_ Git - youtube-dl/blob - 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                         www\.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
  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             r'data-course-id=["\'](\d+)', webpage, 'course id')
  77         return course_id, course.get('title')
  78
  79     def _enroll_course(self, base_url, webpage, course_id):
  80         def combine_url(base_url, url):
  81             return compat_urlparse.urljoin(base_url, url) if not url.startswith('http') else url
  82
  83         checkout_url = unescapeHTML(self._search_regex(
  84             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/(?:payment|cart)/checkout/.+?)\1',
  85             webpage, 'checkout url', group='url', default=None))
  86         if checkout_url:
  87             raise ExtractorError(
  88                 'Course %s is not free. You have to pay for it before you can download. '
  89                 'Use this URL to confirm purchase: %s'
  90                 % (course_id, 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('%s: Successfully enrolled in the course' % course_id)
 103
 104     def _download_lecture(self, course_id, lecture_id):
 105         return self._download_json(
 106             'https://www.udemy.com/api-2.0/users/me/subscribed-courses/%s/lectures/%s?'
 107             % (course_id, lecture_id),
 108             lecture_id, 'Downloading lecture JSON', query={
 109                 'fields[lecture]': 'title,description,view_html,asset',
 110                 'fields[asset]': 'asset_type,stream_url,thumbnail_url,download_urls,stream_urls,captions,data',
 111             })
 112
 113     def _handle_error(self, response):
 114         if not isinstance(response, dict):
 115             return
 116         error = response.get('error')
 117         if error:
 118             error_str = 'Udemy returned error #%s: %s' % (error.get('code'), error.get('message'))
 119             error_data = error.get('data')
 120             if error_data:
 121                 error_str += ' - %s' % error_data.get('formErrors')
 122             raise ExtractorError(error_str, expected=True)
 123
 124     def _download_webpage_handle(self, *args, **kwargs):
 125         headers = kwargs.get('headers', {}).copy()
 126         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'
 127         kwargs['headers'] = headers
 128         ret = super(UdemyIE, self)._download_webpage_handle(
 129             *args, **compat_kwargs(kwargs))
 130         if not ret:
 131             return ret
 132         webpage, _ = ret
 133         if any(p in webpage for p in (
 134                 '>Please verify you are a human',
 135                 'Access to this page has been denied because we believe you are using automation tools to browse the website',
 136                 '"_pxCaptcha"')):
 137             raise ExtractorError(
 138                 'Udemy asks you to solve a CAPTCHA. Login with browser, '
 139                 'solve CAPTCHA, then export cookies and pass cookie file to '
 140                 'youtube-dl with --cookies.', expected=True)
 141         return ret
 142
 143     def _download_json(self, url_or_request, *args, **kwargs):
 144         headers = {
 145             'X-Udemy-Snail-Case': 'true',
 146             'X-Requested-With': 'XMLHttpRequest',
 147         }
 148         for cookie in self._downloader.cookiejar:
 149             if cookie.name == 'client_id':
 150                 headers['X-Udemy-Client-Id'] = cookie.value
 151             elif cookie.name == 'access_token':
 152                 headers['X-Udemy-Bearer-Token'] = cookie.value
 153                 headers['X-Udemy-Authorization'] = 'Bearer %s' % cookie.value
 154
 155         if isinstance(url_or_request, compat_urllib_request.Request):
 156             for header, value in headers.items():
 157                 url_or_request.add_header(header, value)
 158         else:
 159             url_or_request = sanitized_Request(url_or_request, headers=headers)
 160
 161         response = super(UdemyIE, self)._download_json(url_or_request, *args, **kwargs)
 162         self._handle_error(response)
 163         return response
 164
 165     def _real_initialize(self):
 166         self._login()
 167
 168     def _login(self):
 169         username, password = self._get_login_info()
 170         if username is None:
 171             return
 172
 173         login_popup = self._download_webpage(
 174             self._LOGIN_URL, None, 'Downloading login popup')
 175
 176         def is_logged(webpage):
 177             return any(re.search(p, webpage) for p in (
 178                 r'href=["\'](?:https://www\.udemy\.com)?/user/logout/',
 179                 r'>Logout<'))
 180
 181         # already logged in
 182         if is_logged(login_popup):
 183             return
 184
 185         login_form = self._form_hidden_inputs('login-form', login_popup)
 186
 187         login_form.update({
 188             'email': username,
 189             'password': password,
 190         })
 191
 192         response = self._download_webpage(
 193             self._LOGIN_URL, None, 'Logging in',
 194             data=urlencode_postdata(login_form),
 195             headers={
 196                 'Referer': self._ORIGIN_URL,
 197                 'Origin': self._ORIGIN_URL,
 198             })
 199
 200         if not is_logged(response):
 201             error = self._html_search_regex(
 202                 r'(?s)<div[^>]+class="form-errors[^"]*">(.+?)</div>',
 203                 response, 'error message', default=None)
 204             if error:
 205                 raise ExtractorError('Unable to login: %s' % error, expected=True)
 206             raise ExtractorError('Unable to log in')
 207
 208     def _real_extract(self, url):
 209         lecture_id = self._match_id(url)
 210
 211         webpage = self._download_webpage(url, lecture_id)
 212
 213         course_id, _ = self._extract_course_info(webpage, lecture_id)
 214
 215         try:
 216             lecture = self._download_lecture(course_id, lecture_id)
 217         except ExtractorError as e:
 218             # Error could possibly mean we are not enrolled in the course
 219             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 220                 self._enroll_course(url, webpage, course_id)
 221                 lecture = self._download_lecture(course_id, lecture_id)
 222             else:
 223                 raise
 224
 225         title = lecture['title']
 226         description = lecture.get('description')
 227
 228         asset = lecture['asset']
 229
 230         asset_type = asset.get('asset_type') or asset.get('assetType')
 231         if asset_type != 'Video':
 232             raise ExtractorError(
 233                 'Lecture %s is not a video' % lecture_id, expected=True)
 234
 235         stream_url = asset.get('stream_url') or asset.get('streamUrl')
 236         if stream_url:
 237             youtube_url = self._search_regex(
 238                 r'(https?://www\.youtube\.com/watch\?v=.*)', stream_url, 'youtube URL', default=None)
 239             if youtube_url:
 240                 return self.url_result(youtube_url, 'Youtube')
 241
 242         video_id = compat_str(asset['id'])
 243         thumbnail = asset.get('thumbnail_url') or asset.get('thumbnailUrl')
 244         duration = float_or_none(asset.get('data', {}).get('duration'))
 245
 246         subtitles = {}
 247         automatic_captions = {}
 248
 249         formats = []
 250
 251         def extract_output_format(src, f_id):
 252             return {
 253                 'url': src.get('url'),
 254                 'format_id': '%sp' % (src.get('height') or f_id),
 255                 'width': int_or_none(src.get('width')),
 256                 'height': int_or_none(src.get('height')),
 257                 'vbr': int_or_none(src.get('video_bitrate_in_kbps')),
 258                 'vcodec': src.get('video_codec'),
 259                 'fps': int_or_none(src.get('frame_rate')),
 260                 'abr': int_or_none(src.get('audio_bitrate_in_kbps')),
 261                 'acodec': src.get('audio_codec'),
 262                 'asr': int_or_none(src.get('audio_sample_rate')),
 263                 'tbr': int_or_none(src.get('total_bitrate_in_kbps')),
 264                 'filesize': int_or_none(src.get('file_size_in_bytes')),
 265             }
 266
 267         outputs = asset.get('data', {}).get('outputs')
 268         if not isinstance(outputs, dict):
 269             outputs = {}
 270
 271         def add_output_format_meta(f, key):
 272             output = outputs.get(key)
 273             if isinstance(output, dict):
 274                 output_format = extract_output_format(output, key)
 275                 output_format.update(f)
 276                 return output_format
 277             return f
 278
 279         def extract_formats(source_list):
 280             if not isinstance(source_list, list):
 281                 return
 282             for source in source_list:
 283                 video_url = url_or_none(source.get('file') or source.get('src'))
 284                 if not video_url:
 285                     continue
 286                 if source.get('type') == 'application/x-mpegURL' or determine_ext(video_url) == 'm3u8':
 287                     formats.extend(self._extract_m3u8_formats(
 288                         video_url, video_id, 'mp4', entry_protocol='m3u8_native',
 289                         m3u8_id='hls', fatal=False))
 290                     continue
 291                 format_id = source.get('label')
 292                 f = {
 293                     'url': video_url,
 294                     'format_id': '%sp' % format_id,
 295                     'height': int_or_none(format_id),
 296                 }
 297                 if format_id:
 298                     # Some videos contain additional metadata (e.g.
 299                     # https://www.udemy.com/ios9-swift/learn/#/lecture/3383208)
 300                     f = add_output_format_meta(f, format_id)
 301                 formats.append(f)
 302
 303         def extract_subtitles(track_list):
 304             if not isinstance(track_list, list):
 305                 return
 306             for track in track_list:
 307                 if not isinstance(track, dict):
 308                     continue
 309                 if track.get('kind') != 'captions':
 310                     continue
 311                 src = url_or_none(track.get('src'))
 312                 if not src:
 313                     continue
 314                 lang = track.get('language') or track.get(
 315                     'srclang') or track.get('label')
 316                 sub_dict = automatic_captions if track.get(
 317                     'autogenerated') is True else subtitles
 318                 sub_dict.setdefault(lang, []).append({
 319                     'url': src,
 320                 })
 321
 322         for url_kind in ('download', 'stream'):
 323             urls = asset.get('%s_urls' % url_kind)
 324             if isinstance(urls, dict):
 325                 extract_formats(urls.get('Video'))
 326
 327         captions = asset.get('captions')
 328         if isinstance(captions, list):
 329             for cc in captions:
 330                 if not isinstance(cc, dict):
 331                     continue
 332                 cc_url = url_or_none(cc.get('url'))
 333                 if not cc_url:
 334                     continue
 335                 lang = try_get(cc, lambda x: x['locale']['locale'], compat_str)
 336                 sub_dict = (automatic_captions if cc.get('source') == 'auto'
 337                             else subtitles)
 338                 sub_dict.setdefault(lang or 'en', []).append({
 339                     'url': cc_url,
 340                 })
 341
 342         view_html = lecture.get('view_html')
 343         if view_html:
 344             view_html_urls = set()
 345             for source in re.findall(r'<source[^>]+>', view_html):
 346                 attributes = extract_attributes(source)
 347                 src = attributes.get('src')
 348                 if not src:
 349                     continue
 350                 res = attributes.get('data-res')
 351                 height = int_or_none(res)
 352                 if src in view_html_urls:
 353                     continue
 354                 view_html_urls.add(src)
 355                 if attributes.get('type') == 'application/x-mpegURL' or determine_ext(src) == 'm3u8':
 356                     m3u8_formats = self._extract_m3u8_formats(
 357                         src, video_id, 'mp4', entry_protocol='m3u8_native',
 358                         m3u8_id='hls', fatal=False)
 359                     for f in m3u8_formats:
 360                         m = re.search(r'/hls_(?P<height>\d{3,4})_(?P<tbr>\d{2,})/', f['url'])
 361                         if m:
 362                             if not f.get('height'):
 363                                 f['height'] = int(m.group('height'))
 364                             if not f.get('tbr'):
 365                                 f['tbr'] = int(m.group('tbr'))
 366                     formats.extend(m3u8_formats)
 367                 else:
 368                     formats.append(add_output_format_meta({
 369                         'url': src,
 370                         'format_id': '%dp' % height if height else None,
 371                         'height': height,
 372                     }, res))
 373
 374             # react rendition since 2017.04.15 (see
 375             # https://github.com/rg3/youtube-dl/issues/12744)
 376             data = self._parse_json(
 377                 self._search_regex(
 378                     r'videojs-setup-data=(["\'])(?P<data>{.+?})\1', view_html,
 379                     'setup data', default='{}', group='data'), video_id,
