yt_dlp/extractor/tiktok.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import itertools
   5 import random
   6 import string
   7 import time
   8
   9 from .common import InfoExtractor
  10 from ..utils import (
  11     ExtractorError,
  12     int_or_none,
  13     str_or_none,
  14     traverse_obj,
  15     try_get,
  16     qualities,
  17 )
  18
  19
  20 class TikTokIE(InfoExtractor):
  21     _VALID_URL = r'https?://www\.tiktok\.com/@[\w\.-]+/video/(?P<id>\d+)'
  22
  23     _TESTS = [{
  24         'url': 'https://www.tiktok.com/@leenabhushan/video/6748451240264420610',
  25         'md5': '34a7543afd5a151b0840ba6736fb633b',
  26         'info_dict': {
  27             'id': '6748451240264420610',
  28             'ext': 'mp4',
  29             'title': '#jassmanak #lehanga #leenabhushan',
  30             'description': '#jassmanak #lehanga #leenabhushan',
  31             'duration': 13,
  32             'height': 1280,
  33             'width': 720,
  34             'uploader': 'leenabhushan',
  35             'uploader_id': '6691488002098119685',
  36             'uploader_url': 'https://www.tiktok.com/@leenabhushan',
  37             'creator': 'facestoriesbyleenabh',
  38             'thumbnail': r're:^https?://[\w\/\.\-]+(~[\w\-]+\.image)?',
  39             'upload_date': '20191016',
  40             'timestamp': 1571246252,
  41             'view_count': int,
  42             'like_count': int,
  43             'repost_count': int,
  44             'comment_count': int,
  45         }
  46     }, {
  47         'url': 'https://www.tiktok.com/@patroxofficial/video/6742501081818877190?langCountry=en',
  48         'md5': '06b9800d47d5fe51a19e322dd86e61c9',
  49         'info_dict': {
  50             'id': '6742501081818877190',
  51             'ext': 'mp4',
  52             'title': 'md5:5e2a23877420bb85ce6521dbee39ba94',
  53             'description': 'md5:5e2a23877420bb85ce6521dbee39ba94',
  54             'duration': 27,
  55             'height': 960,
  56             'width': 540,
  57             'uploader': 'patrox',
  58             'uploader_id': '18702747',
  59             'uploader_url': 'https://www.tiktok.com/@patrox',
  60             'creator': 'patroX',
  61             'thumbnail': r're:^https?://[\w\/\.\-]+(~[\w\-]+\.image)?',
  62             'upload_date': '20190930',
  63             'timestamp': 1569860870,
  64             'view_count': int,
  65             'like_count': int,
  66             'repost_count': int,
  67             'comment_count': int,
  68         }
  69     }, {
  70         # Promoted content/ad
  71         'url': 'https://www.tiktok.com/@MS4wLjABAAAAAR29F6J2Ktu0Daw03BJyXPNoRQ-W7U5a0Mn3lVCq2rQhjOd_WNLclHUoFgwX8Eno/video/6932675057474981122',
  72         'only_matching': True,
  73     }]
  74     _APP_VERSION = '20.9.3'
  75     _MANIFEST_APP_VERSION = '291'
  76     QUALITIES = ('360p', '540p', '720p')
  77
  78     def _extract_aweme(self, props_data, webpage, url):
  79         video_info = try_get(
  80             props_data, lambda x: x['pageProps']['itemInfo']['itemStruct'], dict)
  81         author_info = try_get(
  82             props_data, lambda x: x['pageProps']['itemInfo']['itemStruct']['author'], dict) or {}
  83         music_info = try_get(
  84             props_data, lambda x: x['pageProps']['itemInfo']['itemStruct']['music'], dict) or {}
  85         stats_info = try_get(props_data, lambda x: x['pageProps']['itemInfo']['itemStruct']['stats'], dict) or {}
  86
  87         user_id = str_or_none(author_info.get('uniqueId'))
  88         download_url = try_get(video_info, (lambda x: x['video']['playAddr'],
  89                                             lambda x: x['video']['downloadAddr']))
  90         height = try_get(video_info, lambda x: x['video']['height'], int)
  91         width = try_get(video_info, lambda x: x['video']['width'], int)
  92         thumbnails = [{
  93             'url': video_info.get('thumbnail') or self._og_search_thumbnail(webpage),
  94             'width': width,
  95             'height': height
  96         }]
  97         tracker = try_get(props_data, lambda x: x['initialProps']['$wid'])
  98
  99         return {
 100             'id': str_or_none(video_info.get('id')),
 101             'url': download_url,
 102             'ext': 'mp4',
 103             'height': height,
 104             'width': width,
 105             'title': video_info.get('desc') or self._og_search_title(webpage),
 106             'duration': try_get(video_info, lambda x: x['video']['duration'], int),
 107             'view_count': int_or_none(stats_info.get('playCount')),
 108             'like_count': int_or_none(stats_info.get('diggCount')),
 109             'repost_count': int_or_none(stats_info.get('shareCount')),
 110             'comment_count': int_or_none(stats_info.get('commentCount')),
 111             'timestamp': try_get(video_info, lambda x: int(x['createTime']), int),
 112             'creator': str_or_none(author_info.get('nickname')),
