youtube_dl/extractor/udemy.py

   1 from __future__ import unicode_literals
   2
   3 import re
   4
   5 from .common import InfoExtractor
   6 from ..compat import (
   7     compat_HTTPError,
   8     compat_kwargs,
   9     compat_str,
  10     compat_urllib_request,
  11     compat_urlparse,
  12 )
  13 from ..utils import (
  14     determine_ext,
  15     extract_attributes,
  16     ExtractorError,
  17     float_or_none,
  18     int_or_none,
  19     js_to_json,
  20     sanitized_Request,
  21     try_get,
  22     unescapeHTML,
  23     url_or_none,
  24     urlencode_postdata,
  25 )
  26
  27
  28 class UdemyIE(InfoExtractor):
  29     IE_NAME = 'udemy'
  30     _VALID_URL = r'''(?x)
  31                     https?://
  32                         (?:[^/]+\.)?udemy\.com/
  33                         (?:
  34                             [^#]+\#/lecture/|
  35                             lecture/view/?\?lectureId=|
  36                             [^/]+/learn/v4/t/lecture/
  37                         )
  38                         (?P<id>\d+)
  39                     '''
  40     _LOGIN_URL = 'https://www.udemy.com/join/login-popup/?displayType=ajax&showSkipButton=1'
  41     _ORIGIN_URL = 'https://www.udemy.com'
  42     _NETRC_MACHINE = 'udemy'
  43
  44     _TESTS = [{
  45         'url': 'https://www.udemy.com/java-tutorial/#/lecture/172757',
  46         'md5': '98eda5b657e752cf945d8445e261b5c5',
  47         'info_dict': {
  48             'id': '160614',
  49             'ext': 'mp4',
  50             'title': 'Introduction and Installation',
  51             'description': 'md5:c0d51f6f21ef4ec65f091055a5eef876',
  52             'duration': 579.29,
  53         },
  54         'skip': 'Requires udemy account credentials',
  55     }, {
  56         # new URL schema
  57         'url': 'https://www.udemy.com/electric-bass-right-from-the-start/learn/v4/t/lecture/4580906',
  58         'only_matching': True,
  59     }, {
  60         # no url in outputs format entry
  61         'url': 'https://www.udemy.com/learn-web-development-complete-step-by-step-guide-to-success/learn/v4/t/lecture/4125812',
  62         'only_matching': True,
  63     }, {
  64         # only outputs rendition
  65         'url': 'https://www.udemy.com/how-you-can-help-your-local-community-5-amazing-examples/learn/v4/t/lecture/3225750?start=0',
  66         'only_matching': True,
  67     }, {
  68         'url': 'https://wipro.udemy.com/java-tutorial/#/lecture/172757',
  69         'only_matching': True,
  70     }]
  71
  72     def _extract_course_info(self, webpage, video_id):
  73         course = self._parse_json(
  74             unescapeHTML(self._search_regex(
  75                 r'ng-init=["\'].*\bcourse=({.+?})[;"\']',
  76                 webpage, 'course', default='{}')),
  77             video_id, fatal=False) or {}
  78         course_id = course.get('id') or self._search_regex(
  79             [
  80                 r'data-course-id=["\'](\d+)',
  81                 r'&quot;courseId&quot;\s*:\s*(\d+)'
  82             ], webpage, 'course id')
  83         return course_id, course.get('title')
  84
  85     def _enroll_course(self, base_url, webpage, course_id):
  86         def combine_url(base_url, url):
  87             return compat_urlparse.urljoin(base_url, url) if not url.startswith('http') else url
  88
  89         checkout_url = unescapeHTML(self._search_regex(
  90             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/(?:payment|cart)/checkout/.+?)\1',
  91             webpage, 'checkout url', group='url', default=None))
  92         if checkout_url:
  93             raise ExtractorError(
  94                 'Course %s is not free. You have to pay for it before you can download. '
  95                 'Use this URL to confirm purchase: %s'
  96                 % (course_id, combine_url(base_url, checkout_url)),
  97                 expected=True)
  98
  99         enroll_url = unescapeHTML(self._search_regex(
 100             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/course/subscribe/.+?)\1',
 101             webpage, 'enroll url', group='url', default=None))
 102         if enroll_url:
 103             webpage = self._download_webpage(
 104                 combine_url(base_url, enroll_url),
 105                 course_id, 'Enrolling in the course',
 106                 headers={'Referer': base_url})
 107             if '>You have enrolled in' in webpage:
 108                 self.to_screen('%s: Successfully enrolled in the course' % course_id)
 109
 110     def _download_lecture(self, course_id, lecture_id):
 111         return self._download_json(
 112             'https://www.udemy.com/api-2.0/users/me/subscribed-courses/%s/lectures/%s?'
