Raphaël G. Git Repositories - youtubedl/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_str,
   9     compat_urllib_request,
  10     compat_urlparse,
  11 )
  12 from ..utils import (
  13     determine_ext,
  14     extract_attributes,
  15     ExtractorError,
  16     float_or_none,
  17     int_or_none,
  18     sanitized_Request,
  19     unescapeHTML,
  20     urlencode_postdata,
  21 )
  22
  23
  24 class UdemyIE(InfoExtractor):
  25     IE_NAME = 'udemy'
  26     _VALID_URL = r'''(?x)
  27                     https?://
  28                         www\.udemy\.com/
  29                         (?:
  30                             [^#]+\#/lecture/|
  31                             lecture/view/?\?lectureId=|
  32                             [^/]+/learn/v4/t/lecture/
  33                         )
  34                         (?P<id>\d+)
  35                     '''
  36     _LOGIN_URL = 'https://www.udemy.com/join/login-popup/?displayType=ajax&showSkipButton=1'
  37     _ORIGIN_URL = 'https://www.udemy.com'
  38     _NETRC_MACHINE = 'udemy'
  39
  40     _TESTS = [{
  41         'url': 'https://www.udemy.com/java-tutorial/#/lecture/172757',
  42         'md5': '98eda5b657e752cf945d8445e261b5c5',
  43         'info_dict': {
  44             'id': '160614',
  45             'ext': 'mp4',
  46             'title': 'Introduction and Installation',
  47             'description': 'md5:c0d51f6f21ef4ec65f091055a5eef876',
  48             'duration': 579.29,
  49         },
  50         'skip': 'Requires udemy account credentials',
  51     }, {
  52         # new URL schema
  53         'url': 'https://www.udemy.com/electric-bass-right-from-the-start/learn/v4/t/lecture/4580906',
  54         'only_matching': True,
  55     }]
  56
  57     def _extract_course_info(self, webpage, video_id):
  58         course = self._parse_json(
  59             unescapeHTML(self._search_regex(
  60                 r'ng-init=["\'].*\bcourse=({.+?});', webpage, 'course', default='{}')),
  61             video_id, fatal=False) or {}
  62         course_id = course.get('id') or self._search_regex(
  63             (r'&quot;id&quot;\s*:\s*(\d+)', r'data-course-id=["\'](\d+)'),
  64             webpage, 'course id')
  65         return course_id, course.get('title')
  66
  67     def _enroll_course(self, base_url, webpage, course_id):
  68         def combine_url(base_url, url):
  69             return compat_urlparse.urljoin(base_url, url) if not url.startswith('http') else url
  70
  71         checkout_url = unescapeHTML(self._search_regex(
  72             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/payment/checkout/.+?)\1',
  73             webpage, 'checkout url', group='url', default=None))
  74         if checkout_url:
  75             raise ExtractorError(
  76                 'Course %s is not free. You have to pay for it before you can download. '
  77                 'Use this URL to confirm purchase: %s'
  78                 % (course_id, combine_url(base_url, checkout_url)),
  79                 expected=True)
  80
  81         enroll_url = unescapeHTML(self._search_regex(
  82             r'href=(["\'])(?P<url>(?:https?://(?:www\.)?udemy\.com)?/course/subscribe/.+?)\1',
  83             webpage, 'enroll url', group='url', default=None))
  84         if enroll_url:
  85             webpage = self._download_webpage(
  86                 combine_url(base_url, enroll_url),
  87                 course_id, 'Enrolling in the course',
  88                 headers={'Referer': base_url})
  89             if '>You have enrolled in' in webpage:
  90                 self.to_screen('%s: Successfully enrolled in the course' % course_id)
  91
  92     def _download_lecture(self, course_id, lecture_id):
  93         return self._download_json(
  94             'https://www.udemy.com/api-2.0/users/me/subscribed-courses/%s/lectures/%s?'
