From 351f37c022b24144c064fab39bd6d134e166c31c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Sergey=20M=E2=80=A4?= Date: Tue, 13 Aug 2019 05:02:52 +0700 Subject: [PATCH] [youtube:playlist] Improve flat extraction (closes #21927) --- youtube_dl/extractor/youtube.py | 49 +++++++++++++++++++++++++++++---- 1 file changed, 43 insertions(+), 6 deletions(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 0a0d2f41a..b63f19bb0 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -31,6 +31,7 @@ from ..utils import ( clean_html, dict_get, error_to_compat_str, + extract_attributes, ExtractorError, float_or_none, get_element_by_attribute, @@ -324,17 +325,18 @@ class YoutubePlaylistBaseInfoExtractor(YoutubeEntryListBaseInfoExtractor): for video_id, video_title in self.extract_videos_from_page(content): yield self.url_result(video_id, 'Youtube', video_id, video_title) - def extract_videos_from_page(self, page): - ids_in_page = [] - titles_in_page = [] - for mobj in re.finditer(self._VIDEO_RE, page): + def extract_videos_from_page_impl(self, video_re, page, ids_in_page, titles_in_page): + for mobj in re.finditer(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')) + video_title = unescapeHTML( + mobj.group('title')) if 'title' in mobj.groupdict() else None if video_title: video_title = video_title.strip() + if video_title == '► Play all': + video_title = None try: idx = ids_in_page.index(video_id) if video_title and not titles_in_page[idx]: @@ -342,6 +344,12 @@ class YoutubePlaylistBaseInfoExtractor(YoutubeEntryListBaseInfoExtractor): except ValueError: ids_in_page.append(video_id) titles_in_page.append(video_title) + + def extract_videos_from_page(self, page): + ids_in_page = [] + titles_in_page = [] + self.extract_videos_from_page_impl( + self._VIDEO_RE, page, ids_in_page, titles_in_page) return zip(ids_in_page, titles_in_page) @@ -2438,7 +2446,8 @@ class YoutubePlaylistIE(YoutubePlaylistBaseInfoExtractor): (%(playlist_id)s) )""" % {'playlist_id': YoutubeBaseInfoExtractor._PLAYLIST_ID_RE} _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+))?(?:[^>]+>(?P[^<]+))?)?' + _VIDEO_RE_TPL = r'href="\s*/watch\?v=%s(?:&(?:[^"]*?index=(?P<index>\d+))?(?:[^>]+>(?P<title>[^<]+))?)?' + _VIDEO_RE = _VIDEO_RE_TPL % r'(?P<id>[0-9A-Za-z_-]{11})' IE_NAME = 'youtube:playlist' _TESTS = [{ 'url': 'https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re', @@ -2603,6 +2612,34 @@ class YoutubePlaylistIE(YoutubePlaylistBaseInfoExtractor): def _real_initialize(self): self._login() + def extract_videos_from_page(self, page): + ids_in_page = [] + titles_in_page = [] + + for item in re.findall( + r'(<[^>]*\bdata-video-id\s*=\s*["\'][0-9A-Za-z_-]{11}[^>]+>)', page): + attrs = extract_attributes(item) + video_id = attrs['data-video-id'] + video_title = unescapeHTML(attrs.get('data-title')) + if video_title: + video_title = video_title.strip() + ids_in_page.append(video_id) + titles_in_page.append(video_title) + + # Fallback with old _VIDEO_RE + self.extract_videos_from_page_impl( + self._VIDEO_RE, page, ids_in_page, titles_in_page) + + # Relaxed fallbacks + self.extract_videos_from_page_impl( + r'href="\s*/watch\?v\s*=\s*(?P<id>[0-9A-Za-z_-]{11})', page, + ids_in_page, titles_in_page) + self.extract_videos_from_page_impl( + r'data-video-ids\s*=\s*["\'](?P<id>[0-9A-Za-z_-]{11})', page, + ids_in_page, titles_in_page) + + return zip(ids_in_page, titles_in_page) + def _extract_mix(self, playlist_id): # The mixes are generated from a single video # the id of the playlist is just 'RD' + video_id