FIX raise error properly when n_features differ in fit and apply

Author: arjoly

sklearn/ensemble/forest.py

@@ -160,7 +160,7 @@ def apply(self, X):
         X = check_array(X, dtype=DTYPE, accept_sparse="csr")
         results = Parallel(n_jobs=self.n_jobs, verbose=self.verbose,
                            backend="threading")(
-            delayed(_parallel_helper)(tree.tree_, 'apply', X)
+            delayed(_parallel_helper)(tree, 'apply', X, check_input=False)
             for tree in self.estimators_)
 
         return np.array(results).T

sklearn/tree/tests/test_tree.py

@@ -541,10 +541,12 @@ def test_error():
         est = TreeEstimator()
         est.fit(np.dot(X, Xt), y)
         assert_raises(ValueError, est.predict, X)
+        assert_raises(ValueError, est.apply, X)
 
         clf = TreeEstimator()
         clf.fit(X, y)
         assert_raises(ValueError, clf.predict, Xt)
+        assert_raises(ValueError, clf.apply, Xt)
 
         # apply before fitting
         est = TreeEstimator()

sklearn/tree/tree.py

@@ -375,7 +375,7 @@ def predict(self, X, check_input=True):
             else:
                 return proba[:, :, 0]
 
-    def apply(self, X):
+    def apply(self, X, check_input=True):
         """
         Returns the index of the leaf that each sample is predicted as.
 
@@ -386,6 +386,10 @@ def apply(self, X):
             ``dtype=np.float32`` and if a sparse matrix is provided
             to a sparse ``csr_matrix``.
 
+        check_input : boolean, (default=True)
+            Allow to bypass several input checking.
+            Don't use this parameter unless you know what you do.
+
         Returns
         -------
         X_leaves : array_like, shape = [n_samples,]
@@ -398,7 +402,15 @@ def apply(self, X):
             raise NotFittedError("Estimator not fitted, "
                                  "call `fit` before `apply`.")
 
-        X = check_array(X, dtype=DTYPE, accept_sparse="csr")
+        if check_input:
+            X = check_array(X, dtype=DTYPE, accept_sparse="csr")
+
+        n_features = X.shape[1]
+        if self.n_features_ != n_features:
+            raise ValueError("Number of features of the model must "
+                             " match the input. Model n_features is %s and "
+                             " input n_features is %s "
+                             % (self.n_features_, n_features))
 
         return self.tree_.apply(X)