yt_dlp/extractor/linkedin.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 from itertools import zip_longest
   5 import re
   6
   7 from .common import InfoExtractor
   8 from ..utils import (
   9     clean_html,
  10     extract_attributes,
  11     ExtractorError,
  12     float_or_none,
  13     get_element_by_class,
  14     int_or_none,
  15     srt_subtitles_timecode,
  16     strip_or_none,
  17     mimetype2ext,
  18     try_get,
  19     urlencode_postdata,
  20     urljoin,
  21 )
  22
  23
  24 class LinkedInBaseIE(InfoExtractor):
  25     _NETRC_MACHINE = 'linkedin'
  26     _logged_in = False
  27
  28     def _perform_login(self, username, password):
  29         if self._logged_in:
  30             return
  31
  32         login_page = self._download_webpage(
  33             self._LOGIN_URL, None, 'Downloading login page')
  34         action_url = urljoin(self._LOGIN_URL, self._search_regex(
  35             r'<form[^>]+action=(["\'])(?P<url>.+?)\1', login_page, 'post url',
  36             default='https://www.linkedin.com/uas/login-submit', group='url'))
  37         data = self._hidden_inputs(login_page)
  38         data.update({
  39             'session_key': username,
  40             'session_password': password,
  41         })
  42         login_submit_page = self._download_webpage(
  43             action_url, None, 'Logging in',
  44             data=urlencode_postdata(data))
  45         error = self._search_regex(
  46             r'<span[^>]+class="error"[^>]*>\s*(.+?)\s*</span>',
  47             login_submit_page, 'error', default=None)
  48         if error:
  49             raise ExtractorError(error, expected=True)
  50         LinkedInBaseIE._logged_in = True
  51
  52
  53 class LinkedInLearningBaseIE(LinkedInBaseIE):
  54     _LOGIN_URL = 'https://www.linkedin.com/uas/login?trk=learning'
  55
  56     def _call_api(self, course_slug, fields, video_slug=None, resolution=None):
  57         query = {
  58             'courseSlug': course_slug,
  59             'fields': fields,
  60             'q': 'slugs',
  61         }
  62         sub = ''
  63         if video_slug:
  64             query.update({
  65                 'videoSlug': video_slug,
  66                 'resolution': '_%s' % resolution,
  67             })
  68             sub = ' %dp' % resolution
  69         api_url = 'https://www.linkedin.com/learning-api/detailedCourses'
  70         if not self._get_cookies(api_url).get('JSESSIONID'):
  71             self.raise_login_required()
  72         return self._download_json(
  73             api_url, video_slug, 'Downloading%s JSON metadata' % sub, headers={
  74                 'Csrf-Token': self._get_cookies(api_url)['JSESSIONID'].value,
  75             }, query=query)['elements'][0]
  76
  77     def _get_urn_id(self, video_data):
  78         urn = video_data.get('urn')
  79         if urn:
  80             mobj = re.search(r'urn:li:lyndaCourse:\d+,(\d+)', urn)
  81             if mobj:
  82                 return mobj.group(1)
  83
  84     def _get_video_id(self, video_data, course_slug, video_slug):
  85         return self._get_urn_id(video_data) or '%s/%s' % (course_slug, video_slug)
  86
  87
  88 class LinkedInIE(LinkedInBaseIE):
  89     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/posts/.+?(?P<id>\d+)'
  90     _TESTS = [{
  91         'url': 'https://www.linkedin.com/posts/mishalkhawaja_sendinblueviews-toronto-digitalmarketing-ugcPost-6850898786781339649-mM20',
  92         'info_dict': {
  93             'id': '6850898786781339649',
  94             'ext': 'mp4',
  95             'title': 'Mishal K. on LinkedIn: #sendinblueviews #toronto #digitalmarketing',
  96             'description': 'md5:be125430bab1c574f16aeb186a4d5b19',
  97             'creator': 'Mishal K.'
