[^"]+)"', more_widget_html)
+ if not mobj:
+ break
+
+ more = self._download_json(
+ 'https://youtube.com/%s' % mobj.group('more'), playlist_id,
+ 'Downloading page #%s' % page_num,
+ transform_source=uppercase_escape)
+ content_html = more['content_html']
+ if not content_html.strip():
+ # Some webpages show a "Load more" button but they don't
+ # have more videos
+ break
+ more_widget_html = more['load_more_widget_html']
+
+ playlist_title = self._html_search_regex(
+ r'(?s)',
+ page, 'title')
+
+ return self.playlist_result(_entries(), playlist_id, playlist_title)
+
def _real_extract(self, url):
# Extract playlist id
mobj = re.match(self._VALID_URL, url)
@@ -1284,53 +1449,13 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
# Mixes require a custom extraction process
return self._extract_mix(playlist_id)
- url = self._TEMPLATE_URL % playlist_id
- page = self._download_webpage(url, playlist_id)
- more_widget_html = content_html = page
-
- # Check if the playlist exists or is private
- if re.search(r'[^<]*?(The|This) playlist (does not exist|is private)[^<]*?
', page) is not None:
- raise ExtractorError(
- 'The playlist doesn\'t exist or is private, use --username or '
- '--netrc to access it.',
- expected=True)
-
- # Extract the video ids from the playlist pages
- ids = []
-
- for page_num in itertools.count(1):
- matches = re.finditer(self._VIDEO_RE, content_html)
- # We remove the duplicates and the link with index 0
- # (it's not the first video of the playlist)
- new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
- ids.extend(new_ids)
-
- mobj = re.search(r'data-uix-load-more-href="/?(?P[^"]+)"', more_widget_html)
- if not mobj:
- break
-
- more = self._download_json(
- 'https://youtube.com/%s' % mobj.group('more'), playlist_id,
- 'Downloading page #%s' % page_num,
- transform_source=uppercase_escape)
- content_html = more['content_html']
- if not content_html.strip():
- # Some webpages show a "Load more" button but they don't
- # have more videos
- break
- more_widget_html = more['load_more_widget_html']
-
- playlist_title = self._html_search_regex(
- r'(?s)',
- page, 'title')
-
- url_results = self._ids_to_results(ids)
- return self.playlist_result(url_results, playlist_id, playlist_title)
+ return self._extract_playlist(playlist_id)
class YoutubeChannelIE(InfoExtractor):
IE_DESC = 'YouTube.com channels'
_VALID_URL = r'https?://(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/(?P[0-9A-Za-z_-]+)'
+ _TEMPLATE_URL = 'https://www.youtube.com/channel/%s/videos'
IE_NAME = 'youtube:channel'
_TESTS = [{
'note': 'paginated channel',
@@ -1341,19 +1466,45 @@ class YoutubeChannelIE(InfoExtractor):
}
}]
- def extract_videos_from_page(self, page):
+ @staticmethod
+ def extract_videos_from_page(page):
ids_in_page = []
- for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
- if mobj.group(1) not in ids_in_page:
- ids_in_page.append(mobj.group(1))
- return ids_in_page
+ titles_in_page = []
+ for mobj in re.finditer(r'(?:title="(?P[^"]+)"[^>]+)?href="/watch\?v=(?P<id>[0-9A-Za-z_-]+)&?', page):
+ video_id = mobj.group('id')
+ video_title = unescapeHTML(mobj.group('title'))
+ try:
+ idx = ids_in_page.index(video_id)
+ if video_title and not titles_in_page[idx]:
+ titles_in_page[idx] = video_title
+ except ValueError:
+ ids_in_page.append(video_id)
+ titles_in_page.append(video_title)
+ return zip(ids_in_page, titles_in_page)
def _real_extract(self, url):
channel_id = self._match_id(url)
- video_ids = []
- url = 'https://www.youtube.com/channel/%s/videos' % channel_id
- channel_page = self._download_webpage(url, channel_id)
+ url = self._TEMPLATE_URL % channel_id
+
+ # Channel by page listing is restricted to 35 pages of 30 items, i.e. 1050 videos total (see #5778)
+ # Workaround by extracting as a playlist if managed to obtain channel playlist URL
+ # otherwise fallback on channel by page extraction
+ channel_page = self._download_webpage(
+ url + '?view=57', channel_id,
+ 'Downloading channel page', fatal=False)
+ channel_playlist_id = self._html_search_meta(
+ 'channelId', channel_page, 'channel id', default=None)
+ if not channel_playlist_id:
+ channel_playlist_id = self._search_regex(
