Header tags added in getFormattedText

muggot · Feb 20, 2013 · 64a80c7 · 64a80c7
1 parent 430116f
commit 64a80c7
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/goose/parsers.py b/goose/parsers.py
@@ -165,7 +165,7 @@ def getText(self, node):
 
     @classmethod
     def getFormattedText(self, node):
-	pars = node.cssselect('p')
+	pars = node.cssselect('h1,h2,h3,h4,h5,p')
         for p in pars:
             if p.text is not None: p.text = u'\ufffc ' + p.text
         return Parser.getText(node)