yt_dlp/extractor/linkedin.py

   1 import itertools
   2 import re
   3
   4 from .common import InfoExtractor
   5 from ..utils import (
   6     ExtractorError,
   7     extract_attributes,
   8     float_or_none,
   9     int_or_none,
  10     mimetype2ext,
  11     srt_subtitles_timecode,
  12     traverse_obj,
  13     try_get,
  14     url_or_none,
  15     urlencode_postdata,
  16     urljoin,
  17 )
  18
  19
  20 class LinkedInBaseIE(InfoExtractor):
  21     _NETRC_MACHINE = 'linkedin'
  22     _logged_in = False
  23
  24     def _perform_login(self, username, password):
  25         if self._logged_in:
  26             return
  27
  28         login_page = self._download_webpage(
  29             self._LOGIN_URL, None, 'Downloading login page')
  30         action_url = urljoin(self._LOGIN_URL, self._search_regex(
  31             r'<form[^>]+action=(["\'])(?P<url>.+?)\1', login_page, 'post url',
  32             default='https://www.linkedin.com/uas/login-submit', group='url'))
  33         data = self._hidden_inputs(login_page)
  34         data.update({
  35             'session_key': username,
  36             'session_password': password,
  37         })
  38         login_submit_page = self._download_webpage(
  39             action_url, None, 'Logging in',
  40             data=urlencode_postdata(data))
  41         error = self._search_regex(
  42             r'<span[^>]+class="error"[^>]*>\s*(.+?)\s*</span>',
  43             login_submit_page, 'error', default=None)
  44         if error:
  45             raise ExtractorError(error, expected=True)
  46         LinkedInBaseIE._logged_in = True
  47
  48
  49 class LinkedInLearningBaseIE(LinkedInBaseIE):
  50     _LOGIN_URL = 'https://www.linkedin.com/uas/login?trk=learning'
  51
  52     def _call_api(self, course_slug, fields, video_slug=None, resolution=None):
  53         query = {
  54             'courseSlug': course_slug,
  55             'fields': fields,
  56             'q': 'slugs',
  57         }
  58         sub = ''
  59         if video_slug:
  60             query.update({
  61                 'videoSlug': video_slug,
  62                 'resolution': f'_{resolution}',
  63             })
  64             sub = ' %dp' % resolution
  65         api_url = 'https://www.linkedin.com/learning-api/detailedCourses'
  66         if not self._get_cookies(api_url).get('JSESSIONID'):
  67             self.raise_login_required()
  68         return self._download_json(
  69             api_url, video_slug, f'Downloading{sub} JSON metadata', headers={
  70                 'Csrf-Token': self._get_cookies(api_url)['JSESSIONID'].value,
  71             }, query=query)['elements'][0]
  72
  73     def _get_urn_id(self, video_data):
  74         urn = video_data.get('urn')
  75         if urn:
  76             mobj = re.search(r'urn:li:lyndaCourse:\d+,(\d+)', urn)
  77             if mobj:
  78                 return mobj.group(1)
  79
  80     def _get_video_id(self, video_data, course_slug, video_slug):
  81         return self._get_urn_id(video_data) or f'{course_slug}/{video_slug}'
  82
  83
  84 class LinkedInIE(LinkedInBaseIE):
  85     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/posts/[^/?#]+-(?P<id>\d+)-\w{4}/?(?:[?#]|$)'
  86     _TESTS = [{
  87         'url': 'https://www.linkedin.com/posts/mishalkhawaja_sendinblueviews-toronto-digitalmarketing-ugcPost-6850898786781339649-mM20',
  88         'info_dict': {
  89             'id': '6850898786781339649',
  90             'ext': 'mp4',
  91             'title': 'Mishal K. on LinkedIn: #sendinblueviews #toronto #digitalmarketing #nowhiring #sendinblue…',
  92             'description': 'md5:2998a31f6f479376dd62831f53a80f71',
  93             'uploader': 'Mishal K.',
  94             'thumbnail': 're:^https?://media.licdn.com/dms/image/.*$',
  95             'like_count': int,
  96         },
  97     }, {
  98         'url': 'https://www.linkedin.com/posts/the-mathworks_2_what-is-mathworks-cloud-center-activity-7151241570371948544-4Gu7',
