[REVIEW][PROPOSAL] Add tags and prefered memory order tags to estimators

python/cuml/common/base.pyx

@@ -36,6 +36,12 @@ from pandas import Series as pdSeries
 from numba import cuda
 
 
+# tag system based on experimental tag system from Scikit-learn >=0.21
+_default_tags = {
+    'preferred_input_order': None
+}
+
+
 class Base:
     """
     Base class for all the ML algos. It handles some of the common operations
@@ -357,6 +363,19 @@ class Base:
         else:
             self.n_features_in_ = X.shape[1]
 
+    def _more_tags(self):
+        return _default_tags
+
+    def _get_tags(self):
+        # method and code from scikit-learn 0.21 _get_tags functionality:
+        # https://scikit-learn.org/stable/developers/develop.html#estimator-tags
+        collected_tags = {}
+        for cl in reversed(inspect.getmro(self.__class__)):
+            if hasattr(cl, '_more_tags'):
+                more_tags = cl._more_tags(self)
+                collected_tags.update(more_tags)
+        return collected_tags
+
 
 class RegressorMixin:
     """Mixin class for regression estimators in cuML"""

python/cuml/linear_model/elastic_net.pyx

@@ -238,3 +238,8 @@ class ElasticNet(Base, RegressorMixin):
             "tol",
             "selection",
         ]
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/linear_model/lasso.pyx

@@ -206,3 +206,8 @@ class Lasso(Base, RegressorMixin):
             "tol",
             "selection",
         ]
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/linear_model/linear_regression.pyx

@@ -353,3 +353,8 @@ class LinearRegression(Base, RegressorMixin):
     def get_param_names(self):
         return super().get_param_names() + \
             ['algorithm', 'fit_intercept', 'normalize']
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/linear_model/logistic_regression.pyx

@@ -421,3 +421,8 @@ class LogisticRegression(Base, ClassifierMixin):
         super(LogisticRegression, self).__init__(handle=None,
                                                  verbose=state["verbose"])
         self.__dict__.update(state)
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/linear_model/mbsgd_classifier.pyx

@@ -216,3 +216,8 @@ class MBSGDClassifier(Base, ClassifierMixin):
             "batch_size",
             "n_iter_no_change",
         ]
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/linear_model/mbsgd_regressor.pyx

@@ -210,3 +210,8 @@ class MBSGDRegressor(Base, RegressorMixin):
             "batch_size",
             "n_iter_no_change",
         ]
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/linear_model/ridge.pyx

@@ -210,7 +210,6 @@ class Ridge(Base, RegressorMixin):
     def __init__(self, alpha=1.0, solver='eig', fit_intercept=True,
                  normalize=False, handle=None, output_type=None,
                  verbose=False):
-
         """
         Initializes the linear ridge regression class.
 
@@ -393,3 +392,8 @@ class Ridge(Base, RegressorMixin):
     def get_param_names(self):
         return super().get_param_names() + \
             ['solver', 'fit_intercept', 'normalize', 'alpha']
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/solvers/cd.pyx

@@ -353,3 +353,8 @@ class CD(Base):
             "tol",
             "shuffle",
         ]
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/solvers/qn.pyx

@@ -523,3 +523,8 @@ class QN(Base):
         return super().get_param_names() + \
             ['loss', 'fit_intercept', 'l1_strength', 'l2_strength',
                 'max_iter', 'tol', 'linesearch_max_iter', 'lbfgs_memory']
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }

python/cuml/solvers/sgd.pyx

@@ -506,3 +506,8 @@ class SGD(Base):
             "batch_size",
             "n_iter_no_change",
         ]
+
+    def _more_tags(self):
+        return {
+            'preferred_input_order': 'F'
+        }