scikit-learn-intelex 2024.3.0__py310-none-manylinux1_x86_64.whl → 2024.4.0__py310-none-manylinux1_x86_64.whl

sklearnex/linear_model/logistic_regression.py

@@ -38,19 +38,27 @@ if daal_check_version((2024, "P", 1)):
     import numpy as np
     from scipy.sparse import issparse
     from sklearn.linear_model import LogisticRegression as sklearn_LogisticRegression
+    from sklearn.metrics import accuracy_score
+    from sklearn.utils.multiclass import type_of_target
     from sklearn.utils.validation import check_X_y
 
     from daal4py.sklearn._n_jobs_support import control_n_jobs
     from daal4py.sklearn._utils import sklearn_check_version
     from onedal.linear_model import LogisticRegression as onedal_LogisticRegression
-    from onedal.utils import _num_features, _num_samples
+    from onedal.utils import _num_samples
 
     from .._device_offload import dispatch, wrap_output_data
     from .._utils import PatchingConditionsChain, get_patch_message
     from ..utils.validation import _assert_all_finite
 
     @control_n_jobs(
-        decorated_methods=["fit", "predict", "predict_proba", "predict_log_proba"]
+        decorated_methods=[
+            "fit",
+            "predict",
+            "predict_proba",
+            "predict_log_proba",
+            "score",
+        ]
     )
     class LogisticRegression(sklearn_LogisticRegression, BaseLogisticRegression):
         __doc__ = sklearn_LogisticRegression.__doc__
@@ -72,9 +80,9 @@ if daal_check_version((2024, "P", 1)):
             intercept_scaling=1,
             class_weight=None,
             random_state=None,
-            solver="lbfgs" if sklearn_check_version("0.22") else "liblinear",
+            solver="lbfgs",
             max_iter=100,
-            multi_class="auto" if sklearn_check_version("0.22") else "ovr",
+            multi_class="auto",
             verbose=0,
             warm_start=False,
             n_jobs=None,
@@ -160,6 +168,27 @@ if daal_check_version((2024, "P", 1)):
                 X,
             )
 
+        @wrap_output_data
+        def score(self, X, y, sample_weight=None):
+            if sklearn_check_version("1.0"):
+                self._check_feature_names(X, reset=False)
+            return dispatch(
+                self,
+                "score",
+                {
+                    "onedal": self.__class__._onedal_score,
+                    "sklearn": sklearn_LogisticRegression.score,
+                },
+                X,
+                y,
+                sample_weight=sample_weight,
+            )
+
+        def _onedal_score(self, X, y, sample_weight=None, queue=None):
+            return accuracy_score(
+                y, self._onedal_predict(X, queue=queue), sample_weight=sample_weight
+            )
+
         def _test_type_and_finiteness(self, X_in):
             X = np.asarray(X_in)
 
@@ -198,6 +227,10 @@ if daal_check_version((2024, "P", 1)):
                     (self.warm_start == False, "Warm start is not supported."),
                     (self.l1_ratio is None, "l1 ratio is not supported."),
                     (sample_weight is None, "Sample weight is not supported."),
+                    (
+                        type_of_target(y) == "binary",
+                        "Only binary classification is supported",
+                    ),
                 ]
             )
 
@@ -216,22 +249,29 @@ if daal_check_version((2024, "P", 1)):
             return patching_status
 
         def _onedal_gpu_predict_supported(self, method_name, *data):
-            assert method_name in ["predict", "predict_proba", "predict_log_proba"]
-            assert len(data) == 1
+            assert method_name in [
+                "predict",
+                "predict_proba",
+                "predict_log_proba",
+                "score",
+            ]
 
             class_name = self.__class__.__name__
             patching_status = PatchingConditionsChain(
                 f"sklearn.linear_model.{class_name}.{method_name}"
             )
 
