get_element_by_id,
int_or_none,
orderedSet,
+ str_to_int,
unescapeHTML,
unified_strdate,
uppercase_escape,
video_description = ''
def _extract_count(count_name):
- count = self._search_regex(
- r'id="watch-%s"[^>]*>.*?([\d,]+)\s*</span>' % re.escape(count_name),
- video_webpage, count_name, default=None)
- if count is not None:
- return int(count.replace(',', ''))
- return None
+ return str_to_int(self._search_regex(
+ r'-%s-button[^>]+><span[^>]+class="yt-uix-button-content"[^>]*>([\d,]+)</span>'
+ % re.escape(count_name),
+ video_webpage, count_name, default=None))
+
like_count = _extract_count('like')
dislike_count = _extract_count('dislike')