yt_dlp/extractor/udemy.py

   1 import re
   2
   3 from .common import InfoExtractor
   4 from ..compat import (
   5     compat_HTTPError,
   6     compat_str,
   7     compat_urllib_request,
   8     compat_urlparse,
   9 )
  10 from ..utils import (
  11     determine_ext,
  12     extract_attributes,
  13     ExtractorError,
  14     float_or_none,
  15     int_or_none,
  16     js_to_json,
  17     sanitized_Request,
  18     try_get,
  19     unescapeHTML,
  20     url_or_none,
  21     urlencode_postdata,
  22 )
  23
  24
  25 class UdemyIE(InfoExtractor):
  26     IE_NAME = 'udemy'
  27     _VALID_URL = r'''(?x)
  28                     https?://
  29                         (?:[^/]+\.)?udemy\.com/
  30                         (?:
  31                             [^#]+\#/lecture/|
  32                             lecture/view/?\?lectureId=|
  33                             [^/]+/learn/v4/t/lecture/
  34                         )
  35                         (?P<id>\d+)
  36                     '''
  37     _LOGIN_URL = 'https://www.udemy.com/join/login-popup/?displayType=ajax&showSkipButton=1'
  38     _ORIGIN_URL = 'https://www.udemy.com'
  39     _NETRC_MACHINE = 'udemy'
  40
  41     _TESTS = [{
  42         'url': 'https://www.udemy.com/java-tutorial/#/lecture/172757',
  43         'md5': '98eda5b657e752cf945d8445e261b5c5',
  44         'info_dict': {
  45             'id': '160614',
  46             'ext': 'mp4',
  47             'title': 'Introduction and Installation',
  48             'description': 'md5:c0d51f6f21ef4ec65f091055a5eef876',
  49             'duration': 579.29,
  50         },
  51         'skip': 'Requires udemy account credentials',
  52     }, {
  53         # new URL schema
  54         'url': 'https://www.udemy.com/electric-bass-right-from-the-start/learn/v4/t/lecture/4580906',
  55         'only_matching': True,
  56     }, {
  57         # no url in outputs format entry
  58         'url': 'https://www.udemy.com/learn-web-development-complete-step-by-step-guide-to-success/learn/v4/t/lecture/4125812',
  59         'only_matching': True,
  60     }, {
  61         # only outputs rendition
  62         'url': 'https://www.udemy.com/how-you-can-help-your-local-community-5-amazing-examples/learn/v4/t/lecture/3225750?start=0',
  63         'only_matching': True,
  64     }, {
  65         'url': 'https://wipro.udemy.com/java-tutorial/#/lecture/172757',
  66         'only_matching': True,
  67     }]
  68
  69     def _extract_course_info(self, webpage, video_id):
  70         course = self._parse_json(
  71             unescapeHTML(self._search_regex(
  72                 r'ng-init=["\'].*\bcourse=({.+?})[;"\']',
  73                 webpage, 'course', default='{}')),
  74             video_id, fatal=False) or {}
  75         course_id = course.get('id') or self._search_regex(
  76             [
  77                 r'data-course-id=["\'](\d+)',
  78                 r'&quot;courseId&quot;\s*:\s*(\d+)'
  79             ], webpage, 'course id')
  80         return course_id, course.get('title')
  81
  82     def _enroll_course(self, base_url, webpage, course_id):
  83         def combine_url(base_url, url):
  84             return compat_urlparse.urljoin(base_url, url) if not url.startswith('http') else url
  85
  86         checkout_url = unescapeHTML(self._search_regex(
  87             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/(?:payment|cart)/checkout/.+?)\1',
  88             webpage, 'checkout url', group='url', default=None))
  89         if checkout_url:
  90             raise ExtractorError(
  91                 'Course %s is not free. You have to pay for it before you can download. '
  92                 'Use this URL to confirm purchase: %s'
  93                 % (course_id, combine_url(base_url, checkout_url)),
  94                 expected=True)
  95
  96         enroll_url = unescapeHTML(self._search_regex(
  97             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/course/subscribe/.+?)\1',
  98             webpage, 'enroll url', group='url', default=None))
  99         if enroll_url:
 100             webpage = self._download_webpage(
 101                 combine_url(base_url, enroll_url),
 102                 course_id, 'Enrolling in the course',
 103                 headers={'Referer': base_url})
 104             if '>You have enrolled in' in webpage:
 105                 self.to_screen('%s: Successfully enrolled in the course' % course_id)
 106
 107     def _download_lecture(self, course_id, lecture_id):
 108         return self._download_json(
 109             'https://www.udemy.com/api-2.0/users/me/subscribed-courses/%s/lectures/%s?'