-            n_samples = _num_samples(*data)
+            n_samples = _num_samples(data[0])
             model_is_sparse = issparse(self.coef_) or (
                 self.fit_intercept and issparse(self.intercept_)
             )
             dal_ready = patching_status.and_conditions(
                 [
                     (n_samples > 0, "Number of samples is less than 1."),
-                    (not issparse(*data), "Sparse input is not supported."),
+                    (
+                        not any([issparse(i) for i in data]),
+                        "Sparse input is not supported.",
+                    ),
                     (not model_is_sparse, "Sparse coefficients are not supported."),
                     (
                         hasattr(self, "_onedal_estimator"),
@@ -251,7 +291,7 @@ if daal_check_version((2024, "P", 1)):
         def _onedal_gpu_supported(self, method_name, *data):
             if method_name == "fit":
                 return self._onedal_gpu_fit_supported(method_name, *data)
-            if method_name in ["predict", "predict_proba", "predict_log_proba"]:
+            if method_name in ["predict", "predict_proba", "predict_log_proba", "score"]:
                 return self._onedal_gpu_predict_supported(method_name, *data)
             raise RuntimeError(
                 f"Unknown method {method_name} in {self.__class__.__name__}"
@@ -334,6 +374,7 @@ if daal_check_version((2024, "P", 1)):
         predict.__doc__ = sklearn_LogisticRegression.predict.__doc__
         predict_proba.__doc__ = sklearn_LogisticRegression.predict_proba.__doc__
         predict_log_proba.__doc__ = sklearn_LogisticRegression.predict_log_proba.__doc__
+        score.__doc__ = sklearn_LogisticRegression.score.__doc__
 
 else:
     LogisticRegression = LogisticRegression_daal4py

sklearnex/linear_model/tests/test_linear.py

@@ -28,26 +28,33 @@ from onedal.tests.utils._dataframes_support import (
 
 
 @pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
 @pytest.mark.parametrize("macro_block", [None, 1024])
-def test_sklearnex_import_linear(dataframe, queue, macro_block):
+def test_sklearnex_import_linear(dataframe, queue, dtype, macro_block):
     from sklearnex.linear_model import LinearRegression
 
     X = np.array([[1, 1], [1, 2], [2, 2], [2, 3]])
     y = np.dot(X, np.array([1, 2])) + 3
+    X = X.astype(dtype=dtype)
+    y = y.astype(dtype=dtype)
     X = _convert_to_dataframe(X, sycl_queue=queue, target_df=dataframe)
     y = _convert_to_dataframe(y, sycl_queue=queue, target_df=dataframe)
+
     linreg = LinearRegression()
     if daal_check_version((2024, "P", 0)) and macro_block is not None:
         hparams = linreg.get_hyperparameters("fit")
         hparams.cpu_macro_block = macro_block
         hparams.gpu_macro_block = macro_block
+
     linreg.fit(X, y)
-    if daal_check_version((2023, "P", 100)):
-        assert hasattr(linreg, "_onedal_estimator")
+
+    assert hasattr(linreg, "_onedal_estimator")
     assert "sklearnex" in linreg.__module__
     assert linreg.n_features_in_ == 2
-    assert_allclose(_as_numpy(linreg.intercept_), 3.0)
-    assert_allclose(_as_numpy(linreg.coef_), [1.0, 2.0])
+
+    tol = 1e-5 if X.dtype == np.float32 else 1e-7
+    assert_allclose(_as_numpy(linreg.intercept_), 3.0, rtol=tol)
+    assert_allclose(_as_numpy(linreg.coef_), [1.0, 2.0], rtol=tol)
 
