MAINT Parameters validation for sklearn.metrics.pairwise_distances_chunked (scikit-learn#26125)

Co-authored-by: jeremiedbb <jeremiedbb@yahoo.fr>

Author: Charlie-XIAO

sklearn/metrics/pairwise.py

@@ -1883,6 +1883,17 @@ def _precompute_metric_params(X, Y, metric=None, **kwds):
     return {}
 
 
+@validate_params(
+    {
+        "X": ["array-like", "sparse matrix"],
+        "Y": ["array-like", "sparse matrix", None],
+        "reduce_func": [callable, None],
+        "metric": [StrOptions({"precomputed"}.union(_VALID_METRICS)), callable],
+        "n_jobs": [Integral, None],
+        "working_memory": [Interval(Real, 0, None, closed="left"), None],
+    },
+    prefer_skip_nested_validation=False,  # metric is not validated yet
+)
 def pairwise_distances_chunked(
     X,
     Y=None,
@@ -1903,13 +1914,13 @@ def pairwise_distances_chunked(
 
     Parameters
     ----------
-    X : ndarray of shape (n_samples_X, n_samples_X) or \
+    X : {array-like, sparse matrix} of shape (n_samples_X, n_samples_X) or \
             (n_samples_X, n_features)
         Array of pairwise distances between samples, or a feature array.
         The shape the array should be (n_samples_X, n_samples_X) if
         metric='precomputed' and (n_samples_X, n_features) otherwise.
 
-    Y : ndarray of shape (n_samples_Y, n_features), default=None
+    Y : {array-like, sparse matrix} of shape (n_samples_Y, n_features), default=None
         An optional second feature array. Only allowed if
         metric != "precomputed".
 
@@ -1946,7 +1957,7 @@ def pairwise_distances_chunked(
         ``-1`` means using all processors. See :term:`Glossary <n_jobs>`
         for more details.
 
-    working_memory : int, default=None
+    working_memory : float, default=None
         The sought maximum memory for temporary distance matrix chunks.
         When None (default), the value of
         ``sklearn.get_config()['working_memory']`` is used.

sklearn/metrics/tests/test_pairwise.py

@@ -748,9 +748,6 @@ def test_pairwise_distances_chunked(global_dtype):
     # "cityblock" uses scikit-learn metric, cityblock (function) is
     # scipy.spatial.
     check_pairwise_distances_chunked(X, Y, working_memory=1, metric="cityblock")
-    # Test that a value error is raised if the metric is unknown
-    with pytest.raises(ValueError):
-        next(pairwise_distances_chunked(X, Y, metric="blah"))
 
     # Test precomputed returns all at once
     D = pairwise_distances(X)

sklearn/tests/test_public_functions.py

@@ -254,6 +254,7 @@ def _check_function_param_validation(
     "sklearn.metrics.pairwise.sigmoid_kernel",
     "sklearn.metrics.pairwise_distances",
     "sklearn.metrics.pairwise_distances_argmin",
+    "sklearn.metrics.pairwise_distances_chunked",
     "sklearn.metrics.precision_recall_curve",
     "sklearn.metrics.precision_recall_fscore_support",
     "sklearn.metrics.precision_score",