  99         'info_dict': {
 100             'id': '7151241570371948544',
 101             'ext': 'mp4',
 102             'title': 'MathWorks on LinkedIn: What Is MathWorks Cloud Center?',
 103             'description': 'md5:95f9d4eeb6337882fb47eefe13d7a40c',
 104             'uploader': 'MathWorks',
 105             'thumbnail': 're:^https?://media.licdn.com/dms/image/.*$',
 106             'like_count': int,
 107             'subtitles': 'mincount:1',
 108         },
 109     }]
 110
 111     def _real_extract(self, url):
 112         video_id = self._match_id(url)
 113         webpage = self._download_webpage(url, video_id)
 114
 115         video_attrs = extract_attributes(self._search_regex(r'(<video[^>]+>)', webpage, 'video'))
 116         sources = self._parse_json(video_attrs['data-sources'], video_id)
 117         formats = [{
 118             'url': source['src'],
 119             'ext': mimetype2ext(source.get('type')),
 120             'tbr': float_or_none(source.get('data-bitrate'), scale=1000),
 121         } for source in sources]
 122         subtitles = {'en': [{
 123             'url': video_attrs['data-captions-url'],
 124             'ext': 'vtt',
 125         }]} if url_or_none(video_attrs.get('data-captions-url')) else {}
 126
 127         return {
 128             'id': video_id,
 129             'formats': formats,
 130             'title': self._og_search_title(webpage, default=None) or self._html_extract_title(webpage),
 131             'like_count': int_or_none(self._search_regex(
 132                 r'\bdata-num-reactions="(\d+)"', webpage, 'reactions', default=None)),
 133             'uploader': traverse_obj(
 134                 self._yield_json_ld(webpage, video_id),
 135                 (lambda _, v: v['@type'] == 'SocialMediaPosting', 'author', 'name', {str}), get_all=False),
 136             'thumbnail': self._og_search_thumbnail(webpage),
 137             'description': self._og_search_description(webpage, default=None),
 138             'subtitles': subtitles,
 139         }
 140
 141
 142 class LinkedInLearningIE(LinkedInLearningBaseIE):
 143     IE_NAME = 'linkedin:learning'
 144     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/learning/(?P<course_slug>[^/]+)/(?P<id>[^/?#]+)'
 145     _TEST = {
 146         'url': 'https://www.linkedin.com/learning/programming-foundations-fundamentals/welcome?autoplay=true',
 147         'md5': 'a1d74422ff0d5e66a792deb996693167',
 148         'info_dict': {
 149             'id': '90426',
 150             'ext': 'mp4',
 151             'title': 'Welcome',
 152             'timestamp': 1430396150.82,
 153             'upload_date': '20150430',
 154         },
 155     }
 156
 157     def json2srt(self, transcript_lines, duration=None):
 158         srt_data = ''
 159         for line, (line_dict, next_dict) in enumerate(itertools.zip_longest(transcript_lines, transcript_lines[1:])):
 160             start_time, caption = line_dict['transcriptStartAt'] / 1000, line_dict['caption']
 161             end_time = next_dict['transcriptStartAt'] / 1000 if next_dict else duration or start_time + 1
 162             srt_data += (
 163                 f'{line + 1}\n'
 164                 f'{srt_subtitles_timecode(start_time)} --> {srt_subtitles_timecode(end_time)}\n'
 165                 f'{caption}\n\n')
 166         return srt_data
 167
 168     def _real_extract(self, url):
 169         course_slug, video_slug = self._match_valid_url(url).groups()
 170
 171         formats = []
 172         for width, height in ((640, 360), (960, 540), (1280, 720)):
 173             video_data = self._call_api(
 174                 course_slug, 'selectedVideo', video_slug, height)['selectedVideo']
 175
 176             video_url_data = video_data.get('url') or {}
 177             progressive_url = video_url_data.get('progressiveUrl')
 178             if progressive_url:
 179                 formats.append({
 180                     'format_id': f'progressive-{height}p',
 181                     'url': progressive_url,
