From 648e6a1ffe45ceae2995c3f9ec6a9413aad55640 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Sun, 18 Oct 2015 00:11:34 +0600 Subject: [PATCH] [youtube] Generalize playlist entries extraction (Closes #6699, closes #6992) --- youtube_dl/extractor/youtube.py | 121 ++++++++++++++------------------ 1 file changed, 52 insertions(+), 69 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index b252e36e11..08e821362e 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -178,6 +178,52 @@ def _real_initialize(self): return +class YoutubePlaylistBaseInfoExtractor(InfoExtractor): + # Extract the video ids from the playlist pages + def _entries(self, page, playlist_id): + more_widget_html = content_html = page + for page_num in itertools.count(1): + 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_title=video_title) + + 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'] + + def extract_videos_from_page(self, page): + ids_in_page = [] + titles_in_page = [] + for mobj in re.finditer(self._VIDEO_RE, page): + # The link with index 0 is not the first video of the playlist (not sure if still actual) + if 'index' in mobj.groupdict() and mobj.group('id') == '0': + continue + video_id = mobj.group('id') + video_title = unescapeHTML(mobj.group('title')) + if video_title: + video_title = video_title.strip() + 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) + + class YoutubeIE(YoutubeBaseInfoExtractor): IE_DESC = 'YouTube.com' _VALID_URL = r"""(?x)^ @@ -1419,7 +1465,7 @@ def _map_to_format_list(urlmap): } -class YoutubePlaylistIE(YoutubeBaseInfoExtractor): +class YoutubePlaylistIE(YoutubeBaseInfoExtractor, YoutubePlaylistBaseInfoExtractor): IE_DESC = 'YouTube.com playlists' _VALID_URL = r"""(?x)(?: (?:https?://)? @@ -1440,7 +1486,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor): ((?:PL|LL|EC|UU|FL|RD|UL)[0-9A-Za-z-_]{10,}) )""" _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s' - _VIDEO_RE = r'href="\s*/watch\?v=(?P[0-9A-Za-z_-]{11})&[^"]*?index=(?P\d+)' + _VIDEO_RE = r'href="\s*/watch\?v=(?P[0-9A-Za-z_-]{11})&[^"]*?index=(?P\d+)(?:[^>]+>(?P[^<]+))?' IE_NAME = 'youtube:playlist' _TESTS = [{ 'url': 'https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re', @@ -1557,37 +1603,11 @@ def _extract_playlist(self, playlist_id): else: self.report_warning('Youtube gives an alert message: ' + match) - # Extract the video ids from the playlist pages - def _entries(): - more_widget_html = content_html = page - 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') - for vid_id in new_ids: - yield self.url_result(vid_id, 'Youtube', video_id=vid_id) - - 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'), 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)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>', page, 'title') - return self.playlist_result(_entries(), playlist_id, playlist_title) + return self.playlist_result(self._entries(page, playlist_id), playlist_id, playlist_title) def _real_extract(self, url): # Extract playlist id @@ -1613,10 +1633,11 @@ def _real_extract(self, url): return self._extract_playlist(playlist_id) -class YoutubeChannelIE(InfoExtractor): +class YoutubeChannelIE(YoutubePlaylistBaseInfoExtractor): IE_DESC = 'YouTube.com channels' _VALID_URL = r'https?://(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/(?P<id>[0-9A-Za-z_-]+)' _TEMPLATE_URL = 'https://www.youtube.com/channel/%s/videos' + _VIDEO_RE = r'(?:title="(?P<title>[^"]+)"[^>]+)?href="/watch\?v=(?P<id>[0-9A-Za-z_-]+)&?' IE_NAME = 'youtube:channel' _TESTS = [{ 'note': 'paginated channel', @@ -1627,22 +1648,6 @@ class YoutubeChannelIE(InfoExtractor): } }] - @staticmethod - def extract_videos_from_page(page): - ids_in_page = [] - titles_in_page = [] - for mobj in re.finditer(r'(?:title="(?P<title>[^"]+)"[^>]+)?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) @@ -1685,29 +1690,7 @@ def _real_extract(self, url): 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): - - 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_title=video_title) - - 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'), channel_id, - 'Downloading page #%s' % (pagenum + 1), - transform_source=uppercase_escape) - content_html = more['content_html'] - more_widget_html = more['load_more_widget_html'] - - return self.playlist_result(_entries(), channel_id) + return self.playlist_result(self._entries(channel_page, channel_id), channel_id) class YoutubeUserIE(YoutubeChannelIE):