patch feedparser which incorrectly detects base64

feedhq/monkey.py

@@ -1,3 +1,4 @@
+import feedparser
 from html5lib.filters import alphabeticalattributes
 
 
@@ -21,3 +22,21 @@ def __iter__(self):
 
 def patch_html5lib():
     alphabeticalattributes.Filter = Filter
+
+
+def _isBase64(self, attrsD, contentparams):
+    if attrsD.get('mode', '') == 'base64':
+        return 1
+    if self.contentparams['type'].startswith('text/'):
+        return 0
+    if self.contentparams['type'].endswith('+xml'):
+        return 0
+    if self.contentparams['type'].endswith('/xml'):
+        return 0
+    if self.contentparams['type'] == 'markdown':
+        return 0
+    return 1
+
+
+def patch_feedparser():
+    feedparser._FeedParserMixin._isBase64 = _isBase64

feedhq/urls.py

@@ -10,6 +10,7 @@
 
 from . import monkey
 monkey.patch_html5lib()
+monkey.patch_feedparser()
 
 from . import views  # noqa
 from .logging import configure_logging  # noqa