yt_dlp/extractor/udemy.py

   1 import re
   2 import urllib.request
   3
   4 from .common import InfoExtractor
   5 from ..compat import compat_HTTPError, compat_str, compat_urlparse
   6 from ..utils import (
   7     ExtractorError,
   8     determine_ext,
   9     extract_attributes,
  10     float_or_none,
  11     int_or_none,
  12     js_to_json,
  13     sanitized_Request,
  14     smuggle_url,
  15     try_get,
  16     unescapeHTML,
  17     unsmuggle_url,
  18     url_or_none,
  19     urlencode_postdata,
  20 )
  21
  22
  23 class UdemyIE(InfoExtractor):
  24     IE_NAME = 'udemy'
  25     _VALID_URL = r'''(?x)
  26                     https?://
  27                         (?:[^/]+\.)?udemy\.com/
  28                         (?:
  29                             [^#]+\#/lecture/|
  30                             lecture/view/?\?lectureId=|
  31                             [^/]+/learn/v4/t/lecture/
  32                         )
  33                         (?P<id>\d+)
  34                     '''
  35     _LOGIN_URL = 'https://www.udemy.com/join/login-popup/?displayType=ajax&showSkipButton=1'
  36     _ORIGIN_URL = 'https://www.udemy.com'
  37     _NETRC_MACHINE = 'udemy'
  38
  39     _TESTS = [{
  40         'url': 'https://www.udemy.com/java-tutorial/#/lecture/172757',
  41         'md5': '98eda5b657e752cf945d8445e261b5c5',
  42         'info_dict': {
  43             'id': '160614',
  44             'ext': 'mp4',
  45             'title': 'Introduction and Installation',
  46             'description': 'md5:c0d51f6f21ef4ec65f091055a5eef876',
  47             'duration': 579.29,
  48         },
  49         'skip': 'Requires udemy account credentials',
  50     }, {
  51         # new URL schema
  52         'url': 'https://www.udemy.com/electric-bass-right-from-the-start/learn/v4/t/lecture/4580906',
  53         'only_matching': True,
  54     }, {
  55         # no url in outputs format entry
  56         'url': 'https://www.udemy.com/learn-web-development-complete-step-by-step-guide-to-success/learn/v4/t/lecture/4125812',
  57         'only_matching': True,
  58     }, {
  59         # only outputs rendition
  60         'url': 'https://www.udemy.com/how-you-can-help-your-local-community-5-amazing-examples/learn/v4/t/lecture/3225750?start=0',
  61         'only_matching': True,
  62     }, {
  63         'url': 'https://wipro.udemy.com/java-tutorial/#/lecture/172757',
  64         'only_matching': True,
  65     }]
  66
  67     def _extract_course_info(self, webpage, video_id):
  68         course = self._parse_json(
  69             unescapeHTML(self._search_regex(
  70                 r'ng-init=["\'].*\bcourse=({.+?})[;"\']',
  71                 webpage, 'course', default='{}')),
  72             video_id, fatal=False) or {}
  73         course_id = course.get('id') or self._search_regex(
  74             [
  75                 r'data-course-id=["\'](\d+)',
  76                 r'&quot;courseId&quot;\s*:\s*(\d+)'
  77             ], webpage, 'course id')
  78         return course_id, course.get('title')
  79
  80     def _enroll_course(self, base_url, webpage, course_id):
  81         def combine_url(base_url, url):
  82             return compat_urlparse.urljoin(base_url, url) if not url.startswith('http') else url
  83
  84         checkout_url = unescapeHTML(self._search_regex(
  85             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/(?:payment|cart)/checkout/.+?)\1',
  86             webpage, 'checkout url', group='url', default=None))
  87         if checkout_url:
  88             raise ExtractorError(
  89                 'Course %s is not free. You have to pay for it before you can download. '
  90                 'Use this URL to confirm purchase: %s'
  91                 % (course_id, combine_url(base_url, checkout_url)),
  92                 expected=True)
  93
  94         enroll_url = unescapeHTML(self._search_regex(
  95             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/course/subscribe/.+?)\1',
  96             webpage, 'enroll url', group='url', default=None))
  97         if enroll_url:
  98             webpage = self._download_webpage(
  99                 combine_url(base_url, enroll_url),
 100                 course_id, 'Enrolling in the course',
 101                 headers={'Referer': base_url})
 102             if '>You have enrolled in' in webpage:
 103                 self.to_screen('%s: Successfully enrolled in the course' % course_id)
 104
 105     def _download_lecture(self, course_id, lecture_id):
 106         return self._download_json(
 107             'https://www.udemy.com/api-2.0/users/me/subscribed-courses/%s/lectures/%s?'
