[enhancement] check that all sklearnex estimators are centrally tested (

#2037)

* add test_all_estimators_covered

* forgotten import

* forgotten underscore

* remove recursion

* root -> path

* forgot to change kwargs

* first attempt

* wrong if statement

* add underscores

* sklearn_ -> _sklearn_

* fix mistake

* reformulate to wrapper

* swaps

* forgotten :

* isort fixes

* remove preview from search

* its a set

* formatting

* manually remove preview from lsit

* add BasicStatistics

* fix mistake

* Update basic_statistics.py

* Delete sklearnex/linear_model/logistic_path.py

* Update test_common.py

* Update test_common.py

* formatting

* add docstrings

* Update sklearnex/tests/test_common.py

Co-authored-by: Samir Nasibli <samir.nasibli@intel.com>

* Update test_common.py

* collect all uncovered estimators for assert

* isort fixes

* add underscores

* fix validate_data checks for IncBS

* Update k_means.py

* Update test_run_to_run_stability.py

* Update basic_statistics.py

* Update test_run_to_run_stability.py

---------

Co-authored-by: Samir Nasibli <samir.nasibli@intel.com>

sklearnex/basic_statistics/basic_statistics.py

@@ -42,6 +42,9 @@ class BasicStatistics(BaseEstimator):
     """
     Estimator for basic statistics.
     Allows to compute basic statistics for provided data.
+    Note, some results can exhibit small variations due to
+    floating point error accumulation and multithreading.
+
     Parameters
     ----------
     result_options: string or list, default='all'

sklearnex/basic_statistics/incremental_basic_statistics.py

@@ -146,21 +146,22 @@ def _onedal_finalize_fit(self, queue=None):
         self._onedal_estimator.finalize_fit(queue=queue)
         self._need_to_finalize = False
 
-    def _onedal_partial_fit(self, X, sample_weight=None, queue=None):
+    def _onedal_partial_fit(self, X, sample_weight=None, queue=None, check_input=True):
         first_pass = not hasattr(self, "n_samples_seen_") or self.n_samples_seen_ == 0
 
-        if sklearn_check_version("1.0"):
-            X = validate_data(
-                self,
-                X,
-                dtype=[np.float64, np.float32],
-                reset=first_pass,
-            )
-        else:
-            X = check_array(
-                X,
-                dtype=[np.float64, np.float32],
-            )
+        if check_input:
+            if sklearn_check_version("1.0"):
+                X = validate_data(
+                    self,
+                    X,
+                    dtype=[np.float64, np.float32],
+                    reset=first_pass,
+                )
+            else:
+                X = check_array(
+                    X,
+                    dtype=[np.float64, np.float32],
+                )
 
         if sample_weight is not None:
             sample_weight = _check_sample_weight(sample_weight, X)
@@ -206,7 +207,9 @@ def _onedal_fit(self, X, sample_weight=None, queue=None):
         for batch in gen_batches(X.shape[0], self.batch_size_):
             X_batch = X[batch]
             weights_batch = sample_weight[batch] if sample_weight is not None else None
-            self._onedal_partial_fit(X_batch, weights_batch, queue=queue)
+            self._onedal_partial_fit(
+                X_batch, weights_batch, queue=queue, check_input=False
+            )
 
         self.n_features_in_ = X.shape[1]
 
@@ -235,7 +238,7 @@ def __getattr__(self, attr):
             f"'{self.__class__.__name__}' object has no attribute '{attr}'"
         )
 
-    def partial_fit(self, X, sample_weight=None):
+    def partial_fit(self, X, sample_weight=None, check_input=True):
         """Incremental fit with X. All of X is processed as a single batch.
 
         Parameters
@@ -250,6 +253,9 @@ def partial_fit(self, X, sample_weight=None):
         sample_weight : array-like of shape (n_samples,), default=None
             Weights for compute weighted statistics, where `n_samples` is the number of samples.
 
+        check_input : bool, default=True
+            Run check_array on X.
+
         Returns
         -------
         self : object
@@ -264,6 +270,7 @@ def partial_fit(self, X, sample_weight=None):
             },
             X,
             sample_weight,
+            check_input=check_input,
         )
         return self
 

sklearnex/cluster/dbscan.py

@@ -18,7 +18,7 @@
 from abc import ABC
 
 from scipy import sparse as sp
-from sklearn.cluster import DBSCAN as sklearn_DBSCAN
+from sklearn.cluster import DBSCAN as _sklearn_DBSCAN
 from sklearn.utils.validation import _check_sample_weight
 
 from daal4py.sklearn._n_jobs_support import control_n_jobs
@@ -34,7 +34,7 @@
 if sklearn_check_version("1.6"):
     from sklearn.utils.validation import validate_data
 else:
-    validate_data = sklearn_DBSCAN._validate_data
+    validate_data = _sklearn_DBSCAN._validate_data
 
 
 class BaseDBSCAN(ABC):
@@ -51,11 +51,11 @@ def _save_attributes(self):
 