 110             % (course_id, lecture_id),
 111             lecture_id, 'Downloading lecture JSON', query={
 112                 'fields[lecture]': 'title,description,view_html,asset',
 113                 'fields[asset]': 'asset_type,stream_url,thumbnail_url,download_urls,stream_urls,captions,data',
 114             })
 115
 116     def _handle_error(self, response):
 117         if not isinstance(response, dict):
 118             return
 119         error = response.get('error')
 120         if error:
 121             error_str = 'Udemy returned error #%s: %s' % (error.get('code'), error.get('message'))
 122             error_data = error.get('data')
 123             if error_data:
 124                 error_str += ' - %s' % error_data.get('formErrors')
 125             raise ExtractorError(error_str, expected=True)
 126
 127     def _download_webpage_handle(self, *args, **kwargs):
 128         headers = kwargs.get('headers', {}).copy()
 129         headers['User-Agent'] = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/72.0.3626.109 Safari/537.36'
 130         kwargs['headers'] = headers
 131         ret = super(UdemyIE, self)._download_webpage_handle(
 132             *args, **kwargs)
 133         if not ret:
 134             return ret
 135         webpage, _ = ret
 136         if any(p in webpage for p in (
 137                 '>Please verify you are a human',
 138                 'Access to this page has been denied because we believe you are using automation tools to browse the website',
 139                 '"_pxCaptcha"')):
 140             raise ExtractorError(
 141                 'Udemy asks you to solve a CAPTCHA. Login with browser, '
 142                 'solve CAPTCHA, then export cookies and pass cookie file to '
 143                 'yt-dlp with --cookies.', expected=True)
 144         return ret
 145
 146     def _download_json(self, url_or_request, *args, **kwargs):
 147         headers = {
 148             'X-Udemy-Snail-Case': 'true',
 149             'X-Requested-With': 'XMLHttpRequest',
 150         }
 151         for cookie in self.cookiejar:
 152             if cookie.name == 'client_id':
 153                 headers['X-Udemy-Client-Id'] = cookie.value
 154             elif cookie.name == 'access_token':
 155                 headers['X-Udemy-Bearer-Token'] = cookie.value
 156                 headers['X-Udemy-Authorization'] = 'Bearer %s' % cookie.value
 157
 158         if isinstance(url_or_request, compat_urllib_request.Request):
 159             for header, value in headers.items():
 160                 url_or_request.add_header(header, value)
 161         else:
 162             url_or_request = sanitized_Request(url_or_request, headers=headers)
 163
 164         response = super(UdemyIE, self)._download_json(url_or_request, *args, **kwargs)
 165         self._handle_error(response)
 166         return response
 167
 168     def _perform_login(self, username, password):
 169         login_popup = self._download_webpage(
 170             self._LOGIN_URL, None, 'Downloading login popup')
 171
 172         def is_logged(webpage):
 173             return any(re.search(p, webpage) for p in (
 174                 r'href=["\'](?:https://www\.udemy\.com)?/user/logout/',
 175                 r'>Logout<'))
 176
 177         # already logged in
 178         if is_logged(login_popup):
 179             return
 180
 181         login_form = self._form_hidden_inputs('login-form', login_popup)
 182
 183         login_form.update({
 184             'email': username,
 185             'password': password,
 186         })
 187
 188         response = self._download_webpage(
 189             self._LOGIN_URL, None, 'Logging in',
 190             data=urlencode_postdata(login_form),
 191             headers={
 192                 'Referer': self._ORIGIN_URL,
 193                 'Origin': self._ORIGIN_URL,
 194             })
 195
 196         if not is_logged(response):
 197             error = self._html_search_regex(
 198                 r'(?s)<div[^>]+class="form-errors[^"]*">(.+?)</div>',
 199                 response, 'error message', default=None)
 200             if error:
 201                 raise ExtractorError('Unable to login: %s' % error, expected=True)
