From ab29e47029e2f5b48abbbab78e82faf7cf6e9506 Mon Sep 17 00:00:00 2001 From: qbnu <93988953+qbnu@users.noreply.github.com> Date: Tue, 18 Apr 2023 02:37:37 +0000 Subject: [PATCH] [extractor/bilibili] Support festival videos (#6547) Closes #6138 Authored by: qbnu --- yt_dlp/extractor/bilibili.py | 86 ++++++++++++++++++++++++++++++------ 1 file changed, 73 insertions(+), 13 deletions(-) diff --git a/yt_dlp/extractor/bilibili.py b/yt_dlp/extractor/bilibili.py index 91d436dd8..faa2218ce 100644 --- a/yt_dlp/extractor/bilibili.py +++ b/yt_dlp/extractor/bilibili.py @@ -134,7 +134,7 @@ def _get_all_children(self, reply): class BiliBiliIE(BilibiliBaseIE): - _VALID_URL = r'https?://www\.bilibili\.com/video/[aAbB][vV](?P[^/?#&]+)' + _VALID_URL = r'https?://www\.bilibili\.com/(?:video/|festival/\w+\?(?:[^#]*&)?bvid=)[aAbB][vV](?P[^/?#&]+)' _TESTS = [{ 'url': 'https://www.bilibili.com/video/BV13x41117TL', @@ -282,19 +282,60 @@ class BiliBiliIE(BilibiliBaseIE): 'thumbnail': r're:^https?://.*\.(jpg|jpeg|png)$', }, 'params': {'skip_download': True}, + }, { + 'note': 'video redirects to festival page', + 'url': 'https://www.bilibili.com/video/BV1wP4y1P72h', + 'info_dict': { + 'id': 'BV1wP4y1P72h', + 'ext': 'mp4', + 'title': '牛虎年相交之际，一首传统民族打击乐《牛斗虎》祝大家新春快乐，虎年大吉！【bilibili音乐虎闹新春】', + 'timestamp': 1643947497, + 'upload_date': '20220204', + 'description': 'md5:8681a0d4d2c06b4ae27e59c8080a7fe6', + 'uploader': '叨叨冯聊音乐', + 'duration': 246.719, + 'uploader_id': '528182630', + 'view_count': int, + 'like_count': int, + 'thumbnail': r're:^https?://.*\.(jpg|jpeg|png)$', + }, + 'params': {'skip_download': True}, + }, { + 'note': 'newer festival video', + 'url': 'https://www.bilibili.com/festival/2023honkaiimpact3gala?bvid=BV1ay4y1d77f', + 'info_dict': { + 'id': 'BV1ay4y1d77f', + 'ext': 'mp4', + 'title': '【崩坏3新春剧场】为特别的你送上祝福！', + 'timestamp': 1674273600, + 'upload_date': '20230121', + 'description': 'md5:58af66d15c6a0122dc30c8adfd828dd8', + 'uploader': '果蝇轰', + 'duration': 1111.722, + 'uploader_id': '8469526', + 'view_count': int, + 'like_count': int, + 'thumbnail': r're:^https?://.*\.(jpg|jpeg|png)$', + }, + 'params': {'skip_download': True}, }] def _real_extract(self, url): video_id = self._match_id(url) webpage = self._download_webpage(url, video_id) initial_state = self._search_json(r'window\.__INITIAL_STATE__\s*=', webpage, 'initial state', video_id) - play_info = self._search_json(r'window\.__playinfo__\s*=', webpage, 'play info', video_id)['data'] - video_data = initial_state['videoData'] + is_festival = 'videoData' not in initial_state + if is_festival: + video_data = initial_state['videoInfo'] + else: + play_info = self._search_json(r'window\.__playinfo__\s*=', webpage, 'play info', video_id)['data'] + video_data = initial_state['videoData'] + video_id, title = video_data['bvid'], video_data.get('title') # Bilibili anthologies are similar to playlists but all videos share the same video ID as the anthology itself. - page_list_json = traverse_obj( + page_list_json = not is_festival and traverse_obj( self._download_json( 'https://api.bilibili.com/x/player/pagelist', video_id, fatal=False, query={'bvid': video_id, 'jsonp': 'jsonp'}, @@ -317,20 +358,39 @@ def _real_extract(self, url): cid = traverse_obj(video_data, ('pages', part_id - 1, 'cid')) if part_id else video_data.get('cid') + festival_info = {} + if is_festival: + play_info = self._download_json( + 'https://api.bilibili.com/x/player/playurl', video_id, + query={'bvid': video_id, 'cid': cid, 'fnval': 4048}, + note='Extracting festival video formats')['data'] + + festival_info = traverse_obj(initial_state, { + 'uploader': ('videoInfo', 'upName'), + 'uploader_id': ('videoInfo', 'upMid', {str_or_none}), + 'like_count': ('videoStatus', 'like', {int_or_none}), + 'thumbnail': ('sectionEpisodes', lambda _, v: v['bvid'] == video_id, 'cover'), + }, get_all=False) + return { + **traverse_obj(initial_state, { + 'uploader': ('upData', 'name'), + 'uploader_id': ('upData', 'mid', {str_or_none}), + 'like_count': ('videoData', 'stat', 'like', {int_or_none}), + 'tags': ('tags', ..., 'tag_name'), + 'thumbnail': ('videoData', 'pic', {url_or_none}), + }), + **festival_info, + **traverse_obj(video_data, { + 'description': 'desc', + 'timestamp': ('pubdate', {int_or_none}), + 'view_count': (('viewCount', ('stat', 'view')), {int_or_none}), + 'comment_count': ('stat', 'reply', {int_or_none}), + }, get_all=False), 'id': f'{video_id}{format_field(part_id, None, "_p%d")}', 'formats': self.extract_formats(play_info), '_old_archive_ids': [make_archive_id(self, old_video_id)] if old_video_id else None, 'title': title, - 'description': traverse_obj(initial_state, ('videoData', 'desc')), - 'view_count': traverse_obj(initial_state, ('videoData', 'stat', 'view')), - 'uploader': traverse_obj(initial_state, ('upData', 'name')), - 'uploader_id': traverse_obj(initial_state, ('upData', 'mid')), - 'like_count': traverse_obj(initial_state, ('videoData', 'stat', 'like')), - 'comment_count': traverse_obj(initial_state, ('videoData', 'stat', 'reply')), - 'tags': traverse_obj(initial_state, ('tags', ..., 'tag_name')), - 'thumbnail': traverse_obj(initial_state, ('videoData', 'pic')), - 'timestamp': traverse_obj(initial_state, ('videoData', 'pubdate')), 'duration': float_or_none(play_info.get('timelength'), scale=1000), 'chapters': self._get_chapters(aid, cid), 'subtitles': self.extract_subtitles(video_id, aid, cid),