 
 @control_n_jobs(decorated_methods=["fit"])
-class DBSCAN(sklearn_DBSCAN, BaseDBSCAN):
-    __doc__ = sklearn_DBSCAN.__doc__
+class DBSCAN(_sklearn_DBSCAN, BaseDBSCAN):
+    __doc__ = _sklearn_DBSCAN.__doc__
 
     if sklearn_check_version("1.2"):
-        _parameter_constraints: dict = {**sklearn_DBSCAN._parameter_constraints}
+        _parameter_constraints: dict = {**_sklearn_DBSCAN._parameter_constraints}
 
     def __init__(
         self,
@@ -185,7 +185,7 @@ def fit(self, X, y=None, sample_weight=None):
             "fit",
             {
                 "onedal": self.__class__._onedal_fit,
-                "sklearn": sklearn_DBSCAN.fit,
+                "sklearn": _sklearn_DBSCAN.fit,
             },
             X,
             y,
@@ -194,4 +194,4 @@ def fit(self, X, y=None, sample_weight=None):
 
         return self
 
-    fit.__doc__ = sklearn_DBSCAN.fit.__doc__
+    fit.__doc__ = _sklearn_DBSCAN.fit.__doc__

sklearnex/cluster/k_means.py

@@ -25,7 +25,7 @@
 
     import numpy as np
     from scipy.sparse import issparse
-    from sklearn.cluster import KMeans as sklearn_KMeans
+    from sklearn.cluster import KMeans as _sklearn_KMeans
     from sklearn.utils._openmp_helpers import _openmp_effective_n_threads
     from sklearn.utils.validation import (
         _check_sample_weight,
@@ -44,14 +44,14 @@
     if sklearn_check_version("1.6"):
         from sklearn.utils.validation import validate_data
     else:
-        validate_data = sklearn_KMeans._validate_data
+        validate_data = _sklearn_KMeans._validate_data
 
     @control_n_jobs(decorated_methods=["fit", "fit_transform", "predict", "score"])
-    class KMeans(sklearn_KMeans):
-        __doc__ = sklearn_KMeans.__doc__
+    class KMeans(_sklearn_KMeans):
+        __doc__ = _sklearn_KMeans.__doc__
 
         if sklearn_check_version("1.2"):
-            _parameter_constraints: dict = {**sklearn_KMeans._parameter_constraints}
+            _parameter_constraints: dict = {**_sklearn_KMeans._parameter_constraints}
 
         def __init__(
             self,
@@ -145,7 +145,7 @@ def fit(self, X, y=None, sample_weight=None):
                 "fit",
                 {
                     "onedal": self.__class__._onedal_fit,
-                    "sklearn": sklearn_KMeans.fit,
+                    "sklearn": _sklearn_KMeans.fit,
                 },
                 X,
                 y,
@@ -254,7 +254,7 @@ def predict(self, X):
                     "predict",
                     {
                         "onedal": self.__class__._onedal_predict,
-                        "sklearn": sklearn_KMeans.predict,
+                        "sklearn": _sklearn_KMeans.predict,
                     },
                     X,
                 )
@@ -286,7 +286,7 @@ def predict(
                     "predict",
                     {
                         "onedal": self.__class__._onedal_predict,
-                        "sklearn": sklearn_KMeans.predict,
+                        "sklearn": _sklearn_KMeans.predict,
                     },
                     X,
                     sample_weight,
@@ -339,7 +339,7 @@ def score(self, X, y=None, sample_weight=None):
                 "score",
                 {
                     "onedal": self.__class__._onedal_score,
-                    "sklearn": sklearn_KMeans.score,
+                    "sklearn": _sklearn_KMeans.score,
                 },
                 X,
                 y,
@@ -384,11 +384,11 @@ def _save_attributes(self):
 
             self._n_init = self._onedal_estimator._n_init
 
-        fit.__doc__ = sklearn_KMeans.fit.__doc__
-        predict.__doc__ = sklearn_KMeans.predict.__doc__
-        transform.__doc__ = sklearn_KMeans.transform.__doc__
-        fit_transform.__doc__ = sklearn_KMeans.fit_transform.__doc__
-        score.__doc__ = sklearn_KMeans.score.__doc__
+        fit.__doc__ = _sklearn_KMeans.fit.__doc__
+        predict.__doc__ = _sklearn_KMeans.predict.__doc__
+        transform.__doc__ = _sklearn_KMeans.transform.__doc__
+        fit_transform.__doc__ = _sklearn_KMeans.fit_transform.__doc__
+        score.__doc__ = _sklearn_KMeans.score.__doc__
 