 108             % (course_id, lecture_id),
 109             lecture_id, 'Downloading lecture JSON', query={
 110                 'fields[lecture]': 'title,description,view_html,asset',
 111                 'fields[asset]': 'asset_type,stream_url,thumbnail_url,download_urls,stream_urls,captions,data,course_is_drmed',
 112             })
 113
 114     def _handle_error(self, response):
 115         if not isinstance(response, dict):
 116             return
 117         error = response.get('error')
 118         if error:
 119             error_str = 'Udemy returned error #%s: %s' % (error.get('code'), error.get('message'))
 120             error_data = error.get('data')
 121             if error_data:
 122                 error_str += ' - %s' % error_data.get('formErrors')
 123             raise ExtractorError(error_str, expected=True)
 124
 125     def _download_webpage_handle(self, *args, **kwargs):
 126         headers = kwargs.get('headers', {}).copy()
 127         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'
 128         kwargs['headers'] = headers
 129         ret = super(UdemyIE, self)._download_webpage_handle(
 130             *args, **kwargs)
 131         if not ret:
 132             return ret
 133         webpage, _ = ret
 134         if any(p in webpage for p in (
 135                 '>Please verify you are a human',
 136                 'Access to this page has been denied because we believe you are using automation tools to browse the website',
 137                 '"_pxCaptcha"')):
 138             raise ExtractorError(
 139                 'Udemy asks you to solve a CAPTCHA. Login with browser, '
 140                 'solve CAPTCHA, then export cookies and pass cookie file to '
 141                 'yt-dlp with --cookies.', expected=True)
 142         return ret
 143
 144     def _download_json(self, url_or_request, *args, **kwargs):
 145         headers = {
 146             'X-Udemy-Snail-Case': 'true',
 147             'X-Requested-With': 'XMLHttpRequest',
 148         }
 149         for cookie in self.cookiejar:
 150             if cookie.name == 'client_id':
 151                 headers['X-Udemy-Client-Id'] = cookie.value
 152             elif cookie.name == 'access_token':
 153                 headers['X-Udemy-Bearer-Token'] = cookie.value
 154                 headers['X-Udemy-Authorization'] = 'Bearer %s' % cookie.value
 155
 156         if isinstance(url_or_request, urllib.request.Request):
 157             for header, value in headers.items():
 158                 url_or_request.add_header(header, value)
 159         else:
 160             url_or_request = sanitized_Request(url_or_request, headers=headers)
 161
 162         response = super(UdemyIE, self)._download_json(url_or_request, *args, **kwargs)
 163         self._handle_error(response)
 164         return response
 165
 166     def _perform_login(self, username, password):
 167         login_popup = self._download_webpage(
 168             self._LOGIN_URL, None, 'Downloading login popup')
 169
 170         def is_logged(webpage):
 171             return any(re.search(p, webpage) for p in (
 172                 r'href=["\'](?:https://www\.udemy\.com)?/user/logout/',
 173                 r'>Logout<'))
 174
 175         # already logged in
 176         if is_logged(login_popup):
 177             return
 178
 179         login_form = self._form_hidden_inputs('login-form', login_popup)
 180
 181         login_form.update({
 182             'email': username,
 183             'password': password,
 184         })
 185
 186         response = self._download_webpage(
 187             self._LOGIN_URL, None, 'Logging in',
 188             data=urlencode_postdata(login_form),
 189             headers={
 190                 'Referer': self._ORIGIN_URL,
 191                 'Origin': self._ORIGIN_URL,
 192             })
 193
 194         if not is_logged(response):
 195             error = self._html_search_regex(
 196                 r'(?s)<div[^>]+class="form-errors[^"]*">(.+?)</div>',
 197                 response, 'error message', default=None)
 198             if error:
 199                 raise ExtractorError('Unable to login: %s' % error, expected=True)
