- preloads = re.findall(r'"preload":\s*"([^"]+)"', webpage)
- if not preloads:
- raise ExtractorError('Preload information could not be extracted')
- preload = max([(len(p), p) for p in preloads])[1]
- data = self._parse_json(
- base64.b64decode(preload.encode('ascii')).decode('utf-8'), video_id)
+ data = None
+
+ preload_codes = self._html_search_regex(
+ r'(function.+)setTimeout\(function\(\)\{playlist',
+ webpage, 'preload codes')
+ base64_fragments = re.findall(r'"([a-zA-z0-9+/=]+)"', preload_codes)
+ base64_fragments.remove('init')
+
+ def _check_sequence(cur_fragments):
+ if not cur_fragments:
+ return
+ for i in range(len(cur_fragments)):
+ cur_sequence = (''.join(cur_fragments[i:] + cur_fragments[:i])).encode('ascii')
+ try:
+ raw_data = base64.b64decode(cur_sequence)
+ if compat_ord(raw_data[0]) == compat_ord('{'):
+ return json.loads(raw_data.decode('utf-8'))
+ except (TypeError, binascii.Error, UnicodeDecodeError, ValueError):
+ continue
+
+ def _check_data():
+ for i in range(len(base64_fragments) + 1):
+ for j in range(i, len(base64_fragments) + 1):
+ data = _check_sequence(base64_fragments[:i] + base64_fragments[j:])
+ if data:
+ return data
+
+ self.to_screen('Try to compute possible data sequence. This may take some time.')
+ data = _check_data()
+
+ if not data:
+ raise ExtractorError(
+ 'Preload information could not be extracted', expected=True)