rebuid and recheck

coverage.txt

@@ -32,10 +32,10 @@ pkg/vtreat/cross_plan.py       50     11    78%
 pkg/vtreat/stats_utils.py      82      3    96%
 pkg/vtreat/transform.py        17      4    76%
 pkg/vtreat/util.py            149     21    86%
-pkg/vtreat/vtreat_api.py      285     57    80%
-pkg/vtreat/vtreat_impl.py     650     86    87%
+pkg/vtreat/vtreat_api.py      285     36    87%
+pkg/vtreat/vtreat_impl.py     655     80    88%
 -----------------------------------------------
-TOTAL                        1239    182    85%
+TOTAL                        1244    155    88%
 
 
-============================= 28 passed in 19.27s ==============================
+============================= 28 passed in 19.38s ==============================

pkg/build/lib/vtreat/vtreat_api.py

@@ -127,7 +127,10 @@ def transform(self, X):
                 "possibly called transform on same data used to fit\n" +
                 "(this causes over-fit, please use fit_transform() instead)")
         res = vtreat_impl.pre_prep_frame(
-            X, col_list=self.var_list_, cols_to_copy=self.cols_to_copy_, cat_cols=self.plan_['cat_list']
+            X,
+            col_list=self.plan_['num_list'] + self.plan_['cat_list'],
+            cols_to_copy=self.cols_to_copy_,
+            cat_cols=self.plan_['cat_list']
         )
         res = vtreat_impl.perform_transform(x=res, transform=self, params=self.params_)
         res = vtreat_impl.limit_to_appropriate_columns(res=res, transform=self)
@@ -274,7 +277,10 @@ def transform(self, X):
                 "possibly called transform on same data used to fit\n" +
                 "(this causes over-fit, please use fit_transform() instead)")
         X = vtreat_impl.pre_prep_frame(
-            X, col_list=self.var_list_, cols_to_copy=self.cols_to_copy_, cat_cols=self.plan_['cat_list']
+            X,
+            col_list=self.plan_['num_list'] + self.plan_['cat_list'],
+            cols_to_copy=self.cols_to_copy_,
+            cat_cols=self.plan_['cat_list']
         )
         res = vtreat_impl.perform_transform(x=X, transform=self, params=self.params_)
         res = vtreat_impl.limit_to_appropriate_columns(res=res, transform=self)
@@ -421,7 +427,10 @@ def transform(self, X):
                 "possibly called transform on same data used to fit\n" +
                 "(this causes over-fit, please use fit_transform() instead)")
         X = vtreat_impl.pre_prep_frame(
-            X, col_list=self.var_list_, cols_to_copy=self.cols_to_copy_, cat_cols=self.plan_['cat_list']
+            X,
+            col_list=self.plan_['num_list'] + self.plan_['cat_list'],
+            cols_to_copy=self.cols_to_copy_,
+            cat_cols=self.plan_['cat_list']
         )
         res = vtreat_impl.perform_transform(x=X, transform=self, params=self.params_)
         res = vtreat_impl.limit_to_appropriate_columns(res=res, transform=self)
@@ -558,7 +567,10 @@ def transform(self, X):
         if self.last_fit_x_id_ is None:
             raise ValueError("called transform on not yet fit treatment")
         X = vtreat_impl.pre_prep_frame(
-            X, col_list=self.var_list_, cols_to_copy=self.cols_to_copy_, cat_cols=self.plan_['cat_list']
+            X,
+            col_list=self.plan_['num_list'] + self.plan_['cat_list'],
+            cols_to_copy=self.cols_to_copy_,
+            cat_cols=self.plan_['cat_list']
         )
         res = vtreat_impl.perform_transform(x=X, transform=self, params=self.params_)
         res = vtreat_impl.limit_to_appropriate_columns(res=res, transform=self)

pkg/build/lib/vtreat/vtreat_impl.py

@@ -1091,7 +1091,12 @@ def predict_proba(self, X):
     def get_feature_names(self, input_features=None):
         if self.score_frame_ is None:
             raise ValueError("get_feature_names called on uninitialized vtreat transform")
-        if self.params_['filter_to_recommended']:
+        filter_to_recommended = False
+        try:
+            filter_to_recommended = self.params_['filter_to_recommended']
+        except KeyError:
+            pass
+        if filter_to_recommended:
             new_vars = [self.score_frame_['variable'][i] for i in range(self.score_frame_.shape[0])
                         if self.score_frame_['has_range'][i] and self.score_frame_['recommended'][i]
                         and (input_features is None or self.score_frame_['orig_variable'][i] in input_features)]