  95             % (course_id, lecture_id),
  96             lecture_id, 'Downloading lecture JSON', query={
  97                 'fields[lecture]': 'title,description,view_html,asset',
  98                 'fields[asset]': 'asset_type,stream_url,thumbnail_url,download_urls,data',
  99             })
 100
 101     def _handle_error(self, response):
 102         if not isinstance(response, dict):
 103             return
 104         error = response.get('error')
 105         if error:
 106             error_str = 'Udemy returned error #%s: %s' % (error.get('code'), error.get('message'))
 107             error_data = error.get('data')
 108             if error_data:
 109                 error_str += ' - %s' % error_data.get('formErrors')
 110             raise ExtractorError(error_str, expected=True)
 111
 112     def _download_json(self, url_or_request, *args, **kwargs):
 113         headers = {
 114             'X-Udemy-Snail-Case': 'true',
 115             'X-Requested-With': 'XMLHttpRequest',
 116         }
 117         for cookie in self._downloader.cookiejar:
 118             if cookie.name == 'client_id':
 119                 headers['X-Udemy-Client-Id'] = cookie.value
 120             elif cookie.name == 'access_token':
 121                 headers['X-Udemy-Bearer-Token'] = cookie.value
 122                 headers['X-Udemy-Authorization'] = 'Bearer %s' % cookie.value
 123
 124         if isinstance(url_or_request, compat_urllib_request.Request):
 125             for header, value in headers.items():
 126                 url_or_request.add_header(header, value)
 127         else:
 128             url_or_request = sanitized_Request(url_or_request, headers=headers)
 129
 130         response = super(UdemyIE, self)._download_json(url_or_request, *args, **kwargs)
 131         self._handle_error(response)
 132         return response
 133
 134     def _real_initialize(self):
 135         self._login()
 136
 137     def _login(self):
 138         (username, password) = self._get_login_info()
 139         if username is None:
 140             return
 141
 142         login_popup = self._download_webpage(
 143             self._LOGIN_URL, None, 'Downloading login popup')
 144
 145         def is_logged(webpage):
 146             return any(re.search(p, webpage) for p in (
 147                 r'href=["\'](?:https://www\.udemy\.com)?/user/logout/',
 148                 r'>Logout<'))
 149
 150         # already logged in
 151         if is_logged(login_popup):
 152             return
 153
 154         login_form = self._form_hidden_inputs('login-form', login_popup)
 155
 156         login_form.update({
 157             'email': username,
 158             'password': password,
 159         })
 160
 161         response = self._download_webpage(
 162             self._LOGIN_URL, None, 'Logging in as %s' % username,
 163             data=urlencode_postdata(login_form),
 164             headers={
 165                 'Referer': self._ORIGIN_URL,
 166                 'Origin': self._ORIGIN_URL,
 167             })
 168
 169         if not is_logged(response):
 170             error = self._html_search_regex(
 171                 r'(?s)<div[^>]+class="form-errors[^"]*">(.+?)</div>',
 172                 response, 'error message', default=None)
 173             if error:
 174                 raise ExtractorError('Unable to login: %s' % error, expected=True)
 175             raise ExtractorError('Unable to log in')
 176
 177     def _real_extract(self, url):
 178         lecture_id = self._match_id(url)
 179
 180         webpage = self._download_webpage(url, lecture_id)
 181
 182         course_id, _ = self._extract_course_info(webpage, lecture_id)
 183
 184         try:
 185             lecture = self._download_lecture(course_id, lecture_id)
 186         except ExtractorError as e:
 187             # Error could possibly mean we are not enrolled in the course
 188             if isinstance(e.cause, compat_HTTPError) and e.cause.code == 403:
 189                 self._enroll_course(url, webpage, course_id)
 190                 lecture = self._download_lecture(course_id, lecture_id)
 191             else:
 192                 raise
 193
 194         title = lecture['title']
 195         description = lecture.get('description')
 196
 197         asset = lecture['asset']
 198
 199         asset_type = asset.get('asset_type') or asset.get('assetType')
 200         if asset_type != 'Video':
 201             raise ExtractorError(
 202                 'Lecture %s is not a video' % lecture_id, expected=True)
 203
 204         stream_url = asset.get('stream_url') or asset.get('streamUrl')
 205         if stream_url:
 206             youtube_url = self._search_regex(
 207                 r'(https?://www\.youtube\.com/watch\?v=.*)', stream_url, 'youtube URL', default=None)
 208             if youtube_url:
 209                 return self.url_result(youtube_url, 'Youtube')
 210
 211         video_id = compat_str(asset['id'])
 212         thumbnail = asset.get('thumbnail_url') or asset.get('thumbnailUrl')
 213         duration = float_or_none(asset.get('data', {}).get('duration'))
 214
 215         subtitles = {}
 216         automatic_captions = {}
 217
 218         formats = []
 219
 220         def extract_output_format(src, f_id):
 221             return {
 222                 'url': src['url'],
 223                 'format_id': '%sp' % (src.get('height') or f_id),
 224                 'width': int_or_none(src.get('width')),
 225                 'height': int_or_none(src.get('height')),
 226                 'vbr': int_or_none(src.get('video_bitrate_in_kbps')),
 227                 'vcodec': src.get('video_codec'),
 228                 'fps': int_or_none(src.get('frame_rate')),
 229                 'abr': int_or_none(src.get('audio_bitrate_in_kbps')),
 230                 'acodec': src.get('audio_codec'),
 231                 'asr': int_or_none(src.get('audio_sample_rate')),
 232                 'tbr': int_or_none(src.get('total_bitrate_in_kbps')),
 233                 'filesize': int_or_none(src.get('file_size_in_bytes')),
 234             }
 235
 236         outputs = asset.get('data', {}).get('outputs')
 237         if not isinstance(outputs, dict):
 238             outputs = {}
 239
 240         def add_output_format_meta(f, key):
 241             output = outputs.get(key)
 242             if isinstance(output, dict):
 243                 output_format = extract_output_format(output, key)
 244                 output_format.update(f)
 245                 return output_format
 246             return f
 247
 248         def extract_formats(source_list):
 249             if not isinstance(source_list, list):
 250                 return
 251             for source in source_list:
 252                 video_url = source.get('file') or source.get('src')
 253                 if not video_url or not isinstance(video_url, compat_str):
 254                     continue
 255                 format_id = source.get('label')
 256                 f = {
 257                     'url': video_url,
 258                     'format_id': '%sp' % format_id,
 259                     'height': int_or_none(format_id),
 260                 }
 261                 if format_id:
 262                     # Some videos contain additional metadata (e.g.
