[REVIEW] Reduce duplicated cython code in RF

CHANGELOG.md

@@ -35,6 +35,7 @@
 - PR #2440: Use Treelite Conda package
 - PR #2403: Support for input and output type consistency in logistic regression predict_proba
 - PR #2468: Add `_n_features_in_` attribute to all single GPU estimators that implement fit
+- PR #2237: Refactor RF cython code
 
 ## Bug Fixes
 - PR #2369: Update RF code to fix set_params memory leak

python/cuml/benchmark/algorithms.py

@@ -396,7 +396,7 @@ def all_algorithms():
                 fil_algo="AUTO",
                 output_class=False,
                 threshold=0.5,
-                storage_type="AUTO",
+                storage_type="auto",
             ),
             name="FIL",
             accepts_labels=False,

python/cuml/dask/ensemble/base.py

@@ -124,10 +124,11 @@ def _concat_treelite_models(self):
         model._concatenate_treelite_handle(all_tl_mod_handles)
         for tl_handle in all_tl_mod_handles:
             TreeliteModel.free_treelite_model(tl_handle)
-
         return model
 
     def _predict_using_fil(self, X, delayed, **kwargs):
+        if self.local_model is None:
+            self.local_model = self._concat_treelite_models()
         data = DistributedDataHandler.create(X, client=self.client)
         self.datatype = data.datatype
         return self._predict(X, delayed=delayed, **kwargs)

python/cuml/dask/ensemble/randomforestclassifier.py

@@ -295,24 +295,15 @@ def predict(self, X, output_class=True, algo='auto', threshold=0.5,
 
         else:
             preds = \
-                self.predict_using_fil(X, output_class=output_class,
-                                       algo=algo,
-                                       threshold=threshold,
-                                       num_classes=self.num_classes,
-                                       convert_dtype=convert_dtype,
-                                       predict_model="GPU",
-                                       fil_sparse_format=fil_sparse_format,
-                                       delayed=delayed)
+                self._predict_using_fil(X, output_class=output_class,
+                                        algo=algo,
+                                        threshold=threshold,
+                                        convert_dtype=convert_dtype,
+                                        fil_sparse_format=fil_sparse_format,
+                                        delayed=delayed)
 
         return preds
 
-    def predict_using_fil(self, X, delayed, **kwargs):
-        if self.local_model is None:
-            self.local_model = self._concat_treelite_models()
-
-        return self._predict_using_fil(X=X,
-                                       delayed=delayed,
-                                       **kwargs)
     """
     TODO : Update function names used for CPU predict.
         Cuml issue #1854 has been created to track this.

python/cuml/dask/ensemble/randomforestregressor.py

@@ -270,20 +270,13 @@ def predict(self, X, predict_model="GPU", algo='auto',
 
         else:
             preds = \
-                self.predict_using_fil(X, predict_model=predict_model,
-                                       algo=algo,
-                                       convert_dtype=convert_dtype,
-                                       fil_sparse_format=fil_sparse_format,
-                                       delayed=delayed)
+                self._predict_using_fil(X,
+                                        algo=algo,
+                                        convert_dtype=convert_dtype,
+                                        fil_sparse_format=fil_sparse_format,
+                                        delayed=delayed)
         return preds
 
-    def predict_using_fil(self, X, delayed, **kwargs):
-        if self.local_model is None:
-            self.local_model = self._concat_treelite_models()
-        return self._predict_using_fil(X=X,
-                                       delayed=delayed,
-                                       **kwargs)
-
     """
     TODO : Update function names used for CPU predict.
            Cuml issue #1854 has been created to track this.

python/cuml/ensemble/__init__.py

@@ -18,4 +18,4 @@
 from cuml.ensemble.randomforestclassifier import RandomForestClassifier
 from cuml.ensemble.randomforestregressor import RandomForestRegressor
 from cuml.ensemble.randomforest_common import _check_fil_parameter_validity, \
-    _check_fil_sparse_format_value, _obtain_treelite_model, _obtain_fil_model
+    _obtain_fil_model