 
 def test_sklearnex_import_ridge():
@@ -80,3 +87,31 @@ def test_sklearnex_import_elastic():
     assert "daal4py" in elasticnet.__module__
     assert_allclose(elasticnet.intercept_, 1.451, atol=1e-3)
     assert_allclose(elasticnet.coef_, [18.838, 64.559], atol=1e-3)
+
+
+@pytest.mark.parametrize("dataframe,queue", get_dataframes_and_queues())
+@pytest.mark.parametrize("dtype", [np.float32, np.float64])
+def test_sklearnex_reconstruct_model(dataframe, queue, dtype):
+    from sklearnex.linear_model import LinearRegression
+
+    seed = 42
+    num_samples = 3500
+    num_features, num_targets = 14, 9
+
+    gen = np.random.default_rng(seed)
+    intercept = gen.random(size=num_targets, dtype=dtype)
+    coef = gen.random(size=(num_targets, num_features), dtype=dtype).T
+
+    X = gen.random(size=(num_samples, num_features), dtype=dtype)
+    gtr = X @ coef + intercept[np.newaxis, :]
+
+    X = _convert_to_dataframe(X, sycl_queue=queue, target_df=dataframe)
+
+    linreg = LinearRegression(fit_intercept=True)
+    linreg.coef_ = coef.T
+    linreg.intercept_ = intercept
+
+    y_pred = linreg.predict(X)
+
+    tol = 1e-5 if X.dtype == np.float32 else 1e-7
+    assert_allclose(gtr, _as_numpy(y_pred), rtol=tol)

sklearnex/neighbors/_lof.py

@@ -23,13 +23,13 @@ from sklearn.utils.validation import check_is_fitted
 
 from daal4py.sklearn._n_jobs_support import control_n_jobs
 from daal4py.sklearn._utils import sklearn_check_version
+from sklearnex._device_offload import dispatch, wrap_output_data
+from sklearnex.neighbors.common import KNeighborsDispatchingBase
+from sklearnex.neighbors.knn_unsupervised import NearestNeighbors
+from sklearnex.utils import get_namespace
 
-from .._device_offload import dispatch, wrap_output_data
-from .common import KNeighborsDispatchingBase
-from .knn_unsupervised import NearestNeighbors
 
-
-@control_n_jobs(decorated_methods=["fit", "kneighbors"])
+@control_n_jobs(decorated_methods=["fit", "_kneighbors"])
 class LocalOutlierFactor(KNeighborsDispatchingBase, sklearn_LocalOutlierFactor):
     __doc__ = (
         sklearn_LocalOutlierFactor.__doc__
@@ -100,7 +100,6 @@ class LocalOutlierFactor(KNeighborsDispatchingBase, sklearn_LocalOutlierFactor):
         return self
 
     def fit(self, X, y=None):
-        self._fit_validation(X, y)
         result = dispatch(
             self,
             "fit",
@@ -113,16 +112,13 @@ class LocalOutlierFactor(KNeighborsDispatchingBase, sklearn_LocalOutlierFactor):
         )
         return result
 
-    # Subtle order change to remove check_array and preserve dpnp and
-    # dpctl conformance. decision_function will return a dpnp or dpctl
-    # instance via kneighbors and an equivalent check_array exists in
-    # that call already in sklearn so no loss of functionality occurs
     def _predict(self, X=None):
         check_is_fitted(self)
 
         if X is not None:
+            xp, _ = get_namespace(X)
             output = self.decision_function(X) < 0
-            is_inlier = np.ones(output.shape[0], dtype=int)
+            is_inlier = xp.ones_like(output, dtype=int)
             is_inlier[output] = -1
         else:
             is_inlier = np.ones(self.n_samples_fit_, dtype=int)
@@ -159,16 +155,40 @@ class LocalOutlierFactor(KNeighborsDispatchingBase, sklearn_LocalOutlierFactor):
         """
         return self.fit(X)._predict()
 
-    @available_if(sklearn_LocalOutlierFactor._check_novelty_predict)
+    def _kneighbors(self, X=None, n_neighbors=None, return_distance=True):
+        check_is_fitted(self)
+        if sklearn_check_version("1.0") and X is not None:
+            self._check_feature_names(X, reset=False)
+        return dispatch(
+            self,
+            "kneighbors",
+            {
+                "onedal": self.__class__._onedal_kneighbors,
+                "sklearn": sklearn_LocalOutlierFactor.kneighbors,
+            },
+            X,
+            n_neighbors=n_neighbors,
+            return_distance=return_distance,
+        )
+
+    kneighbors = wrap_output_data(_kneighbors)
+
+    @available_if(sklearn_LocalOutlierFactor._check_novelty_score_samples)
     @wrap_output_data
-    def predict(self, X=None):
-        """Predict the labels (1 inlier, -1 outlier) of X according to LOF.
+    def score_samples(self, X):
+        """Opposite of the Local Outlier Factor of X.
+
+        It is the opposite as bigger is better, i.e. large values correspond
+        to inliers.
 