 113             'uploader': user_id,
 114             'uploader_id': str_or_none(author_info.get('id')),
 115             'uploader_url': f'https://www.tiktok.com/@{user_id}',
 116             'track': str_or_none(music_info.get('title')),
 117             'album': str_or_none(music_info.get('album')) or None,
 118             'artist': str_or_none(music_info.get('authorName')),
 119             'thumbnails': thumbnails,
 120             'description': str_or_none(video_info.get('desc')),
 121             'webpage_url': self._og_search_url(webpage),
 122             'http_headers': {
 123                 'Referer': url,
 124                 'Cookie': 'tt_webid=%s; tt_webid_v2=%s' % (tracker, tracker),
 125             }
 126         }
 127
 128     def _extract_aweme_app(self, aweme_id):
 129         query = {
 130             'aweme_id': aweme_id,
 131             'version_name': self._APP_VERSION,
 132             'version_code': self._MANIFEST_APP_VERSION,
 133             'build_number': self._APP_VERSION,
 134             'manifest_version_code': self._MANIFEST_APP_VERSION,
 135             'update_version_code': self._MANIFEST_APP_VERSION,
 136             'openudid': ''.join(random.choice('0123456789abcdef') for i in range(16)),
 137             'uuid': ''.join([random.choice(string.digits) for num in range(16)]),
 138             '_rticket': int(time.time() * 1000),
 139             'ts': int(time.time()),
 140             'device_brand': 'Google',
 141             'device_type': 'Pixel 4',
 142             'device_platform': 'android',
 143             'resolution': '1080*1920',
 144             'dpi': 420,
 145             'os_version': '10',
 146             'os_api': '29',
 147             'carrier_region': 'US',
 148             'sys_region': 'US',
 149             'region': 'US',
 150             'app_name': 'trill',
 151             'app_language': 'en',
 152             'language': 'en',
 153             'timezone_name': 'America/New_York',
 154             'timezone_offset': '-14400',
 155             'channel': 'googleplay',
 156             'ac': 'wifi',
 157             'mcc_mnc': '310260',
 158             'is_my_cn': 0,
 159             'aid': 1180,
 160             'ssmix': 'a',
 161             'as': 'a1qwert123',
 162             'cp': 'cbfhckdckkde1',
 163         }
 164
 165         self._set_cookie('.tiktokv.com', 'odin_tt', ''.join(random.choice('0123456789abcdef') for i in range(160)))
 166
 167         aweme_detail = self._download_json(
 168             'https://api-t2.tiktokv.com/aweme/v1/aweme/detail/', aweme_id,
 169             'Downloading video details', 'Unable to download video details',
 170             headers={
 171                 'User-Agent': f'com.ss.android.ugc.trill/{self._MANIFEST_APP_VERSION} (Linux; U; Android 10; en_US; Pixel 4; Build/QQ3A.200805.001; Cronet/58.0.2991.0)',
 172             }, query=query)['aweme_detail']
 173         video_info = aweme_detail['video']
 174
 175         def parse_url_key(url_key):
 176             format_id, codec, res, bitrate = self._search_regex(
 177                 r'v[^_]+_(?P<id>(?P<codec>[^_]+)_(?P<res>\d+p)_(?P<bitrate>\d+))', url_key,
 178                 'url key', default=(None, None, None, None), group=('id', 'codec', 'res', 'bitrate'))
 179             if not format_id:
 180                 return {}, None
 181             return {
 182                 'format_id': format_id,
 183                 'vcodec': 'h265' if codec == 'bytevc1' else codec,
 184                 'tbr': int_or_none(bitrate, scale=1000) or None,
 185                 'quality': qualities(self.QUALITIES)(res),
 186             }, res
 187
 188         known_resolutions = {}
 189
 190         def extract_addr(addr, add_meta={}):
 191             parsed_meta, res = parse_url_key(addr.get('url_key', ''))
 192             if res:
 193                 known_resolutions.setdefault(res, {}).setdefault('height', add_meta.get('height'))
 194                 known_resolutions[res].setdefault('width', add_meta.get('width'))
 195                 parsed_meta.update(known_resolutions.get(res, {}))
 196                 add_meta.setdefault('height', int_or_none(res[:-1]))
 197             return [{
 198                 'url': url,
 199                 'filesize': int_or_none(addr.get('data_size')),
 200                 'ext': 'mp4',
 201                 'acodec': 'aac',
 202                 **add_meta, **parsed_meta
 203             } for url in addr.get('url_list') or []]
 204
 205         # Hack: Add direct video links first to prioritize them when removing duplicate formats
 206         formats = []
 207         if video_info.get('play_addr'):
 208             formats.extend(extract_addr(video_info['play_addr'], {
 209                 'format_id': 'play_addr',
 210                 'format_note': 'Direct video',
 211                 'vcodec': 'h265' if traverse_obj(
 212                     video_info, 'is_bytevc1', 'is_h265') else 'h264',  # Always h264?
 213                 'width': video_info.get('width'),
 214                 'height': video_info.get('height'),
 215             }))
 216         if video_info.get('download_addr'):
 217             formats.extend(extract_addr(video_info['download_addr'], {
 218                 'format_id': 'download_addr',
 219                 'format_note': 'Download video%s' % (', watermarked' if video_info.get('has_watermark') else ''),
 220                 'vcodec': 'h264',
 221                 'width': video_info.get('width'),
 222                 'height': video_info.get('height'),
 223                 'source_preference': -2 if video_info.get('has_watermark') else -1,
 224             }))
 225         if video_info.get('play_addr_h264'):
 226             formats.extend(extract_addr(video_info['play_addr_h264'], {
 227                 'format_id': 'play_addr_h264',
 228                 'format_note': 'Direct video',
 229                 'vcodec': 'h264',
 230             }))
 231         if video_info.get('play_addr_bytevc1'):
 232             formats.extend(extract_addr(video_info['play_addr_bytevc1'], {
 233                 'format_id': 'play_addr_bytevc1',
 234                 'format_note': 'Direct video',
 235                 'vcodec': 'h265',
 236             }))
 237
 238         for bitrate in video_info.get('bit_rate', []):
 239             if bitrate.get('play_addr'):
 240                 formats.extend(extract_addr(bitrate['play_addr'], {
 241                     'format_id': bitrate.get('gear_name'),
 242                     'format_note': 'Playback video',
 243                     'tbr': try_get(bitrate, lambda x: x['bit_rate'] / 1000),
 244                     'vcodec': 'h265' if traverse_obj(
 245                         bitrate, 'is_bytevc1', 'is_h265') else 'h264',
 246                 }))
 247
 248         self._remove_duplicate_formats(formats)
 249         self._sort_formats(formats, ('quality', 'source', 'codec', 'size', 'br'))
 250
 251         thumbnails = []
 252         for cover_id in ('cover', 'ai_dynamic_cover', 'animated_cover', 'ai_dynamic_cover_bak',
 253                          'origin_cover', 'dynamic_cover'):
 254             cover = video_info.get(cover_id)
 255             if cover:
 256                 for cover_url in cover['url_list']:
 257                     thumbnails.append({
 258                         'id': cover_id,
 259                         'url': cover_url,
 260                     })
 261
 262         stats_info = aweme_detail.get('statistics', {})
 263         author_info = aweme_detail.get('author', {})
 264         music_info = aweme_detail.get('music', {})
 265         user_id = str_or_none(author_info.get('nickname'))
 266
 267         contained_music_track = traverse_obj(
 268             music_info, ('matched_song', 'title'), ('matched_pgc_sound', 'title'), expected_type=str)
 269         contained_music_author = traverse_obj(
 270             music_info, ('matched_song', 'author'), ('matched_pgc_sound', 'author'), 'author', expected_type=str)
 271
 272         is_generic_og_trackname = music_info.get('is_original_sound') and music_info.get('title') == 'original sound - %s' % music_info.get('owner_handle')
 273         if is_generic_og_trackname:
 274             music_track, music_author = contained_music_track or 'original sound', contained_music_author
 275         else:
 276             music_track, music_author = music_info.get('title'), music_info.get('author')
 277
 278         return {
 279             'id': aweme_id,
 280             'title': aweme_detail['desc'],
 281             'description': aweme_detail['desc'],
 282             'view_count': int_or_none(stats_info.get('play_count')),
 283             'like_count': int_or_none(stats_info.get('digg_count')),
 284             'repost_count': int_or_none(stats_info.get('share_count')),
 285             'comment_count': int_or_none(stats_info.get('comment_count')),
 286             'uploader': str_or_none(author_info.get('unique_id')),
 287             'creator': user_id,
 288             'uploader_id': str_or_none(author_info.get('uid')),
 289             'uploader_url': f'https://www.tiktok.com/@{user_id}' if user_id else None,
 290             'track': music_track,
 291             'album': str_or_none(music_info.get('album')) or None,
 292             'artist': music_author,
 293             'timestamp': int_or_none(aweme_detail.get('create_time')),
 294             'formats': formats,
 295             'thumbnails': thumbnails,
 296             'duration': int_or_none(traverse_obj(video_info, 'duration', ('download_addr', 'duration')), scale=1000)
 297         }
 298
 299     def _real_extract(self, url):
 300         video_id = self._match_id(url)
 301
 302         try:
 303             return self._extract_aweme_app(video_id)
 304         except ExtractorError as e:
 305             self.report_warning(f'{e}; Retrying with webpage')
 306
 307         # If we only call once, we get a 403 when downlaoding the video.
