fix ranger class mapping on predictions (#45)

* fix ranger class mapping on predictions

* rename class order var

* upgrade poetry and fix test and docs

* rm post

* revert updates

* revert poetry update

* make docs build

docs/requirements.txt

@@ -66,7 +66,7 @@ qtconsole==4.7.5
 qtpy==1.9.0
 regex==2020.6.8
 requests==2.24.0
-scikit-learn==0.23.1
+scikit-learn==0.22.0
 scikit-survival==0.12.0
 scipy==1.5.1
 send2trash==1.5.0

skranger/ensemble/ranger_forest_classifier.py

@@ -80,6 +80,7 @@ class specific values.
         regularization factor input parameter.
     :ivar int importance_mode\_: The importance mode integer corresponding to ranger
         enum ``ImportanceMode``.
+    :ivar list ranger_class_order\_: The class reference ordering derived from ranger.
     """
 
     def __init__(
@@ -156,7 +157,7 @@ def fit(self, X, y, sample_weight=None):
         self._validate_parameters(X, y, sample_weight)
 
         # Map classes to indices
-        y = y.copy()
+        y = np.copy(y)
         self.classes_, y = np.unique(y, return_inverse=True)
         self.n_classes_ = len(self.classes_)
 
@@ -217,6 +218,7 @@ def fit(self, X, y, sample_weight=None):
             False,  # use_regularization_factor
             self.regularization_usedepth,
         )
+        self.ranger_class_order_ = np.argsort(np.array(self.ranger_forest_["forest"]["class_values"]).astype(int))
         return self
 
     def predict(self, X):
@@ -282,7 +284,8 @@ def predict_proba(self, X):
             self.use_regularization_factor_,
             self.regularization_usedepth,
         )
-        return np.atleast_2d(np.array(result["predictions"]))
+        predictions = np.atleast_2d(np.array(result["predictions"]))
+        return predictions[:, self.ranger_class_order_]
 
     def predict_log_proba(self, X):
         """Predict log probabilities for classes from X.

tests/ensemble/test_ranger_forest_classifier.py

@@ -5,7 +5,10 @@
 import numpy as np
 import pytest
 from sklearn.base import clone
+from sklearn.ensemble import RandomForestClassifier
 from sklearn.exceptions import NotFittedError
+from sklearn.metrics import accuracy_score
+from sklearn.model_selection import train_test_split
 from sklearn.utils.validation import check_is_fitted
 
 from skranger.ensemble import RangerForestClassifier
@@ -24,6 +27,7 @@ def test_fit(self, iris_X, iris_y):
         assert hasattr(rfc, "classes_")
         assert hasattr(rfc, "n_classes_")
         assert hasattr(rfc, "ranger_forest_")
+        assert hasattr(rfc, "ranger_class_order_")
         assert hasattr(rfc, "n_features_")
 
     def test_predict(self, iris_X, iris_y):
@@ -232,3 +236,22 @@ def test_always_split_features(self, iris_X, iris_y):
         # feature 0 is in every tree split
         for tree in rfc.ranger_forest_["forest"]["split_var_ids"]:
             assert 0 in tree
+
+    def test_accuracy(self, iris_X, iris_y):
+        X_train, X_test, y_train, y_test = train_test_split(iris_X, iris_y, test_size=0.33, random_state=42)
+
+        # train and test a random forest classifier
+        rf = RandomForestClassifier()
+        rf.fit(X_train, y_train)
+        y_pred_rf = rf.predict(X_test)
+        rf_acc = accuracy_score(y_test, y_pred_rf)
+
+        # train and test a ranger classifier
+        ra = RangerForestClassifier()
+        ra.fit(X_train, y_train)
+        y_pred_ra = ra.predict(X_test)
+        ranger_acc = accuracy_score(y_test, y_pred_ra)
+
+        # the accuracy should be good
+        assert rf_acc > 0.9
+        assert ranger_acc > 0.9