         **Only available for novelty detection (when novelty is set to True).**
-        This method allows to generalize prediction to *new observations* (not
-        in the training set). Note that the result of ``clf.fit(X)`` then
-        ``clf.predict(X)`` with ``novelty=True`` may differ from the result
-        obtained by ``clf.fit_predict(X)`` with ``novelty=False``.
+        The argument X is supposed to contain *new data*: if X contains a
+        point from training, it considers the later in its own neighborhood.
+        Also, the samples in X are not considered in the neighborhood of any
+        point. Because of this, the scores obtained via ``score_samples`` may
+        differ from the standard LOF scores.
+        The standard LOF scores for the training data is available via the
+        ``negative_outlier_factor_`` attribute.
 
         Parameters
         ----------
@@ -178,27 +198,24 @@ class LocalOutlierFactor(KNeighborsDispatchingBase, sklearn_LocalOutlierFactor):
 
         Returns
         -------
-        is_inlier : ndarray of shape (n_samples,)
-            Returns -1 for anomalies/outliers and +1 for inliers.
+        opposite_lof_scores : ndarray of shape (n_samples,)
+            The opposite of the Local Outlier Factor of each input samples.
+            The lower, the more abnormal.
         """
-        return self._predict(X)
-
-    @wrap_output_data
-    def kneighbors(self, X=None, n_neighbors=None, return_distance=True):
         check_is_fitted(self)
-        if sklearn_check_version("1.0") and X is not None:
-            self._check_feature_names(X, reset=False)
-        return dispatch(
-            self,
-            "kneighbors",
-            {
-                "onedal": self.__class__._onedal_kneighbors,
-                "sklearn": sklearn_LocalOutlierFactor.kneighbors,
-            },
-            X,
-            n_neighbors=n_neighbors,
-            return_distance=return_distance,
+
+        distances_X, neighbors_indices_X = self._kneighbors(
+            X, n_neighbors=self.n_neighbors_
         )
 
+        X_lrd = self._local_reachability_density(
+            distances_X,
+            neighbors_indices_X,
+        )
+
+        lrd_ratios_array = self._lrd[neighbors_indices_X] / X_lrd[:, np.newaxis]
+
+        return -np.mean(lrd_ratios_array, axis=1)
+
     fit.__doc__ = sklearn_LocalOutlierFactor.fit.__doc__
     kneighbors.__doc__ = sklearn_LocalOutlierFactor.kneighbors.__doc__

sklearnex/neighbors/common.py

@@ -137,6 +137,9 @@ class KNeighborsDispatchingBase:
             self.n_features_in_ = X.data.shape[1]
 
     def _onedal_supported(self, device, method_name, *data):
+        if method_name == "fit":
+            self._fit_validation(data[0], data[1])
+
         class_name = self.__class__.__name__
         is_classifier = "Classifier" in class_name
         is_regressor = "Regressor" in class_name
@@ -249,7 +252,7 @@ class KNeighborsDispatchingBase:
                     class_count >= 2, "One-class case is not supported."
                 )
             return patching_status
-        if method_name in ["predict", "predict_proba", "kneighbors"]:
+        if method_name in ["predict", "predict_proba", "kneighbors", "score"]:
             patching_status.and_condition(
                 hasattr(self, "_onedal_estimator"), "oneDAL model was not trained."
             )

sklearnex/neighbors/knn_classification.py

@@ -14,129 +14,30 @@
 # limitations under the License.
 # ===============================================================================
 