+ r'data-channel-external-id="([^"]+)"',
+ channel_page, 'channel id', default=None)
+ if channel_playlist_id and channel_playlist_id.startswith('UC'):
+ playlist_id = 'UU' + channel_playlist_id[2:]
+ return self.url_result(
+ compat_urlparse.urljoin(url, '/playlist?list=%s' % playlist_id), 'YoutubePlaylist')
+
+ channel_page = self._download_webpage(url, channel_id, 'Downloading page #1')
autogenerated = re.search(r'''(?x)
class="[^"]*?(?:
channel-header-autogenerated-label|
@@ -1363,20 +1514,21 @@ class YoutubeChannelIE(InfoExtractor):
if autogenerated:
# The videos are contained in a single page
# the ajax pages can't be used, they are empty
- video_ids = self.extract_videos_from_page(channel_page)
entries = [
- self.url_result(video_id, 'Youtube', video_id=video_id)
- for video_id in video_ids]
+ self.url_result(
+ video_id, 'Youtube', video_id=video_id,
+ video_title=video_title)
+ for video_id, video_title in self.extract_videos_from_page(channel_page)]
return self.playlist_result(entries, channel_id)
def _entries():
more_widget_html = content_html = channel_page
for pagenum in itertools.count(1):
- ids_in_page = self.extract_videos_from_page(content_html)
- for video_id in ids_in_page:
+ for video_id, video_title in self.extract_videos_from_page(content_html):
yield self.url_result(
- video_id, 'Youtube', video_id=video_id)
+ video_id, 'Youtube', video_id=video_id,
+ video_title=video_title)
mobj = re.search(
r'data-uix-load-more-href="/?(?P<more>[^"]+)"',
@@ -1394,12 +1546,10 @@ class YoutubeChannelIE(InfoExtractor):
return self.playlist_result(_entries(), channel_id)
-class YoutubeUserIE(InfoExtractor):
+class YoutubeUserIE(YoutubeChannelIE):
IE_DESC = 'YouTube.com user videos (URL or "ytuser" keyword)'
_VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)(?P<id>[A-Za-z0-9_-]+)'
- _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s'
- _GDATA_PAGE_SIZE = 50
- _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
+ _TEMPLATE_URL = 'https://www.youtube.com/user/%s/videos'
IE_NAME = 'youtube:user'
_TESTS = [{
@@ -1423,95 +1573,57 @@ class YoutubeUserIE(InfoExtractor):
else:
return super(YoutubeUserIE, cls).suitable(url)
- def _real_extract(self, url):
- username = self._match_id(url)
-
- # Download video ids using YouTube Data API. Result size per
- # query is limited (currently to 50 videos) so we need to query
- # page by page until there are no video ids - it means we got
- # all of them.
-
- def download_page(pagenum):
- start_index = pagenum * self._GDATA_PAGE_SIZE + 1
-
- gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
- page = self._download_webpage(
- gdata_url, username,
- 'Downloading video ids from %d to %d' % (
- start_index, start_index + self._GDATA_PAGE_SIZE))
-
- try:
- response = json.loads(page)
- except ValueError as err:
- raise ExtractorError('Invalid JSON in API response: ' + compat_str(err))
- if 'entry' not in response['feed']:
- return
-
- # Extract video identifiers
- entries = response['feed']['entry']
- for entry in entries:
- title = entry['title']['$t']
- video_id = entry['id']['$t'].split('/')[-1]
- yield {
- '_type': 'url',
- 'url': video_id,
- 'ie_key': 'Youtube',
- 'id': video_id,
- 'title': title,
- }
- url_results = OnDemandPagedList(download_page, self._GDATA_PAGE_SIZE)
-
- return self.playlist_result(url_results, playlist_title=username)
-
-class YoutubeSearchIE(SearchInfoExtractor):
+class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE):
IE_DESC = 'YouTube.com searches'
- _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
- _MAX_RESULTS = 1000
+ # there doesn't appear to be a real limit, for example if you search for
+ # 'python' you get more than 8.000.000 results
+ _MAX_RESULTS = float('inf')
IE_NAME = 'youtube:search'
_SEARCH_KEY = 'ytsearch'
+ _EXTRA_QUERY_ARGS = {}
+ _TESTS = []
def _get_n_results(self, query, n):
"""Get a specified number of results for a query"""
- video_ids = []
- pagenum = 0
+ videos = []
limit = n
- PAGE_SIZE = 50
- while (PAGE_SIZE * pagenum) < limit:
- result_url = self._API_URL % (
- compat_urllib_parse.quote_plus(query.encode('utf-8')),
- (PAGE_SIZE * pagenum) + 1)