 200             raise ExtractorError('Unable to log in')
 201
 202     def _real_extract(self, url):
 203         lecture_id = self._match_id(url)
 204         course_id = unsmuggle_url(url, {})[1].get('course_id')
 205
 206         webpage = None
 207         if not course_id:
 208             webpage = self._download_webpage(url, lecture_id)
 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                 webpage = webpage or self._download_webpage(url, lecture_id)
 217                 self._enroll_course(url, webpage, course_id)
 218                 lecture = self._download_lecture(course_id, lecture_id)
 219             else:
 220                 raise
 221
 222         title = lecture['title']
 223         description = lecture.get('description')
 224
 225         asset = lecture['asset']
 226
 227         asset_type = asset.get('asset_type') or asset.get('assetType')
 228         if asset_type != 'Video':
 229             raise ExtractorError(
 230                 'Lecture %s is not a video' % lecture_id, expected=True)
 231
 232         stream_url = asset.get('stream_url') or asset.get('streamUrl')
 233         if stream_url:
 234             youtube_url = self._search_regex(
 235                 r'(https?://www\.youtube\.com/watch\?v=.*)', stream_url, 'youtube URL', default=None)
 236             if youtube_url:
 237                 return self.url_result(youtube_url, 'Youtube')
 238
 239         video_id = compat_str(asset['id'])
 240         thumbnail = asset.get('thumbnail_url') or asset.get('thumbnailUrl')
 241         duration = float_or_none(asset.get('data', {}).get('duration'))
 242
 243         subtitles = {}
 244         automatic_captions = {}
 245
 246         formats = []
 247
 248         def extract_output_format(src, f_id):
 249             return {
 250                 'url': src.get('url'),
 251                 'format_id': '%sp' % (src.get('height') or f_id),
 252                 'width': int_or_none(src.get('width')),
 253                 'height': int_or_none(src.get('height')),
 254                 'vbr': int_or_none(src.get('video_bitrate_in_kbps')),
 255                 'vcodec': src.get('video_codec'),
 256                 'fps': int_or_none(src.get('frame_rate')),
 257                 'abr': int_or_none(src.get('audio_bitrate_in_kbps')),
 258                 'acodec': src.get('audio_codec'),
 259                 'asr': int_or_none(src.get('audio_sample_rate')),
 260                 'tbr': int_or_none(src.get('total_bitrate_in_kbps')),
 261                 'filesize': int_or_none(src.get('file_size_in_bytes')),
 262             }
 263
 264         outputs = asset.get('data', {}).get('outputs')
 265         if not isinstance(outputs, dict):
 266             outputs = {}
 267
 268         def add_output_format_meta(f, key):
 269             output = outputs.get(key)
 270             if isinstance(output, dict):
 271                 output_format = extract_output_format(output, key)
 272                 output_format.update(f)
 273                 return output_format
 274             return f
 275
 276         def extract_formats(source_list):
 277             if not isinstance(source_list, list):
 278                 return
 279             for source in source_list:
 280                 video_url = url_or_none(source.get('file') or source.get('src'))
 281                 if not video_url:
 282                     continue
 283                 if source.get('type') == 'application/x-mpegURL' or determine_ext(video_url) == 'm3u8':
 284                     formats.extend(self._extract_m3u8_formats(
 285                         video_url, video_id, 'mp4', entry_protocol='m3u8_native',
 286                         m3u8_id='hls', fatal=False))
 287                     continue
 288                 format_id = source.get('label')
 289                 f = {
 290                     'url': video_url,
 291                     'format_id': '%sp' % format_id,
 292                     'height': int_or_none(format_id),
 293                 }
 294                 if format_id:
 295                     # Some videos contain additional metadata (e.g.