 else:
     from daal4py.sklearn.cluster import KMeans

sklearnex/covariance/incremental_covariance.py

@@ -20,7 +20,7 @@
 import numpy as np
 from scipy import linalg
 from sklearn.base import BaseEstimator, clone
-from sklearn.covariance import EmpiricalCovariance as sklearn_EmpiricalCovariance
+from sklearn.covariance import EmpiricalCovariance as _sklearn_EmpiricalCovariance
 from sklearn.covariance import log_likelihood
 from sklearn.utils import check_array, gen_batches
 from sklearn.utils.validation import _num_features
@@ -103,8 +103,8 @@ class IncrementalEmpiricalCovariance(BaseEstimator):
             "copy": ["boolean"],
         }
 
-    get_precision = sklearn_EmpiricalCovariance.get_precision
-    error_norm = wrap_output_data(sklearn_EmpiricalCovariance.error_norm)
+    get_precision = _sklearn_EmpiricalCovariance.get_precision
+    error_norm = wrap_output_data(_sklearn_EmpiricalCovariance.error_norm)
 
     def __init__(
         self, *, store_precision=False, assume_centered=False, batch_size=None, copy=True
@@ -374,6 +374,6 @@ def mahalanobis(self, X):
     _onedal_cpu_supported = _onedal_supported
     _onedal_gpu_supported = _onedal_supported
 
-    mahalanobis.__doc__ = sklearn_EmpiricalCovariance.mahalanobis.__doc__
-    error_norm.__doc__ = sklearn_EmpiricalCovariance.error_norm.__doc__
-    score.__doc__ = sklearn_EmpiricalCovariance.score.__doc__
+    mahalanobis.__doc__ = _sklearn_EmpiricalCovariance.mahalanobis.__doc__
+    error_norm.__doc__ = _sklearn_EmpiricalCovariance.error_norm.__doc__
+    score.__doc__ = _sklearn_EmpiricalCovariance.score.__doc__

sklearnex/decomposition/pca.py

@@ -40,21 +40,21 @@
     if sklearn_check_version("1.2"):
         from sklearn.utils._param_validation import StrOptions
 
-    from sklearn.decomposition import PCA as sklearn_PCA
+    from sklearn.decomposition import PCA as _sklearn_PCA
 
     from onedal.decomposition import PCA as onedal_PCA
 
     if sklearn_check_version("1.6"):
         from sklearn.utils.validation import validate_data
     else:
-        validate_data = sklearn_PCA._validate_data
+        validate_data = _sklearn_PCA._validate_data
 
     @control_n_jobs(decorated_methods=["fit", "transform", "fit_transform"])
-    class PCA(sklearn_PCA):
-        __doc__ = sklearn_PCA.__doc__
+    class PCA(_sklearn_PCA):
+        __doc__ = _sklearn_PCA.__doc__
 
         if sklearn_check_version("1.2"):
-            _parameter_constraints: dict = {**sklearn_PCA._parameter_constraints}
+            _parameter_constraints: dict = {**_sklearn_PCA._parameter_constraints}
             # "onedal_svd" solver uses oneDAL's PCA-SVD algorithm
             # and required for testing purposes to fully enable it in future.
             # "covariance_eigh" solver is added for ability to explicitly request
@@ -132,7 +132,7 @@ def _fit(self, X):
                 "fit",
                 {
                     "onedal": self.__class__._onedal_fit,
-                    "sklearn": sklearn_PCA._fit,
+                    "sklearn": _sklearn_PCA._fit,
                 },
                 X,
             )
@@ -175,7 +175,7 @@ def transform(self, X):
                 "transform",
                 {
                     "onedal": self.__class__._onedal_transform,
-                    "sklearn": sklearn_PCA.transform,
+                    "sklearn": _sklearn_PCA.transform,
                 },
                 X,
             )
@@ -412,10 +412,10 @@ def _validate_n_features_in_after_fitting(self, X):
                     )
                 )
 
-        fit.__doc__ = sklearn_PCA.fit.__doc__
-        transform.__doc__ = sklearn_PCA.transform.__doc__
-        fit_transform.__doc__ = sklearn_PCA.fit_transform.__doc__
-        inverse_transform.__doc__ = sklearn_PCA.inverse_transform.__doc__
+        fit.__doc__ = _sklearn_PCA.fit.__doc__
+        transform.__doc__ = _sklearn_PCA.transform.__doc__
+        fit_transform.__doc__ = _sklearn_PCA.fit_transform.__doc__
+        inverse_transform.__doc__ = _sklearn_PCA.inverse_transform.__doc__
 
 else:
     from daal4py.sklearn.decomposition import PCA