- data_json = self._download_webpage(
+ for pagenum in itertools.count(1):
+ url_query = {
+ 'search_query': query.encode('utf-8'),
+ 'page': pagenum,
+ 'spf': 'navigate',
+ }
+ url_query.update(self._EXTRA_QUERY_ARGS)
+ result_url = 'https://www.youtube.com/results?' + compat_urllib_parse.urlencode(url_query)
+ data = self._download_json(
result_url, video_id='query "%s"' % query,
- note='Downloading page %s' % (pagenum + 1),
+ note='Downloading page %s' % pagenum,
errnote='Unable to download API page')
- data = json.loads(data_json)
- api_response = data['data']
+ html_content = data[1]['body']['content']
- if 'items' not in api_response:
+ if 'class="search-message' in html_content:
raise ExtractorError(
'[youtube] No video results', expected=True)
- new_ids = list(video['id'] for video in api_response['items'])
- video_ids += new_ids
-
- limit = min(n, api_response['totalItems'])
- pagenum += 1
+ new_videos = self._ids_to_results(orderedSet(re.findall(
+ r'href="/watch\?v=(.{11})', html_content)))
+ videos += new_videos
+ if not new_videos or len(videos) > limit:
+ break
- if len(video_ids) > n:
- video_ids = video_ids[:n]
- videos = [self.url_result(video_id, 'Youtube', video_id=video_id)
- for video_id in video_ids]
+ if len(videos) > n:
+ videos = videos[:n]
return self.playlist_result(videos, query)
class YoutubeSearchDateIE(YoutubeSearchIE):
IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
- _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published'
_SEARCH_KEY = 'ytsearchdate'
IE_DESC = 'YouTube.com searches, newest videos first'
+ _EXTRA_QUERY_ARGS = {'search_sort': 'video_date_uploaded'}
class YoutubeSearchURLIE(InfoExtractor):
@@ -1528,11 +1640,11 @@ class YoutubeSearchURLIE(InfoExtractor):
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
- query = compat_urllib_parse.unquote_plus(mobj.group('query'))
+ query = compat_urllib_parse_unquote_plus(mobj.group('query'))
webpage = self._download_webpage(url, query)
result_code = self._search_regex(
- r'(?s)<ol class="item-section"(.*?)</ol>', webpage, 'result HTML')
+ r'(?s)<ol[^>]+class="item-section"(.*?)</ol>', webpage, 'result HTML')
part_codes = re.findall(
r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code)
@@ -1595,20 +1707,10 @@ class YoutubeShowIE(InfoExtractor):
class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
"""
- Base class for extractors that fetch info from
- http://www.youtube.com/feed_ajax
+ Base class for feed extractors
Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
"""
_LOGIN_REQUIRED = True
- # use action_load_personal_feed instead of action_load_system_feed
- _PERSONAL_FEED = False
-
- @property
- def _FEED_TEMPLATE(self):
- action = 'action_load_system_feed'
- if self._PERSONAL_FEED:
- action = 'action_load_personal_feed'
- return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
@property
def IE_NAME(self):
@@ -1618,51 +1720,49 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
self._login()
def _real_extract(self, url):
- feed_entries = []
- paging = 0
- for i in itertools.count(1):
- info = self._download_json(
- self._FEED_TEMPLATE % paging,
- '%s feed' % self._FEED_NAME,
- 'Downloading page %s' % i,
- transform_source=uppercase_escape)
- feed_html = info.get('feed_html') or info.get('content_html')
- load_more_widget_html = info.get('load_more_widget_html') or feed_html
- m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
- ids = orderedSet(m.group(1) for m in m_ids)
- feed_entries.extend(
- self.url_result(video_id, 'Youtube', video_id=video_id)
- for video_id in ids)
- mobj = re.search(
- r'data-uix-load-more-href="/?[^"]+paging=(?P<paging>\d+)',
- load_more_widget_html)
- if mobj is None:
+ page = self._download_webpage(
+ 'https://www.youtube.com/feed/%s' % self._FEED_NAME, self._PLAYLIST_TITLE)
+
+ # The extraction process is the same as for playlists, but the regex
+ # for the video ids doesn't contain an index
+ ids = []
+ more_widget_html = content_html = page
+ for page_num in itertools.count(1):
+ matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html)
+
+ # 'recommended' feed has infinite 'load more' and each new portion spins