 182                     'ext': 'mp4',
 183                     'height': height,
 184                     'width': width,
 185                     'source_preference': 1,
 186                 })
 187
 188         title = video_data['title']
 189
 190         audio_url = video_data.get('audio', {}).get('progressiveUrl')
 191         if audio_url:
 192             formats.append({
 193                 'abr': 64,
 194                 'ext': 'm4a',
 195                 'format_id': 'audio',
 196                 'url': audio_url,
 197                 'vcodec': 'none',
 198             })
 199
 200         streaming_url = video_url_data.get('streamingUrl')
 201         if streaming_url:
 202             formats.extend(self._extract_m3u8_formats(
 203                 streaming_url, video_slug, 'mp4',
 204                 'm3u8_native', m3u8_id='hls', fatal=False))
 205
 206         subtitles = {}
 207         duration = int_or_none(video_data.get('durationInSeconds'))
 208         transcript_lines = try_get(video_data, lambda x: x['transcript']['lines'], expected_type=list)
 209         if transcript_lines:
 210             subtitles['en'] = [{
 211                 'ext': 'srt',
 212                 'data': self.json2srt(transcript_lines, duration),
 213             }]
 214
 215         return {
 216             'id': self._get_video_id(video_data, course_slug, video_slug),
 217             'title': title,
 218             'formats': formats,
 219             'thumbnail': video_data.get('defaultThumbnail'),
 220             'timestamp': float_or_none(video_data.get('publishedOn'), 1000),
 221             'duration': duration,
 222             'subtitles': subtitles,
 223             # It seems like this would be correctly handled by default
 224             # However, unless someone can confirm this, the old
 225             # behaviour is being kept as-is
 226             '_format_sort_fields': ('res', 'source_preference'),
 227         }
 228
 229
 230 class LinkedInLearningCourseIE(LinkedInLearningBaseIE):
 231     IE_NAME = 'linkedin:learning:course'
 232     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/learning/(?P<id>[^/?#]+)'
 233     _TEST = {
 234         'url': 'https://www.linkedin.com/learning/programming-foundations-fundamentals',
 235         'info_dict': {
 236             'id': 'programming-foundations-fundamentals',
 237             'title': 'Programming Foundations: Fundamentals',
 238             'description': 'md5:76e580b017694eb89dc8e8923fff5c86',
 239         },
 240         'playlist_mincount': 61,
 241     }
 242
 243     @classmethod
 244     def suitable(cls, url):
 245         return False if LinkedInLearningIE.suitable(url) else super().suitable(url)
 246
 247     def _real_extract(self, url):
 248         course_slug = self._match_id(url)
 249         course_data = self._call_api(course_slug, 'chapters,description,title')
 250
 251         entries = []
 252         for chapter_number, chapter in enumerate(course_data.get('chapters', []), 1):
 253             chapter_title = chapter.get('title')
 254             chapter_id = self._get_urn_id(chapter)
 255             for video in chapter.get('videos', []):
 256                 video_slug = video.get('slug')
 257                 if not video_slug:
 258                     continue
 259                 entries.append({
 260                     '_type': 'url_transparent',
 261                     'id': self._get_video_id(video, course_slug, video_slug),
 262                     'title': video.get('title'),
 263                     'url': f'https://www.linkedin.com/learning/{course_slug}/{video_slug}',
 264                     'chapter': chapter_title,
 265                     'chapter_number': chapter_number,
 266                     'chapter_id': chapter_id,
 267                     'ie_key': LinkedInLearningIE.ie_key(),
 268                 })
 269
 270         return self.playlist_result(
 271             entries, course_slug,
 272             course_data.get('title'),
 273             course_data.get('description'))