  98         },
  99     }]
 100
 101     def _real_extract(self, url):
 102         video_id = self._match_id(url)
 103         webpage = self._download_webpage(url, video_id)
 104
 105         title = self._html_search_regex(r'<title>([^<]+)</title>', webpage, 'title')
 106         description = clean_html(get_element_by_class('share-update-card__update-text', webpage))
 107         like_count = int_or_none(get_element_by_class('social-counts-reactions__social-counts-numRections', webpage))
 108         creator = strip_or_none(clean_html(get_element_by_class('comment__actor-name', webpage)))
 109
 110         sources = self._parse_json(extract_attributes(self._search_regex(r'(<video[^>]+>)', webpage, 'video'))['data-sources'], video_id)
 111         formats = [{
 112             'url': source['src'],
 113             'ext': mimetype2ext(source.get('type')),
 114             'tbr': float_or_none(source.get('data-bitrate'), scale=1000),
 115         } for source in sources]
 116
 117         self._sort_formats(formats)
 118
 119         return {
 120             'id': video_id,
 121             'formats': formats,
 122             'title': title,
 123             'like_count': like_count,
 124             'creator': creator,
 125             'thumbnail': self._og_search_thumbnail(webpage),
 126             'description': description,
 127         }
 128
 129
 130 class LinkedInLearningIE(LinkedInLearningBaseIE):
 131     IE_NAME = 'linkedin:learning'
 132     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/learning/(?P<course_slug>[^/]+)/(?P<id>[^/?#]+)'
 133     _TEST = {
 134         'url': 'https://www.linkedin.com/learning/programming-foundations-fundamentals/welcome?autoplay=true',
 135         'md5': 'a1d74422ff0d5e66a792deb996693167',
 136         'info_dict': {
 137             'id': '90426',
 138             'ext': 'mp4',
 139             'title': 'Welcome',
 140             'timestamp': 1430396150.82,
 141             'upload_date': '20150430',
 142         },
 143     }
 144
 145     def json2srt(self, transcript_lines, duration=None):
 146         srt_data = ''
 147         for line, (line_dict, next_dict) in enumerate(zip_longest(transcript_lines, transcript_lines[1:])):
 148             start_time, caption = line_dict['transcriptStartAt'] / 1000, line_dict['caption']
 149             end_time = next_dict['transcriptStartAt'] / 1000 if next_dict else duration or start_time + 1
 150             srt_data += '%d\n%s --> %s\n%s\n\n' % (line + 1, srt_subtitles_timecode(start_time),
 151                                                    srt_subtitles_timecode(end_time),
 152                                                    caption)
 153         return srt_data
 154
 155     def _real_extract(self, url):
 156         course_slug, video_slug = self._match_valid_url(url).groups()
 157
 158         formats = []
 159         for width, height in ((640, 360), (960, 540), (1280, 720)):
 160             video_data = self._call_api(
 161                 course_slug, 'selectedVideo', video_slug, height)['selectedVideo']
 162
 163             video_url_data = video_data.get('url') or {}
 164             progressive_url = video_url_data.get('progressiveUrl')
 165             if progressive_url:
 166                 formats.append({
 167                     'format_id': 'progressive-%dp' % height,
 168                     'url': progressive_url,
 169                     'ext': 'mp4',
 170                     'height': height,
 171                     'width': width,
 172                     'source_preference': 1,
 173                 })
 174
 175         title = video_data['title']
 176
 177         audio_url = video_data.get('audio', {}).get('progressiveUrl')
 178         if audio_url:
 179             formats.append({
 180                 'abr': 64,
 181                 'ext': 'm4a',
 182                 'format_id': 'audio',
 183                 'url': audio_url,
 184                 'vcodec': 'none',
 185             })
 186
 187         streaming_url = video_url_data.get('streamingUrl')
 188         if streaming_url:
 189             formats.extend(self._extract_m3u8_formats(
 190                 streaming_url, video_slug, 'mp4',
 191                 'm3u8_native', m3u8_id='hls', fatal=False))
 192
 193         # It seems like this would be correctly handled by default
 194         # However, unless someone can confirm this, the old
 195         # behaviour is being kept as-is
 196         self._sort_formats(formats, ('res', 'source_preference'))
 197         subtitles = {}
 198         duration = int_or_none(video_data.get('durationInSeconds'))
 199         transcript_lines = try_get(video_data, lambda x: x['transcript']['lines'], expected_type=list)
 200         if transcript_lines:
 201             subtitles['en'] = [{
 202                 'ext': 'srt',
 203                 'data': self.json2srt(transcript_lines, duration)
 204             }]
 205
 206         return {
 207             'id': self._get_video_id(video_data, course_slug, video_slug),
 208             'title': title,
 209             'formats': formats,
 210             'thumbnail': video_data.get('defaultThumbnail'),
 211             'timestamp': float_or_none(video_data.get('publishedOn'), 1000),
 212             'duration': duration,
 213             'subtitles': subtitles,
 214         }
 215
 216
 217 class LinkedInLearningCourseIE(LinkedInLearningBaseIE):
 218     IE_NAME = 'linkedin:learning:course'
 219     _VALID_URL = r'https?://(?:www\.)?linkedin\.com/learning/(?P<id>[^/?#]+)'
 220     _TEST = {
 221         'url': 'https://www.linkedin.com/learning/programming-foundations-fundamentals',
 222         'info_dict': {
 223             'id': 'programming-foundations-fundamentals',
 224             'title': 'Programming Foundations: Fundamentals',
 225             'description': 'md5:76e580b017694eb89dc8e8923fff5c86',
 226         },
 227         'playlist_mincount': 61,
 228     }
 229
 230     @classmethod
 231     def suitable(cls, url):
 232         return False if LinkedInLearningIE.suitable(url) else super(LinkedInLearningCourseIE, cls).suitable(url)
 233
 234     def _real_extract(self, url):
 235         course_slug = self._match_id(url)
 236         course_data = self._call_api(course_slug, 'chapters,description,title')
 237
 238         entries = []
 239         for chapter_number, chapter in enumerate(course_data.get('chapters', []), 1):
 240             chapter_title = chapter.get('title')
 241             chapter_id = self._get_urn_id(chapter)
 242             for video in chapter.get('videos', []):
 243                 video_slug = video.get('slug')
 244                 if not video_slug:
 245                     continue
 246                 entries.append({
 247                     '_type': 'url_transparent',
 248                     'id': self._get_video_id(video, course_slug, video_slug),
 249                     'title': video.get('title'),
 250                     'url': 'https://www.linkedin.com/learning/%s/%s' % (course_slug, video_slug),
 251                     'chapter': chapter_title,
 252                     'chapter_number': chapter_number,
 253                     'chapter_id': chapter_id,
 254                     'ie_key': LinkedInLearningIE.ie_key(),
 255                 })
 256
 257         return self.playlist_result(
 258             entries, course_slug,
 259             course_data.get('title'),
 260             course_data.get('description'))