 113             % (course_id, lecture_id),
 114             lecture_id, 'Downloading lecture JSON', query={
 115                 'fields[lecture]': 'title,description,view_html,asset',
 116                 'fields[asset]': 'asset_type,stream_url,thumbnail_url,download_urls,stream_urls,captions,data',
 117             })
 118
 119     def _handle_error(self, response):
 120         if not isinstance(response, dict):
 121             return
 122         error = response.get('error')
 123         if error:
 124             error_str = 'Udemy returned error #%s: %s' % (error.get('code'), error.get('message'))
 125             error_data = error.get('data')
 126             if error_data:
 127                 error_str += ' - %s' % error_data.get('formErrors')
 128             raise ExtractorError(error_str, expected=True)
 129
 130     def _download_webpage_handle(self, *args, **kwargs):
 131         headers = kwargs.get('headers', {}).copy()
 132         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'
 133         kwargs['headers'] = headers
 134         ret = super(UdemyIE, self)._download_webpage_handle(
 135             *args, **compat_kwargs(kwargs))
 136         if not ret:
 137             return ret
 138         webpage, _ = ret
 139         if any(p in webpage for p in (
 140                 '>Please verify you are a human',
 141                 'Access to this page has been denied because we believe you are using automation tools to browse the website',
 142                 '"_pxCaptcha"')):
 143             raise ExtractorError(
 144                 'Udemy asks you to solve a CAPTCHA. Login with browser, '
 145                 'solve CAPTCHA, then export cookies and pass cookie file to '
 146                 'youtube-dl with --cookies.', expected=True)
 147         return ret
 148
 149     def _download_json(self, url_or_request, *args, **kwargs):
 150         headers = {
 151             'X-Udemy-Snail-Case': 'true',
 152             'X-Requested-With': 'XMLHttpRequest',
 153         }
 154         for cookie in self._downloader.cookiejar:
 155             if cookie.name == 'client_id':
 156                 headers['X-Udemy-Client-Id'] = cookie.value
 157             elif cookie.name == 'access_token':
 158                 headers['X-Udemy-Bearer-Token'] = cookie.value
 159                 headers['X-Udemy-Authorization'] = 'Bearer %s' % cookie.value
 160
 161         if isinstance(url_or_request, compat_urllib_request.Request):
 162             for header, value in headers.items():
 163                 url_or_request.add_header(header, value)
 164         else:
 165             url_or_request = sanitized_Request(url_or_request, headers=headers)
 166
 167         response = super(UdemyIE, self)._download_json(url_or_request, *args, **kwargs)
 168         self._handle_error(response)
 169         return response
 170
 171     def _real_initialize(self):
 172         self._login()
 173
 174     def _login(self):
 175         username, password = self._get_login_info()
 176         if username is None:
 177             return
 178
 179         login_popup = self._download_webpage(
 180             self._LOGIN_URL, None, 'Downloading login popup')
 181
 182         def is_logged(webpage):
 183             return any(re.search(p, webpage) for p in (
 184                 r'href=["\'](?:https://www\.udemy\.com)?/user/logout/',
 185                 r'>Logout<'))
 186
 187         # already logged in
 188         if is_logged(login_popup):
 189             return
 190
 191         login_form = self._form_hidden_inputs('login-form', login_popup)
 192
 193         login_form.update({
 194             'email': username,
 195             'password': password,
 196         })
 197
 198         response = self._download_webpage(
 199             self._LOGIN_URL, None, 'Logging in',
 200             data=urlencode_postdata(login_form),
 201             headers={
 202                 'Referer': self._ORIGIN_URL,
 203                 'Origin': self._ORIGIN_URL,
 204             })
 205
 206         if not is_logged(response):
 207             error = self._html_search_regex(
 208                 r'(?s)<div[^>]+class="form-errors[^"]*">(.+?)</div>',
 209                 response, 'error message', default=None)
 210             if error:
