Keep text formatting in convertToText

muggot · Feb 17, 2013 · 7420632 · 7420632
1 parent a2e7869
commit 7420632
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 1 deletion.
diff --git a/goose/outputformatters.py b/goose/outputformatters.py
@@ -64,7 +64,7 @@ def convertToText(self):
             txt = HTMLParser().unescape(txt)
             txts.append(innerTrim(txt))
         for node in list(self.getTopNode()):
-            txt = Parser.getText(node)
+            txt = Parser.getFormattedText(node)
             if txt:
                 txt = HTMLParser().unescape(txt)
                 txts.append(innerTrim(txt))

diff --git a/goose/parsers.py b/goose/parsers.py
@@ -163,6 +163,13 @@ def getText(self, node):
         txts = [i for i in node.itertext()]
         return innerTrim(u' '.join(txts).strip())
 
+    @classmethod
+    def getFormattedText(self, node):
+	pars = node.cssselect('p')
+        for p in pars:
+            if p.text is not None: p.text = u'\ufffc ' + p.text
+        return Parser.getText(node)
+
     @classmethod
     def previousSiblings(self, node):
         nodes = []