 202             raise ExtractorError('Unable to log in')
 203
 204     def _real_extract(self, url):
 205         lecture_id = self._match_id(url)
 206
 207         webpage = self._download_webpage(url, lecture_id)
 208
 209         course_id, _ = self._extract_course_info(webpage, lecture_id)
 210
 211         try:
 212             lecture = self._download_lecture(course_id, lecture_id)
 213         except ExtractorError as e:
 214             # Error could possibly mean we are not enrolled in the course
 215             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 216                 self._enroll_course(url, webpage, course_id)
 217                 lecture = self._download_lecture(course_id, lecture_id)
 218             else:
 219                 raise
 220
 221         title = lecture['title']
 222         description = lecture.get('description')
 223
 224         asset = lecture['asset']
 225
 226         asset_type = asset.get('asset_type') or asset.get('assetType')
 227         if asset_type != 'Video':
 228             raise ExtractorError(
 229                 'Lecture %s is not a video' % lecture_id, expected=True)
 230
 231         stream_url = asset.get('stream_url') or asset.get('streamUrl')
 232         if stream_url:
 233             youtube_url = self._search_regex(
 234                 r'(https?://www\.youtube\.com/watch\?v=.*)', stream_url, 'youtube URL', default=None)
 235             if youtube_url:
 236                 return self.url_result(youtube_url, 'Youtube')
 237
 238         video_id = compat_str(asset['id'])
 239         thumbnail = asset.get('thumbnail_url') or asset.get('thumbnailUrl')
 240         duration = float_or_none(asset.get('data', {}).get('duration'))
 241
 242         subtitles = {}
 243         automatic_captions = {}
 244
 245         formats = []
 246
 247         def extract_output_format(src, f_id):
 248             return {
 249                 'url': src.get('url'),
 250                 'format_id': '%sp' % (src.get('height') or f_id),
 251                 'width': int_or_none(src.get('width')),
 252                 'height': int_or_none(src.get('height')),
 253                 'vbr': int_or_none(src.get('video_bitrate_in_kbps')),
 254                 'vcodec': src.get('video_codec'),
 255                 'fps': int_or_none(src.get('frame_rate')),
 256                 'abr': int_or_none(src.get('audio_bitrate_in_kbps')),
 257                 'acodec': src.get('audio_codec'),
 258                 'asr': int_or_none(src.get('audio_sample_rate')),
 259                 'tbr': int_or_none(src.get('total_bitrate_in_kbps')),
 260                 'filesize': int_or_none(src.get('file_size_in_bytes')),
 261             }
 262
 263         outputs = asset.get('data', {}).get('outputs')
 264         if not isinstance(outputs, dict):
 265             outputs = {}
 266
 267         def add_output_format_meta(f, key):
 268             output = outputs.get(key)
 269             if isinstance(output, dict):
 270                 output_format = extract_output_format(output, key)
 271                 output_format.update(f)
 272                 return output_format
 273             return f
 274
 275         def extract_formats(source_list):
 276             if not isinstance(source_list, list):
 277                 return
 278             for source in source_list:
 279                 video_url = url_or_none(source.get('file') or source.get('src'))
 280                 if not video_url:
 281                     continue
 282                 if source.get('type') == 'application/x-mpegURL' or determine_ext(video_url) == 'm3u8':
 283                     formats.extend(self._extract_m3u8_formats(
 284                         video_url, video_id, 'mp4', entry_protocol='m3u8_native',
 285                         m3u8_id='hls', fatal=False))
 286                     continue
 287                 format_id = source.get('label')
 288                 f = {
 289                     'url': video_url,
 290                     'format_id': '%sp' % format_id,
 291                     'height': int_or_none(format_id),
 292                 }
 293                 if format_id:
 294                     # Some videos contain additional metadata (e.g.