 211                 raise ExtractorError('Unable to login: %s' % error, expected=True)
 212             raise ExtractorError('Unable to log in')
 213
 214     def _real_extract(self, url):
 215         lecture_id = self._match_id(url)
 216
 217         webpage = self._download_webpage(url, lecture_id)
 218
 219         course_id, _ = self._extract_course_info(webpage, lecture_id)
 220
 221         try:
 222             lecture = self._download_lecture(course_id, lecture_id)
 223         except ExtractorError as e:
 224             # Error could possibly mean we are not enrolled in the course
 225             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 226                 self._enroll_course(url, webpage, course_id)
 227                 lecture = self._download_lecture(course_id, lecture_id)
 228             else:
 229                 raise
 230
 231         title = lecture['title']
 232         description = lecture.get('description')
 233
 234         asset = lecture['asset']
 235
 236         asset_type = asset.get('asset_type') or asset.get('assetType')
 237         if asset_type != 'Video':
 238             raise ExtractorError(
 239                 'Lecture %s is not a video' % lecture_id, expected=True)
 240
 241         stream_url = asset.get('stream_url') or asset.get('streamUrl')
 242         if stream_url:
 243             youtube_url = self._search_regex(
 244                 r'(https?://www\.youtube\.com/watch\?v=.*)', stream_url, 'youtube URL', default=None)
 245             if youtube_url:
 246                 return self.url_result(youtube_url, 'Youtube')
 247
 248         video_id = compat_str(asset['id'])
 249         thumbnail = asset.get('thumbnail_url') or asset.get('thumbnailUrl')
 250         duration = float_or_none(asset.get('data', {}).get('duration'))
 251
 252         subtitles = {}
 253         automatic_captions = {}
 254
 255         formats = []
 256
 257         def extract_output_format(src, f_id):
 258             return {
 259                 'url': src.get('url'),
 260                 'format_id': '%sp' % (src.get('height') or f_id),
 261                 'width': int_or_none(src.get('width')),
 262                 'height': int_or_none(src.get('height')),
 263                 'vbr': int_or_none(src.get('video_bitrate_in_kbps')),
 264                 'vcodec': src.get('video_codec'),
 265                 'fps': int_or_none(src.get('frame_rate')),
 266                 'abr': int_or_none(src.get('audio_bitrate_in_kbps')),
 267                 'acodec': src.get('audio_codec'),
 268                 'asr': int_or_none(src.get('audio_sample_rate')),
 269                 'tbr': int_or_none(src.get('total_bitrate_in_kbps')),
 270                 'filesize': int_or_none(src.get('file_size_in_bytes')),
 271             }
 272
 273         outputs = asset.get('data', {}).get('outputs')
 274         if not isinstance(outputs, dict):
 275             outputs = {}
 276
 277         def add_output_format_meta(f, key):
 278             output = outputs.get(key)
 279             if isinstance(output, dict):
 280                 output_format = extract_output_format(output, key)
 281                 output_format.update(f)
 282                 return output_format
 283             return f
 284
 285         def extract_formats(source_list):
 286             if not isinstance(source_list, list):
 287                 return
 288             for source in source_list:
 289                 video_url = url_or_none(source.get('file') or source.get('src'))
 290                 if not video_url:
 291                     continue
 292                 if source.get('type') == 'application/x-mpegURL' or determine_ext(video_url) == 'm3u8':
 293                     formats.extend(self._extract_m3u8_formats(
 294                         video_url, video_id, 'mp4', entry_protocol='m3u8_native',
 295                         m3u8_id='hls', fatal=False))
 296                     continue
 297                 format_id = source.get('label')
 298                 f = {
 299                     'url': video_url,
 300                     'format_id': '%sp' % format_id,
 301                     'height': int_or_none(format_id),
 302                 }
 303                 if format_id:
 304                     # Some videos contain additional metadata (e.g.
