yt_dlp/extractor/linkedin.py

   1 from itertools import zip_longest
   2 import re
   3
   4 from .common import InfoExtractor
   5 from ..utils import (
   6     clean_html,
   7     extract_attributes,
   8     ExtractorError,
   9     float_or_none,
  10     get_element_by_class,
  11     int_or_none,
  12     srt_subtitles_timecode,
  13     strip_or_none,
  14     mimetype2ext,
  15     try_get,
  16     urlencode_postdata,
  17     urljoin,
  18 )
  19
  20
  21 class LinkedInBaseIE(InfoExtractor):
  22     _NETRC_MACHINE = 'linkedin'
  23     _logged_in = False
  24
  25     def _perform_login(self, username, password):
  26         if self._logged_in:
  27             return
  28
  29         login_page = self._download_webpage(
  30             self._LOGIN_URL, None, 'Downloading login page')
  31         action_url = urljoin(self._LOGIN_URL, self._search_regex(
  32             r'<form[^>]+action=(["\'])(?P<url>.+?)\1', login_page, 'post url',
  33             default='https://www.linkedin.com/uas/login-submit', group='url'))
  34         data = self._hidden_inputs(login_page)
  35         data.update({
  36             'session_key': username,
  37             'session_password': password,
  38         })
  39         login_submit_page = self._download_webpage(
  40             action_url, None, 'Logging in',
  41             data=urlencode_postdata(data))
  42         error = self._search_regex(
  43             r'<span[^>]+class="error"[^>]*>\s*(.+?)\s*</span>',
  44             login_submit_page, 'error', default=None)
  45         if error:
  46             raise ExtractorError(error, expected=True)
  47         LinkedInBaseIE._logged_in = True
  48
  49
  50 class LinkedInLearningBaseIE(LinkedInBaseIE):
  51     _LOGIN_URL = 'https://www.linkedin.com/uas/login?trk=learning'
  52
  53     def _call_api(self, course_slug, fields, video_slug=None, resolution=None):
  54         query = {
  55             'courseSlug': course_slug,
  56             'fields': fields,
  57             'q': 'slugs',
  58         }
  59         sub = ''
  60         if video_slug:
  61             query.update({
  62                 'videoSlug': video_slug,
  63                 'resolution': '_%s' % resolution,
  64             })
  65             sub = ' %dp' % resolution
  66         api_url = 'https://www.linkedin.com/learning-api/detailedCourses'
  67         if not self._get_cookies(api_url).get('JSESSIONID'):
  68             self.raise_login_required()
  69         return self._download_json(
  70             api_url, video_slug, 'Downloading%s JSON metadata' % sub, headers={
  71                 'Csrf-Token': self._get_cookies(api_url)['JSESSIONID'].value,
  72             }, query=query)['elements'][0]
  73
  74     def _get_urn_id(self, video_data):
  75         urn = video_data.get('urn')
  76         if urn:
  77             mobj = re.search(r'urn:li:lyndaCourse:\d+,(\d+)', urn)
  78             if mobj:
  79                 return mobj.group(1)
  80
  81     def _get_video_id(self, video_data, course_slug, video_slug):
  82         return self._get_urn_id(video_data) or '%s/%s' % (course_slug, video_slug)
  83
  84
  85 class LinkedInIE(LinkedInBaseIE):
  86     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/posts/.+?(?P<id>\d+)'
  87     _TESTS = [{
  88         'url': 'https://www.linkedin.com/posts/mishalkhawaja_sendinblueviews-toronto-digitalmarketing-ugcPost-6850898786781339649-mM20',
  89         'info_dict': {
  90             'id': '6850898786781339649',
  91             'ext': 'mp4',
  92             'title': 'Mishal K. on LinkedIn: #sendinblueviews #toronto #digitalmarketing',
  93             'description': 'md5:be125430bab1c574f16aeb186a4d5b19',
  94             'creator': 'Mishal K.'
