    DistanceMetric.get_metric("seuclidean", V=weights)
+    DistanceMetric.get_metric("wminkowski", p=1, w=weights)
+    DistanceMetric.get_metric("mahalanobis", VI=VI)