 295                     # https://www.udemy.com/ios9-swift/learn/#/lecture/3383208)
 296                     f = add_output_format_meta(f, format_id)
 297                 formats.append(f)
 298
 299         def extract_subtitles(track_list):
 300             if not isinstance(track_list, list):
 301                 return
 302             for track in track_list:
 303                 if not isinstance(track, dict):
 304                     continue
 305                 if track.get('kind') != 'captions':
 306                     continue
 307                 src = url_or_none(track.get('src'))
 308                 if not src:
 309                     continue
 310                 lang = track.get('language') or track.get(
 311                     'srclang') or track.get('label')
 312                 sub_dict = automatic_captions if track.get(
 313                     'autogenerated') is True else subtitles
 314                 sub_dict.setdefault(lang, []).append({
 315                     'url': src,
 316                 })
 317
 318         for url_kind in ('download', 'stream'):
 319             urls = asset.get('%s_urls' % url_kind)
 320             if isinstance(urls, dict):
 321                 extract_formats(urls.get('Video'))
 322
 323         captions = asset.get('captions')
 324         if isinstance(captions, list):
 325             for cc in captions:
 326                 if not isinstance(cc, dict):
 327                     continue
 328                 cc_url = url_or_none(cc.get('url'))
 329                 if not cc_url:
 330                     continue
 331                 lang = try_get(cc, lambda x: x['locale']['locale'], compat_str)
 332                 sub_dict = (automatic_captions if cc.get('source') == 'auto'
 333                             else subtitles)
 334                 sub_dict.setdefault(lang or 'en', []).append({
 335                     'url': cc_url,
 336                 })
 337
 338         view_html = lecture.get('view_html')
 339         if view_html:
 340             view_html_urls = set()
 341             for source in re.findall(r'<source[^>]+>', view_html):
 342                 attributes = extract_attributes(source)
 343                 src = attributes.get('src')
 344                 if not src:
 345                     continue
 346                 res = attributes.get('data-res')
 347                 height = int_or_none(res)
 348                 if src in view_html_urls:
 349                     continue
 350                 view_html_urls.add(src)
 351                 if attributes.get('type') == 'application/x-mpegURL' or determine_ext(src) == 'm3u8':
 352                     m3u8_formats = self._extract_m3u8_formats(
 353                         src, video_id, 'mp4', entry_protocol='m3u8_native',
 354                         m3u8_id='hls', fatal=False)
 355                     for f in m3u8_formats:
 356                         m = re.search(r'/hls_(?P<height>\d{3,4})_(?P<tbr>\d{2,})/', f['url'])
 357                         if m:
 358                             if not f.get('height'):
 359                                 f['height'] = int(m.group('height'))
 360                             if not f.get('tbr'):
 361                                 f['tbr'] = int(m.group('tbr'))
 362                     formats.extend(m3u8_formats)
 363                 else:
 364                     formats.append(add_output_format_meta({
 365                         'url': src,
 366                         'format_id': '%dp' % height if height else None,
 367                         'height': height,
 368                     }, res))
 369
 370             # react rendition since 2017.04.15 (see
 371             # https://github.com/ytdl-org/youtube-dl/issues/12744)
 372             data = self._parse_json(
 373                 self._search_regex(
 374                     r'videojs-setup-data=(["\'])(?P<data>{.+?})\1', view_html,
 375                     'setup data', default='{}', group='data'), video_id,
 376                 transform_source=unescapeHTML, fatal=False)