 305                     # https://www.udemy.com/ios9-swift/learn/#/lecture/3383208)
 306                     f = add_output_format_meta(f, format_id)
 307                 formats.append(f)
 308
 309         def extract_subtitles(track_list):
 310             if not isinstance(track_list, list):
 311                 return
 312             for track in track_list:
 313                 if not isinstance(track, dict):
 314                     continue
 315                 if track.get('kind') != 'captions':
 316                     continue
 317                 src = url_or_none(track.get('src'))
 318                 if not src:
 319                     continue
 320                 lang = track.get('language') or track.get(
 321                     'srclang') or track.get('label')
 322                 sub_dict = automatic_captions if track.get(
 323                     'autogenerated') is True else subtitles
 324                 sub_dict.setdefault(lang, []).append({
 325                     'url': src,
 326                 })
 327
 328         for url_kind in ('download', 'stream'):
 329             urls = asset.get('%s_urls' % url_kind)
 330             if isinstance(urls, dict):
 331                 extract_formats(urls.get('Video'))
 332
 333         captions = asset.get('captions')
 334         if isinstance(captions, list):
 335             for cc in captions:
 336                 if not isinstance(cc, dict):
 337                     continue
 338                 cc_url = url_or_none(cc.get('url'))
 339                 if not cc_url:
 340                     continue
 341                 lang = try_get(cc, lambda x: x['locale']['locale'], compat_str)
 342                 sub_dict = (automatic_captions if cc.get('source') == 'auto'
 343                             else subtitles)
 344                 sub_dict.setdefault(lang or 'en', []).append({
 345                     'url': cc_url,
 346                 })
 347
 348         view_html = lecture.get('view_html')
 349         if view_html:
 350             view_html_urls = set()
 351             for source in re.findall(r'<source[^>]+>', view_html):
 352                 attributes = extract_attributes(source)
 353                 src = attributes.get('src')
 354                 if not src:
 355                     continue
 356                 res = attributes.get('data-res')
 357                 height = int_or_none(res)
 358                 if src in view_html_urls:
 359                     continue
 360                 view_html_urls.add(src)
 361                 if attributes.get('type') == 'application/x-mpegURL' or determine_ext(src) == 'm3u8':
 362                     m3u8_formats = self._extract_m3u8_formats(
 363                         src, video_id, 'mp4', entry_protocol='m3u8_native',
 364                         m3u8_id='hls', fatal=False)
 365                     for f in m3u8_formats:
 366                         m = re.search(r'/hls_(?P<height>\d{3,4})_(?P<tbr>\d{2,})/', f['url'])
 367                         if m:
 368                             if not f.get('height'):
 369                                 f['height'] = int(m.group('height'))
 370                             if not f.get('tbr'):
 371                                 f['tbr'] = int(m.group('tbr'))
 372                     formats.extend(m3u8_formats)
 373                 else:
 374                     formats.append(add_output_format_meta({
 375                         'url': src,
 376                         'format_id': '%dp' % height if height else None,
 377                         'height': height,
 378                     }, res))
 379
 380             # react rendition since 2017.04.15 (see
 381             # https://github.com/ytdl-org/youtube-dl/issues/12744)
 382             data = self._parse_json(
 383                 self._search_regex(
 384                     r'videojs-setup-data=(["\'])(?P<data>{.+?})\1', view_html,
 385                     'setup data', default='{}', group='data'), video_id,
 386                 transform_source=unescapeHTML, fatal=False)
 387             if data and isinstance(data, dict):
