News download: Add support for images embedded in the HTML

src/calibre/web/fetch/simple.py

@@ -12,6 +12,7 @@ from urllib import url2pathname, quote
 from httplib import responses
 from PIL import Image
 from cStringIO import StringIO
+from base64 import b64decode
 
 from calibre import browser, relpath, unicode_path
 from calibre.constants import filesystem_encoding, iswindows
@@ -346,6 +347,13 @@ class RecursiveFetcher(object):
         c = 0
         for tag in soup.findAll(lambda tag: tag.name.lower()=='img' and tag.has_key('src')):
             iurl = tag['src']
+            if iurl.startswith('data:image/'):
+                try:
+                    data = b64decode(iurl.partition(',')[-1])
+                except:
+                    self.log.exception('Failed to decode embedded image')
+                    continue
+            else:
                 if callable(self.image_url_processor):
                     iurl = self.image_url_processor(baseurl, iurl)
                 if not urlparse.urlsplit(iurl).scheme: