yt_dlp/extractor/udemy.py

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