-from daal4py.sklearn._n_jobs_support import control_n_jobs
-from daal4py.sklearn._utils import sklearn_check_version
-
-if not sklearn_check_version("1.2"):
-    from sklearn.neighbors._base import _check_weights
-
+from sklearn.metrics import accuracy_score
 from sklearn.neighbors._classification import (
     KNeighborsClassifier as sklearn_KNeighborsClassifier,
 )
 from sklearn.neighbors._unsupervised import NearestNeighbors as sklearn_NearestNeighbors
 from sklearn.utils.validation import _deprecate_positional_args, check_is_fitted
 
+from daal4py.sklearn._n_jobs_support import control_n_jobs
+from daal4py.sklearn._utils import sklearn_check_version
 from onedal.neighbors import KNeighborsClassifier as onedal_KNeighborsClassifier
 
 from .._device_offload import dispatch, wrap_output_data
 from .common import KNeighborsDispatchingBase
 
-if sklearn_check_version("0.24"):
-
-    class KNeighborsClassifier_(sklearn_KNeighborsClassifier):
-        if sklearn_check_version("1.2"):
-            _parameter_constraints: dict = {
-                **sklearn_KNeighborsClassifier._parameter_constraints
-            }
 
-        @_deprecate_positional_args
-        def __init__(
-            self,
-            n_neighbors=5,
-            *,
-            weights="uniform",
-            algorithm="auto",
-            leaf_size=30,
-            p=2,
-            metric="minkowski",
-            metric_params=None,
-            n_jobs=None,
-            **kwargs,
-        ):
-            super().__init__(
-                n_neighbors=n_neighbors,
-                algorithm=algorithm,
-                leaf_size=leaf_size,
-                metric=metric,
-                p=p,
-                metric_params=metric_params,
-                n_jobs=n_jobs,
-                **kwargs,
-            )
-            self.weights = (
-                weights if sklearn_check_version("1.0") else _check_weights(weights)
-            )
-
-elif sklearn_check_version("0.22"):
-    from sklearn.neighbors._base import (
-        SupervisedIntegerMixin as BaseSupervisedIntegerMixin,
-    )
-
-    class KNeighborsClassifier_(sklearn_KNeighborsClassifier, BaseSupervisedIntegerMixin):
-        @_deprecate_positional_args
-        def __init__(
-            self,
-            n_neighbors=5,
-            *,
-            weights="uniform",
-            algorithm="auto",
-            leaf_size=30,
-            p=2,
-            metric="minkowski",
-            metric_params=None,
-            n_jobs=None,
-            **kwargs,
-        ):
-            super().__init__(
-                n_neighbors=n_neighbors,
-                algorithm=algorithm,
-                leaf_size=leaf_size,
-                metric=metric,
-                p=p,
-                metric_params=metric_params,
-                n_jobs=n_jobs,
-                **kwargs,
-            )
-            self.weights = _check_weights(weights)
-
-else:
-    from sklearn.neighbors.base import (
-        SupervisedIntegerMixin as BaseSupervisedIntegerMixin,
-    )
-
-    class KNeighborsClassifier_(sklearn_KNeighborsClassifier, BaseSupervisedIntegerMixin):
-        @_deprecate_positional_args
-        def __init__(
-            self,
-            n_neighbors=5,
-            *,
-            weights="uniform",
-            algorithm="auto",
-            leaf_size=30,
-            p=2,
-            metric="minkowski",
-            metric_params=None,
-            n_jobs=None,
-            **kwargs,
-        ):
-            super().__init__(
-                n_neighbors=n_neighbors,
-                algorithm=algorithm,
-                leaf_size=leaf_size,
-                metric=metric,
-                p=p,
-                metric_params=metric_params,
-                n_jobs=n_jobs,
-                **kwargs,
-            )
-            self.weights = _check_weights(weights)
-
-
-@control_n_jobs(decorated_methods=["fit", "predict", "predict_proba", "kneighbors"])
-class KNeighborsClassifier(KNeighborsClassifier_, KNeighborsDispatchingBase):
+@control_n_jobs(
+    decorated_methods=["fit", "predict", "predict_proba", "kneighbors", "score"]
+)
+class KNeighborsClassifier(sklearn_KNeighborsClassifier, KNeighborsDispatchingBase):
     __doc__ = sklearn_KNeighborsClassifier.__doc__
     if sklearn_check_version("1.2"):
-        _parameter_constraints: dict = {**KNeighborsClassifier_._parameter_constraints}
+        _parameter_constraints: dict = {
+            **sklearn_KNeighborsClassifier._parameter_constraints
+        }
 