 296                     # https://www.udemy.com/ios9-swift/learn/#/lecture/3383208)
 297                     f = add_output_format_meta(f, format_id)
 298                 formats.append(f)
 299
 300         def extract_subtitles(track_list):
 301             if not isinstance(track_list, list):
 302                 return
 303             for track in track_list:
 304                 if not isinstance(track, dict):
 305                     continue
 306                 if track.get('kind') != 'captions':
 307                     continue
 308                 src = url_or_none(track.get('src'))
 309                 if not src:
 310                     continue
 311                 lang = track.get('language') or track.get(
 312                     'srclang') or track.get('label')
 313                 sub_dict = automatic_captions if track.get(
 314                     'autogenerated') is True else subtitles
 315                 sub_dict.setdefault(lang, []).append({
 316                     'url': src,
 317                 })
 318
 319         for url_kind in ('download', 'stream'):
 320             urls = asset.get('%s_urls' % url_kind)
 321             if isinstance(urls, dict):
 322                 extract_formats(urls.get('Video'))
 323
 324         captions = asset.get('captions')
 325         if isinstance(captions, list):
 326             for cc in captions:
 327                 if not isinstance(cc, dict):
 328                     continue
 329                 cc_url = url_or_none(cc.get('url'))
 330                 if not cc_url:
 331                     continue
 332                 lang = try_get(cc, lambda x: x['locale']['locale'], compat_str)
 333                 sub_dict = (automatic_captions if cc.get('source') == 'auto'
 334                             else subtitles)
 335                 sub_dict.setdefault(lang or 'en', []).append({
 336                     'url': cc_url,
 337                 })
 338
 339         view_html = lecture.get('view_html')
 340         if view_html:
 341             view_html_urls = set()
 342             for source in re.findall(r'<source[^>]+>', view_html):
 343                 attributes = extract_attributes(source)
 344                 src = attributes.get('src')
 345                 if not src:
 346                     continue
 347                 res = attributes.get('data-res')
 348                 height = int_or_none(res)
 349                 if src in view_html_urls:
 350                     continue
 351                 view_html_urls.add(src)
 352                 if attributes.get('type') == 'application/x-mpegURL' or determine_ext(src) == 'm3u8':
 353                     m3u8_formats = self._extract_m3u8_formats(
 354                         src, video_id, 'mp4', entry_protocol='m3u8_native',
 355                         m3u8_id='hls', fatal=False)
 356                     for f in m3u8_formats:
 357                         m = re.search(r'/hls_(?P<height>\d{3,4})_(?P<tbr>\d{2,})/', f['url'])
 358                         if m:
 359                             if not f.get('height'):
 360                                 f['height'] = int(m.group('height'))
 361                             if not f.get('tbr'):
 362                                 f['tbr'] = int(m.group('tbr'))
 363                     formats.extend(m3u8_formats)
 364                 else:
 365                     formats.append(add_output_format_meta({
 366                         'url': src,
 367                         'format_id': '%dp' % height if height else None,
 368                         'height': height,
 369                     }, res))
 370
 371             # react rendition since 2017.04.15 (see
 372             # https://github.com/ytdl-org/youtube-dl/issues/12744)
 373             data = self._parse_json(
 374                 self._search_regex(
 375                     r'videojs-setup-data=(["\'])(?P<data>{.+?})\1', view_html,
 376                     'setup data', default='{}', group='data'), video_id,
 377                 transform_source=unescapeHTML, fatal=False)
 378             if data and isinstance(data, dict):
 379                 extract_formats(data.get('sources'))