+ # the same videos in (sometimes) slightly different order, so we'll check
+ # for unicity and break when portion has no new videos
+ new_ids = filter(lambda video_id: video_id not in ids, orderedSet(matches))
+ if not new_ids:
break
- paging = mobj.group('paging')
- return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
+ ids.extend(new_ids)
-class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
- IE_DESC = 'YouTube.com recommended videos, ":ytrec" for short (requires authentication)'
- _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
- _FEED_NAME = 'recommended'
- _PLAYLIST_TITLE = 'Youtube Recommended videos'
+ mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
+ if not mobj:
+ break
+ more = self._download_json(
+ 'https://youtube.com/%s' % mobj.group('more'), self._PLAYLIST_TITLE,
+ 'Downloading page #%s' % page_num,
+ transform_source=uppercase_escape)
+ content_html = more['content_html']
+ more_widget_html = more['load_more_widget_html']
-class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
+ return self.playlist_result(
+ self._ids_to_results(ids), playlist_title=self._PLAYLIST_TITLE)
+
+
+class YoutubeWatchLaterIE(YoutubePlaylistIE):
+ IE_NAME = 'youtube:watchlater'
IE_DESC = 'Youtube watch later list, ":ytwatchlater" for short (requires authentication)'
- _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
- _FEED_NAME = 'watch_later'
- _PLAYLIST_TITLE = 'Youtube Watch Later'
- _PERSONAL_FEED = True
+ _VALID_URL = r'https?://www\.youtube\.com/(?:feed/watch_later|playlist\?list=WL)|:ytwatchlater'
+ _TESTS = [] # override PlaylistIE tests
-class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
- IE_DESC = 'Youtube watch history, ":ythistory" for short (requires authentication)'
- _VALID_URL = 'https?://www\.youtube\.com/feed/history|:ythistory'
- _FEED_NAME = 'history'
- _PERSONAL_FEED = True
- _PLAYLIST_TITLE = 'Youtube Watch History'
+ def _real_extract(self, url):
+ return self._extract_playlist('WL')
class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
@@ -1677,42 +1777,25 @@ class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
return self.url_result(playlist_id, 'YoutubePlaylist')
-class YoutubeSubscriptionsIE(YoutubePlaylistIE):
- IE_NAME = 'youtube:subscriptions'
- IE_DESC = 'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
- _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
- _TESTS = []
-
- def _real_extract(self, url):
- title = 'Youtube Subscriptions'
- page = self._download_webpage('https://www.youtube.com/feed/subscriptions', title)
-
- # The extraction process is the same as for playlists, but the regex
- # for the video ids doesn't contain an index
- ids = []
- more_widget_html = content_html = page
+class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
+ IE_DESC = 'YouTube.com recommended videos, ":ytrec" for short (requires authentication)'
+ _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
+ _FEED_NAME = 'recommended'
+ _PLAYLIST_TITLE = 'Youtube Recommended videos'
- for page_num in itertools.count(1):
- matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html)
- new_ids = orderedSet(matches)
- ids.extend(new_ids)
- mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
- if not mobj:
- break
+class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
+ IE_DESC = 'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
+ _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
+ _FEED_NAME = 'subscriptions'
+ _PLAYLIST_TITLE = 'Youtube Subscriptions'
- more = self._download_json(
- 'https://youtube.com/%s' % mobj.group('more'), title,
- 'Downloading page #%s' % page_num,
- transform_source=uppercase_escape)
- content_html = more['content_html']
- more_widget_html = more['load_more_widget_html']
- return {
- '_type': 'playlist',
- 'title': title,
- 'entries': self._ids_to_results(ids),
- }
+class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
+ IE_DESC = 'Youtube watch history, ":ythistory" for short (requires authentication)'
+ _VALID_URL = 'https?://www\.youtube\.com/feed/history|:ythistory'
+ _FEED_NAME = 'history'
+ _PLAYLIST_TITLE = 'Youtube History'
class YoutubeTruncatedURLIE(InfoExtractor):