     if sklearn_check_version("1.0"):
 
@@ -192,7 +93,6 @@ class KNeighborsClassifier(KNeighborsClassifier_, KNeighborsDispatchingBase):
             )
 
     def fit(self, X, y):
-        self._fit_validation(X, y)
         dispatch(
             self,
             "fit",
@@ -235,6 +135,23 @@ class KNeighborsClassifier(KNeighborsClassifier_, KNeighborsDispatchingBase):
             X,
         )
 
+    @wrap_output_data
+    def score(self, X, y, sample_weight=None):
+        check_is_fitted(self)
+        if sklearn_check_version("1.0"):
+            self._check_feature_names(X, reset=False)
+        return dispatch(
+            self,
+            "score",
+            {
+                "onedal": self.__class__._onedal_score,
+                "sklearn": sklearn_KNeighborsClassifier.score,
+            },
+            X,
+            y,
+            sample_weight=sample_weight,
+        )
+
     @wrap_output_data
     def kneighbors(self, X=None, n_neighbors=None, return_distance=True):
         check_is_fitted(self)
@@ -263,18 +180,10 @@ class KNeighborsClassifier(KNeighborsClassifier_, KNeighborsDispatchingBase):
             or getattr(self, "_tree", 0) is None
             and self._fit_method == "kd_tree"
         ):
-            if sklearn_check_version("0.24"):
-                sklearn_NearestNeighbors.fit(self, self._fit_X, getattr(self, "_y", None))
-            else:
-                sklearn_NearestNeighbors.fit(self, self._fit_X)
-        if sklearn_check_version("0.22"):
-            result = sklearn_NearestNeighbors.radius_neighbors(
-                self, X, radius, return_distance, sort_results
-            )
-        else:
-            result = sklearn_NearestNeighbors.radius_neighbors(
-                self, X, radius, return_distance
-            )
+            sklearn_NearestNeighbors.fit(self, self._fit_X, getattr(self, "_y", None))
+        result = sklearn_NearestNeighbors.radius_neighbors(
+            self, X, radius, return_distance, sort_results
+        )
 
         return result
 
@@ -313,6 +222,11 @@ class KNeighborsClassifier(KNeighborsClassifier_, KNeighborsDispatchingBase):
             X, n_neighbors, return_distance, queue=queue
         )
 
+    def _onedal_score(self, X, y, sample_weight=None, queue=None):
+        return accuracy_score(
+            y, self._onedal_predict(X, queue=queue), sample_weight=sample_weight
+        )
+
     def _save_attributes(self):
         self.classes_ = self._onedal_estimator.classes_
         self.n_features_in_ = self._onedal_estimator.n_features_in_
@@ -326,5 +240,6 @@ class KNeighborsClassifier(KNeighborsClassifier_, KNeighborsDispatchingBase):
     fit.__doc__ = sklearn_KNeighborsClassifier.fit.__doc__
     predict.__doc__ = sklearn_KNeighborsClassifier.predict.__doc__
     predict_proba.__doc__ = sklearn_KNeighborsClassifier.predict_proba.__doc__
+    score.__doc__ = sklearn_KNeighborsClassifier.score.__doc__
     kneighbors.__doc__ = sklearn_KNeighborsClassifier.kneighbors.__doc__
     radius_neighbors.__doc__ = sklearn_NearestNeighbors.radius_neighbors.__doc__