dont take pdf metadata if title starts with Microsoft Word

oml/media/pdf.py

@@ -220,6 +220,10 @@ def info(pdf):
         data['author'] = [ox.normalize_name(data['author'])]
     if 'description' in data:
         data['description'] = ox.strip_tags(ox.decode_html(data['description'])).strip()
+    if data.get('title', '').startswith('Microsoft Word'):
+        for key in ('title', 'author', 'producer', 'creator'):
+            if key in data:
+                del data[key]
     return data
 
 '''