 380                 if not duration:
 381                     duration = int_or_none(data.get('duration'))
 382                 extract_subtitles(data.get('tracks'))
 383
 384             if not subtitles and not automatic_captions:
 385                 text_tracks = self._parse_json(
 386                     self._search_regex(
 387                         r'text-tracks=(["\'])(?P<data>\[.+?\])\1', view_html,
 388                         'text tracks', default='{}', group='data'), video_id,
 389                     transform_source=lambda s: js_to_json(unescapeHTML(s)),
 390                     fatal=False)
 391                 extract_subtitles(text_tracks)
 392
 393         if not formats and outputs:
 394             for format_id, output in outputs.items():
 395                 f = extract_output_format(output, format_id)
 396                 if f.get('url'):
 397                     formats.append(f)
 398
 399         if not formats and asset.get('course_is_drmed'):
 400             self.report_drm(video_id)
 401
 402         return {
 403             'id': video_id,
 404             'title': title,
 405             'description': description,
 406             'thumbnail': thumbnail,
 407             'duration': duration,
 408             'formats': formats,
 409             'subtitles': subtitles,
 410             'automatic_captions': automatic_captions,
 411         }
 412
 413
 414 class UdemyCourseIE(UdemyIE):  # XXX: Do not subclass from concrete IE
 415     IE_NAME = 'udemy:course'
 416     _VALID_URL = r'https?://(?:[^/]+\.)?udemy\.com/(?P<id>[^/?#&]+)'
 417     _TESTS = [{
 418         'url': 'https://www.udemy.com/java-tutorial/',
 419         'only_matching': True,
 420     }, {
 421         'url': 'https://wipro.udemy.com/java-tutorial/',
 422         'only_matching': True,
 423     }]
 424
 425     @classmethod
 426     def suitable(cls, url):
 427         return False if UdemyIE.suitable(url) else super(UdemyCourseIE, cls).suitable(url)
 428
 429     def _real_extract(self, url):
 430         course_path = self._match_id(url)
 431
 432         webpage = self._download_webpage(url, course_path)
 433
 434         course_id, title = self._extract_course_info(webpage, course_path)
 435
 436         self._enroll_course(url, webpage, course_id)
 437
 438         response = self._download_json(
 439             'https://www.udemy.com/api-2.0/courses/%s/cached-subscriber-curriculum-items' % course_id,
 440             course_id, 'Downloading course curriculum', query={
 441                 'fields[chapter]': 'title,object_index',
 442                 'fields[lecture]': 'title,asset',
 443                 'page_size': '1000',
 444             })
 445
 446         entries = []
 447         chapter, chapter_number = [None] * 2
 448         for entry in response['results']:
 449             clazz = entry.get('_class')
 450             if clazz == 'lecture':
 451                 asset = entry.get('asset')
 452                 if isinstance(asset, dict):
 453                     asset_type = asset.get('asset_type') or asset.get('assetType')
 454                     if asset_type != 'Video':
 455                         continue
 456                 lecture_id = entry.get('id')
 457                 if lecture_id:
 458                     entry = {
 459                         '_type': 'url_transparent',
 460                         'url': smuggle_url(
 461                             f'https://www.udemy.com/{course_path}/learn/v4/t/lecture/{entry["id"]}',
 462                             {'course_id': course_id}),
 463                         'title': entry.get('title'),
 464                         'ie_key': UdemyIE.ie_key(),
 465                     }
 466                     if chapter_number:
 467                         entry['chapter_number'] = chapter_number
 468                     if chapter:
 469                         entry['chapter'] = chapter
 470                     entries.append(entry)
 471             elif clazz == 'chapter':
 472                 chapter_number = entry.get('object_index')
 473                 chapter = entry.get('title')
 474
 475         return self.playlist_result(entries, course_id, title)