 380                 transform_source=unescapeHTML, fatal=False)
 381             if data and isinstance(data, dict):
 382                 extract_formats(data.get('sources'))
 383                 if not duration:
 384                     duration = int_or_none(data.get('duration'))
 385                 extract_subtitles(data.get('tracks'))
 386
 387             if not subtitles and not automatic_captions:
 388                 text_tracks = self._parse_json(
 389                     self._search_regex(
 390                         r'text-tracks=(["\'])(?P<data>\[.+?\])\1', view_html,
 391                         'text tracks', default='{}', group='data'), video_id,
 392                     transform_source=lambda s: js_to_json(unescapeHTML(s)),
 393                     fatal=False)
 394                 extract_subtitles(text_tracks)
 395
 396         if not formats and outputs:
 397             for format_id, output in outputs.items():
 398                 f = extract_output_format(output, format_id)
 399                 if f.get('url'):
 400                     formats.append(f)
 401
 402         self._sort_formats(formats, field_preference=('height', 'width', 'tbr', 'format_id'))
 403
 404         return {
 405             'id': video_id,
 406             'title': title,
 407             'description': description,
 408             'thumbnail': thumbnail,
 409             'duration': duration,
 410             'formats': formats,
 411             'subtitles': subtitles,
 412             'automatic_captions': automatic_captions,
 413         }
 414
 415
 416 class UdemyCourseIE(UdemyIE):
 417     IE_NAME = 'udemy:course'
 418     _VALID_URL = r'https?://(?:www\.)?udemy\.com/(?P<id>[^/?#&]+)'
 419     _TESTS = []
 420
 421     @classmethod
 422     def suitable(cls, url):
 423         return False if UdemyIE.suitable(url) else super(UdemyCourseIE, cls).suitable(url)
 424
 425     def _real_extract(self, url):
 426         course_path = self._match_id(url)
 427
 428         webpage = self._download_webpage(url, course_path)
 429
 430         course_id, title = self._extract_course_info(webpage, course_path)
 431
 432         self._enroll_course(url, webpage, course_id)
 433
 434         response = self._download_json(
 435             'https://www.udemy.com/api-2.0/courses/%s/cached-subscriber-curriculum-items' % course_id,
 436             course_id, 'Downloading course curriculum', query={
 437                 'fields[chapter]': 'title,object_index',
 438                 'fields[lecture]': 'title,asset',
 439                 'page_size': '1000',
 440             })
 441
 442         entries = []
 443         chapter, chapter_number = [None] * 2
 444         for entry in response['results']:
 445             clazz = entry.get('_class')
 446             if clazz == 'lecture':
 447                 asset = entry.get('asset')
 448                 if isinstance(asset, dict):
 449                     asset_type = asset.get('asset_type') or asset.get('assetType')
 450                     if asset_type != 'Video':
 451                         continue
 452                 lecture_id = entry.get('id')
 453                 if lecture_id:
 454                     entry = {
 455                         '_type': 'url_transparent',
 456                         'url': 'https://www.udemy.com/%s/learn/v4/t/lecture/%s' % (course_path, entry['id']),
 457                         'title': entry.get('title'),
 458                         'ie_key': UdemyIE.ie_key(),
 459                     }
 460                     if chapter_number:
 461                         entry['chapter_number'] = chapter_number
 462                     if chapter:
 463                         entry['chapter'] = chapter
 464                     entries.append(entry)
 465             elif clazz == 'chapter':
 466                 chapter_number = entry.get('object_index')
 467                 chapter = entry.get('title')
 468
 469         return self.playlist_result(entries, course_id, title)