 388                 extract_formats(data.get('sources'))
 389                 if not duration:
 390                     duration = int_or_none(data.get('duration'))
 391                 extract_subtitles(data.get('tracks'))
 392
 393             if not subtitles and not automatic_captions:
 394                 text_tracks = self._parse_json(
 395                     self._search_regex(
 396                         r'text-tracks=(["\'])(?P<data>\[.+?\])\1', view_html,
 397                         'text tracks', default='{}', group='data'), video_id,
 398                     transform_source=lambda s: js_to_json(unescapeHTML(s)),
 399                     fatal=False)
 400                 extract_subtitles(text_tracks)
 401
 402         if not formats and outputs:
 403             for format_id, output in outputs.items():
 404                 f = extract_output_format(output, format_id)
 405                 if f.get('url'):
 406                     formats.append(f)
 407
 408         self._sort_formats(formats, field_preference=('height', 'width', 'tbr', 'format_id'))
 409
 410         return {
 411             'id': video_id,
 412             'title': title,
 413             'description': description,
 414             'thumbnail': thumbnail,
 415             'duration': duration,
 416             'formats': formats,
 417             'subtitles': subtitles,
 418             'automatic_captions': automatic_captions,
 419         }
 420
 421
 422 class UdemyCourseIE(UdemyIE):
 423     IE_NAME = 'udemy:course'
 424     _VALID_URL = r'https?://(?:[^/]+\.)?udemy\.com/(?P<id>[^/?#&]+)'
 425     _TESTS = [{
 426         'url': 'https://www.udemy.com/java-tutorial/',
 427         'only_matching': True,
 428     }, {
 429         'url': 'https://wipro.udemy.com/java-tutorial/',
 430         'only_matching': True,
 431     }]
 432
 433     @classmethod
 434     def suitable(cls, url):
 435         return False if UdemyIE.suitable(url) else super(UdemyCourseIE, cls).suitable(url)
 436
 437     def _real_extract(self, url):
 438         course_path = self._match_id(url)
 439
 440         webpage = self._download_webpage(url, course_path)
 441
 442         course_id, title = self._extract_course_info(webpage, course_path)
 443
 444         self._enroll_course(url, webpage, course_id)
 445
 446         response = self._download_json(
 447             'https://www.udemy.com/api-2.0/courses/%s/cached-subscriber-curriculum-items' % course_id,
 448             course_id, 'Downloading course curriculum', query={
 449                 'fields[chapter]': 'title,object_index',
 450                 'fields[lecture]': 'title,asset',
 451                 'page_size': '1000',
 452             })
 453
 454         entries = []
 455         chapter, chapter_number = [None] * 2
 456         for entry in response['results']:
 457             clazz = entry.get('_class')
 458             if clazz == 'lecture':
 459                 asset = entry.get('asset')
 460                 if isinstance(asset, dict):
 461                     asset_type = asset.get('asset_type') or asset.get('assetType')
 462                     if asset_type != 'Video':
 463                         continue
 464                 lecture_id = entry.get('id')
 465                 if lecture_id:
 466                     entry = {
 467                         '_type': 'url_transparent',
 468                         'url': 'https://www.udemy.com/%s/learn/v4/t/lecture/%s' % (course_path, entry['id']),
 469                         'title': entry.get('title'),
 470                         'ie_key': UdemyIE.ie_key(),
 471                     }
 472                     if chapter_number:
 473                         entry['chapter_number'] = chapter_number
 474                     if chapter:
 475                         entry['chapter'] = chapter
 476                     entries.append(entry)
 477             elif clazz == 'chapter':
 478                 chapter_number = entry.get('object_index')
 479                 chapter = entry.get('title')
 480
 481         return self.playlist_result(entries, course_id, title)