 377             if data and isinstance(data, dict):
 378                 extract_formats(data.get('sources'))
 379                 if not duration:
 380                     duration = int_or_none(data.get('duration'))
 381                 extract_subtitles(data.get('tracks'))
 382
 383             if not subtitles and not automatic_captions:
 384                 text_tracks = self._parse_json(
 385                     self._search_regex(
 386                         r'text-tracks=(["\'])(?P<data>\[.+?\])\1', view_html,
 387                         'text tracks', default='{}', group='data'), video_id,
 388                     transform_source=lambda s: js_to_json(unescapeHTML(s)),
 389                     fatal=False)
 390                 extract_subtitles(text_tracks)
 391
 392         if not formats and outputs:
 393             for format_id, output in outputs.items():
 394                 f = extract_output_format(output, format_id)
 395                 if f.get('url'):
 396                     formats.append(f)
 397
 398         self._sort_formats(formats)
 399
 400         return {
 401             'id': video_id,
 402             'title': title,
 403             'description': description,
 404             'thumbnail': thumbnail,
 405             'duration': duration,
 406             'formats': formats,
 407             'subtitles': subtitles,
 408             'automatic_captions': automatic_captions,
 409         }
 410
 411
 412 class UdemyCourseIE(UdemyIE):
 413     IE_NAME = 'udemy:course'
 414     _VALID_URL = r'https?://(?:[^/]+\.)?udemy\.com/(?P<id>[^/?#&]+)'
 415     _TESTS = [{
 416         'url': 'https://www.udemy.com/java-tutorial/',
 417         'only_matching': True,
 418     }, {
 419         'url': 'https://wipro.udemy.com/java-tutorial/',
 420         'only_matching': True,
 421     }]
 422
 423     @classmethod
 424     def suitable(cls, url):
 425         return False if UdemyIE.suitable(url) else super(UdemyCourseIE, cls).suitable(url)
 426
 427     def _real_extract(self, url):
 428         course_path = self._match_id(url)
 429
 430         webpage = self._download_webpage(url, course_path)
 431
 432         course_id, title = self._extract_course_info(webpage, course_path)
 433
 434         self._enroll_course(url, webpage, course_id)
 435
 436         response = self._download_json(
 437             'https://www.udemy.com/api-2.0/courses/%s/cached-subscriber-curriculum-items' % course_id,
 438             course_id, 'Downloading course curriculum', query={
 439                 'fields[chapter]': 'title,object_index',
 440                 'fields[lecture]': 'title,asset',
 441                 'page_size': '1000',
 442             })
 443
 444         entries = []
 445         chapter, chapter_number = [None] * 2
 446         for entry in response['results']:
 447             clazz = entry.get('_class')
 448             if clazz == 'lecture':
 449                 asset = entry.get('asset')
 450                 if isinstance(asset, dict):
 451                     asset_type = asset.get('asset_type') or asset.get('assetType')
 452                     if asset_type != 'Video':
 453                         continue
 454                 lecture_id = entry.get('id')
 455                 if lecture_id:
 456                     entry = {
 457                         '_type': 'url_transparent',
 458                         'url': 'https://www.udemy.com/%s/learn/v4/t/lecture/%s' % (course_path, entry['id']),
 459                         'title': entry.get('title'),
 460                         'ie_key': UdemyIE.ie_key(),
 461                     }
 462                     if chapter_number:
 463                         entry['chapter_number'] = chapter_number
 464                     if chapter:
 465                         entry['chapter'] = chapter
 466                     entries.append(entry)
 467             elif clazz == 'chapter':
 468                 chapter_number = entry.get('object_index')
 469                 chapter = entry.get('title')
 470
 471         return self.playlist_result(entries, course_id, title)