  95         },
  96     }]
  97
  98     def _real_extract(self, url):
  99         video_id = self._match_id(url)
 100         webpage = self._download_webpage(url, video_id)
 101
 102         title = self._html_extract_title(webpage)
 103         description = clean_html(get_element_by_class('share-update-card__update-text', webpage))
 104         like_count = int_or_none(get_element_by_class('social-counts-reactions__social-counts-numRections', webpage))
 105         creator = strip_or_none(clean_html(get_element_by_class('comment__actor-name', webpage)))
 106
 107         sources = self._parse_json(extract_attributes(self._search_regex(r'(<video[^>]+>)', webpage, 'video'))['data-sources'], video_id)
 108         formats = [{
 109             'url': source['src'],
 110             'ext': mimetype2ext(source.get('type')),
 111             'tbr': float_or_none(source.get('data-bitrate'), scale=1000),
 112         } for source in sources]
 113
 114         return {
 115             'id': video_id,
 116             'formats': formats,
 117             'title': title,
 118             'like_count': like_count,
 119             'creator': creator,
 120             'thumbnail': self._og_search_thumbnail(webpage),
 121             'description': description,
 122         }
 123
 124
 125 class LinkedInLearningIE(LinkedInLearningBaseIE):
 126     IE_NAME = 'linkedin:learning'
 127     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/learning/(?P<course_slug>[^/]+)/(?P<id>[^/?#]+)'
 128     _TEST = {
 129         'url': 'https://www.linkedin.com/learning/programming-foundations-fundamentals/welcome?autoplay=true',
 130         'md5': 'a1d74422ff0d5e66a792deb996693167',
 131         'info_dict': {
 132             'id': '90426',
 133             'ext': 'mp4',
 134             'title': 'Welcome',
 135             'timestamp': 1430396150.82,
 136             'upload_date': '20150430',
 137         },
 138     }
 139
 140     def json2srt(self, transcript_lines, duration=None):
 141         srt_data = ''
 142         for line, (line_dict, next_dict) in enumerate(zip_longest(transcript_lines, transcript_lines[1:])):
 143             start_time, caption = line_dict['transcriptStartAt'] / 1000, line_dict['caption']
 144             end_time = next_dict['transcriptStartAt'] / 1000 if next_dict else duration or start_time + 1
 145             srt_data += '%d\n%s --> %s\n%s\n\n' % (line + 1, srt_subtitles_timecode(start_time),
 146                                                    srt_subtitles_timecode(end_time),
 147                                                    caption)
 148         return srt_data
 149
 150     def _real_extract(self, url):
 151         course_slug, video_slug = self._match_valid_url(url).groups()
 152
 153         formats = []
 154         for width, height in ((640, 360), (960, 540), (1280, 720)):
 155             video_data = self._call_api(
 156                 course_slug, 'selectedVideo', video_slug, height)['selectedVideo']
 157
 158             video_url_data = video_data.get('url') or {}
 159             progressive_url = video_url_data.get('progressiveUrl')
 160             if progressive_url:
 161                 formats.append({
 162                     'format_id': 'progressive-%dp' % height,
 163                     'url': progressive_url,
 164                     'ext': 'mp4',
 165                     'height': height,
 166                     'width': width,
 167                     'source_preference': 1,
 168                 })
 169
 170         title = video_data['title']
 171
 172         audio_url = video_data.get('audio', {}).get('progressiveUrl')
 173         if audio_url:
 174             formats.append({
 175                 'abr': 64,
 176                 'ext': 'm4a',
 177                 'format_id': 'audio',
 178                 'url': audio_url,
 179                 'vcodec': 'none',
 180             })
 181
 182         streaming_url = video_url_data.get('streamingUrl')
 183         if streaming_url:
 184             formats.extend(self._extract_m3u8_formats(
 185                 streaming_url, video_slug, 'mp4',
 186                 'm3u8_native', m3u8_id='hls', fatal=False))
 187
 188         subtitles = {}
 189         duration = int_or_none(video_data.get('durationInSeconds'))
 190         transcript_lines = try_get(video_data, lambda x: x['transcript']['lines'], expected_type=list)
 191         if transcript_lines:
 192             subtitles['en'] = [{
 193                 'ext': 'srt',
 194                 'data': self.json2srt(transcript_lines, duration)
 195             }]
 196
 197         return {
 198             'id': self._get_video_id(video_data, course_slug, video_slug),
 199             'title': title,
 200             'formats': formats,
 201             'thumbnail': video_data.get('defaultThumbnail'),
 202             'timestamp': float_or_none(video_data.get('publishedOn'), 1000),
 203             'duration': duration,
 204             'subtitles': subtitles,
 205             # It seems like this would be correctly handled by default
 206             # However, unless someone can confirm this, the old
 207             # behaviour is being kept as-is
 208             '_format_sort_fields': ('res', 'source_preference')
 209         }
 210
 211
 212 class LinkedInLearningCourseIE(LinkedInLearningBaseIE):
 213     IE_NAME = 'linkedin:learning:course'
 214     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/learning/(?P<id>[^/?#]+)'
 215     _TEST = {
 216         'url': 'https://www.linkedin.com/learning/programming-foundations-fundamentals',
 217         'info_dict': {
 218             'id': 'programming-foundations-fundamentals',
 219             'title': 'Programming Foundations: Fundamentals',
 220             'description': 'md5:76e580b017694eb89dc8e8923fff5c86',
 221         },
 222         'playlist_mincount': 61,
 223     }
 224
 225     @classmethod
 226     def suitable(cls, url):
 227         return False if LinkedInLearningIE.suitable(url) else super(LinkedInLearningCourseIE, cls).suitable(url)
 228
 229     def _real_extract(self, url):
 230         course_slug = self._match_id(url)
 231         course_data = self._call_api(course_slug, 'chapters,description,title')
 232
 233         entries = []
 234         for chapter_number, chapter in enumerate(course_data.get('chapters', []), 1):
 235             chapter_title = chapter.get('title')
 236             chapter_id = self._get_urn_id(chapter)
 237             for video in chapter.get('videos', []):
 238                 video_slug = video.get('slug')
 239                 if not video_slug:
 240                     continue
 241                 entries.append({
 242                     '_type': 'url_transparent',
 243                     'id': self._get_video_id(video, course_slug, video_slug),
 244                     'title': video.get('title'),
 245                     'url': 'https://www.linkedin.com/learning/%s/%s' % (course_slug, video_slug),
 246                     'chapter': chapter_title,
 247                     'chapter_number': chapter_number,
 248                     'chapter_id': chapter_id,
 249                     'ie_key': LinkedInLearningIE.ie_key(),
 250                 })
 251
 252         return self.playlist_result(
 253             entries, course_slug,
 254             course_data.get('title'),
 255             course_data.get('description'))