Strip tags from post titles

WordPress allows to include markup on titles and we want to get rid of them.

transmogrify/wordpress/blueprints/csvsource.py

@@ -3,6 +3,7 @@
 from collective.transmogrifier.interfaces import ISectionBlueprint
 from DateTime import DateTime
 from OFS.ObjectManager import bad_id
+from transmogrify.wordpress.utils import strip_tags
 from transmogrify.wordpress.blueprints.csvutils import get_display_names
 from transmogrify.wordpress.blueprints.csvutils import get_taxonomies
 from transmogrify.wordpress.logger import logger
@@ -140,7 +141,7 @@ def __iter__(self):
                     item_id = row['post_name']
                     # Zope ids need to be ASCII
                     item_id = unquote_plus(item_id).decode('utf-8').encode('ascii', 'ignore')
-                    item['title'] = row['post_title']
+                    item['title'] = strip_tags(row['post_title'])
                 else:
                     # for attachments we need to parse the guid
                     # and use the file name as title

transmogrify/wordpress/utils.py

@@ -0,0 +1,27 @@
+# -*- coding: utf-8 -*-
+from HTMLParser import HTMLParser
+
+
+class MLStripper(HTMLParser):
+
+    """Taken from: http://stackoverflow.com/a/925630/644075
+
+    Why don't you use RegEx? http://stackoverflow.com/a/1732454/644075
+    """
+
+    def __init__(self):
+        self.reset()
+        self.fed = []
+
+    def handle_data(self, d):
+        self.fed.append(d)
+
+    def get_data(self):
+        return ''.join(self.fed)
+
+
+def strip_tags(html):
+    """Remove HTML tags from a string."""
+    s = MLStripper()
+    s.feed(html)
+    return s.get_data()