 308         self._download_webpage(url, video_id)
 309         webpage = self._download_webpage(url, video_id, note='Downloading video webpage')
 310         json_string = self._search_regex(
 311             r'id=\"__NEXT_DATA__\"\s+type=\"application\/json\"\s*[^>]+>\s*(?P<json_string_ld>[^<]+)',
 312             webpage, 'json_string', group='json_string_ld')
 313         json_data = self._parse_json(json_string, video_id)
 314         props_data = try_get(json_data, lambda x: x['props'], expected_type=dict)
 315
 316         # Chech statusCode for success
 317         status = props_data.get('pageProps').get('statusCode')
 318         if status == 0:
 319             return self._extract_aweme(props_data, webpage, url)
 320         elif status == 10216:
 321             raise ExtractorError('This video is private', expected=True)
 322
 323         raise ExtractorError('Video not available', video_id=video_id)
 324
 325
 326 class TikTokUserIE(InfoExtractor):
 327     IE_NAME = 'tiktok:user'
 328     _VALID_URL = r'https?://(?:www\.)?tiktok\.com/@(?P<id>[\w\._]+)/?(?:$|[#?])'
 329     _TESTS = [{
 330         'url': 'https://tiktok.com/@corgibobaa?lang=en',
 331         'playlist_mincount': 45,
 332         'info_dict': {
 333             'id': '6935371178089399301',
 334         },
 335         'skip': 'Cookies (not necessarily logged in) are needed.'
 336     }, {
 337         'url': 'https://www.tiktok.com/@meme',
 338         'playlist_mincount': 593,
 339         'info_dict': {
 340             'id': '79005827461758976',
 341         },
 342         'skip': 'Cookies (not necessarily logged in) are needed.'
 343     }]
 344
 345     def _entries(self, url, user_id):
 346         webpage = self._download_webpage(url, user_id)
 347         own_id = self._search_regex(r'\"id\":\"(?P<userid>\d+)', webpage, user_id, default=None)
 348         if not own_id:
 349             raise ExtractorError('Cookies (not necessarily logged in) are needed.', expected=True)
 350         secuid = self._search_regex(r'\"secUid\":\"(?P<secUid>[^\"]+)', webpage, user_id)
 351         verifyfp_cookie = self._get_cookies('https://www.tiktok.com').get('s_v_web_id')
 352         if not verifyfp_cookie:
 353             raise ExtractorError('Improper cookies (missing s_v_web_id).', expected=True)
 354         api_url = f'https://m.tiktok.com/api/post/item_list/?aid=1988&cookie_enabled=true&count=30&verifyFp={verifyfp_cookie.value}&secUid={secuid}&cursor='
 355         cursor = '0'
 356         for page in itertools.count():
 357             data_json = self._download_json(api_url + cursor, user_id, note='Downloading Page %d' % page)
 358             for video in data_json.get('itemList', []):
 359                 video_id = video['id']
 360                 video_url = f'https://www.tiktok.com/@{user_id}/video/{video_id}'
 361                 yield self._url_result(video_url, 'TikTok', video_id, str_or_none(video.get('desc')))
 362             if not data_json.get('hasMore'):
 363                 break
 364             cursor = data_json['cursor']
 365
 366     def _real_extract(self, url):
 367         user_id = self._match_id(url)
 368         return self.playlist_result(self._entries(url, user_id), user_id)