Undocumented HtmlFeatureExtractor post-processing step is removed to simplify code and make it faster.

If needed, it can be implemented as a global feature.

Author: kmike

webstruct/feature_extraction.py

@@ -415,8 +415,7 @@ def transform_single(self, html_tokens):
         for feat in self.global_features:
             feat(token_data)
 
-        return [{k: fd[k] for k in fd if not k.startswith('_')}
-                for tok, fd in token_data]
+        return [featdict for tok, featdict in token_data]
 
     def _pruned(self, X, low=None):
         if low is None or low <= 1: