Conversion: Convert images encoded as data URIs in the input document into normal images, as some ebook reader software cannot handle data URIs. Fixes #1405803 [Can't see images in *.mobi in Kindle after convertation](https://bugs.launchpad.net/calibre/+bug/1405803)

src/calibre/ebooks/conversion/plumber.py

@@ -1067,6 +1067,8 @@ OptionRecommendation(name='search_replace',
 
         self.oeb.plumber_output_format = self.output_fmt or ''
 
+        from calibre.ebooks.oeb.transforms.data_url import DataURL
+        DataURL()(self.oeb, self.opts)
         from calibre.ebooks.oeb.transforms.guide import Clean
         Clean()(self.oeb, self.opts)
         pr(0.1)

src/calibre/ebooks/oeb/transforms/data_url.py

@@ -0,0 +1,50 @@
+#!/usr/bin/env python
+# vim:fileencoding=utf-8
+from __future__ import (unicode_literals, division, absolute_import,
+                        print_function)
+
+__license__ = 'GPL v3'
+__copyright__ = '2014, Kovid Goyal <kovid at kovidgoyal.net>'
+
+import re
+from calibre.ebooks.oeb.base import XPath, urlunquote
+
+class DataURL(object):
+
+    def __call__(self, oeb, opts):
+        self.log = oeb.log
+        attr_path = XPath('//h:img[@src]')
+        for item in oeb.spine:
+            root = item.data
+            if not hasattr(root, 'xpath'):
+                continue
+            for img in attr_path(root):
+                raw = img.get('src')
+                if not raw.startswith('data:'):
+                    continue
+                header, data = raw.partition(',')[0::2]
+                if not header.startswith('data:image/') or not data:
+                    continue
+                if ';base64' in header:
+                    data = re.sub(r'\s+', '', data)
+                    from base64 import b64decode
+                    try:
+                        data = b64decode(data)
+                    except Exception:
+                        self.log.error('Found invalid base64 encoded data URI, ignoring it')
+                        continue
+                else:
+                    data = urlunquote(data)
+                from imghdr import what
+                fmt = what(None, data)
+                if not fmt:
+                    self.log.warn('Image encoded as data URL has unknown format, ignoring')
+                    continue
+                img.set('src', item.relhref(self.convert_image_data_uri(data, fmt, oeb)))
+
+    def convert_image_data_uri(self, data, fmt, oeb):
+        self.log('Found image encoded as data URI converting it to normal image')
+        from calibre import guess_type
+        item_id, item_href = oeb.manifest.generate('data-url-image', 'data-url-image.' + fmt)
+        oeb.manifest.add(item_id, item_href, guess_type(item_href)[0], data=data)
+        return item_href