 263                     # https://www.udemy.com/ios9-swift/learn/#/lecture/3383208)
 264                     f = add_output_format_meta(f, format_id)
 265                 formats.append(f)
 266
 267         download_urls = asset.get('download_urls')
 268         if isinstance(download_urls, dict):
 269             extract_formats(download_urls.get('Video'))
 270
 271         view_html = lecture.get('view_html')
 272         if view_html:
 273             view_html_urls = set()
 274             for source in re.findall(r'<source[^>]+>', view_html):
 275                 attributes = extract_attributes(source)
 276                 src = attributes.get('src')
 277                 if not src:
 278                     continue
 279                 res = attributes.get('data-res')
 280                 height = int_or_none(res)
 281                 if src in view_html_urls:
 282                     continue
 283                 view_html_urls.add(src)
 284                 if attributes.get('type') == 'application/x-mpegURL' or determine_ext(src) == 'm3u8':
 285                     m3u8_formats = self._extract_m3u8_formats(
 286                         src, video_id, 'mp4', entry_protocol='m3u8_native',
 287                         m3u8_id='hls', fatal=False)
 288                     for f in m3u8_formats:
 289                         m = re.search(r'/hls_(?P<height>\d{3,4})_(?P<tbr>\d{2,})/', f['url'])
 290                         if m:
 291                             if not f.get('height'):
 292                                 f['height'] = int(m.group('height'))
 293                             if not f.get('tbr'):
 294                                 f['tbr'] = int(m.group('tbr'))
 295                     formats.extend(m3u8_formats)
 296                 else:
 297                     formats.append(add_output_format_meta({
 298                         'url': src,
 299                         'format_id': '%dp' % height if height else None,
 300                         'height': height,
 301                     }, res))
 302
 303             # react rendition since 2017.04.15 (see
 304             # https://github.com/rg3/youtube-dl/issues/12744)
 305             data = self._parse_json(
 306                 self._search_regex(
 307                     r'videojs-setup-data=(["\'])(?P<data>{.+?})\1', view_html,
 308                     'setup data', default='{}', group='data'), video_id,
 309                 transform_source=unescapeHTML, fatal=False)
 310             if data and isinstance(data, dict):
 311                 extract_formats(data.get('sources'))
 312                 if not duration:
 313                     duration = int_or_none(data.get('duration'))
 314                 tracks = data.get('tracks')
 315                 if isinstance(tracks, list):
 316                     for track in tracks:
 317                         if not isinstance(track, dict):
 318                             continue
 319                         if track.get('kind') != 'captions':
 320                             continue
 321                         src = track.get('src')
 322                         if not src or not isinstance(src, compat_str):
 323                             continue
 324                         lang = track.get('language') or track.get(
 325                             'srclang') or track.get('label')
 326                         sub_dict = automatic_captions if track.get(
 327                             'autogenerated') is True else subtitles
 328                         sub_dict.setdefault(lang, []).append({
 329                             'url': src,
 330                         })
 331
 332         self._sort_formats(formats, field_preference=('height', 'width', 'tbr', 'format_id'))
 333
 334         return {
 335             'id': video_id,
 336             'title': title,
 337             'description': description,
 338             'thumbnail': thumbnail,
 339             'duration': duration,
 340             'formats': formats,
 341             'subtitles': subtitles,
 342             'automatic_captions': automatic_captions,
 343         }
 344
 345
 346 class UdemyCourseIE(UdemyIE):
 347     IE_NAME = 'udemy:course'
 348     _VALID_URL = r'https?://(?:www\.)?udemy\.com/(?P<id>[^/?#&]+)'
 349     _TESTS = []
 350
 351     @classmethod
 352     def suitable(cls, url):
 353         return False if UdemyIE.suitable(url) else super(UdemyCourseIE, cls).suitable(url)
 354
 355     def _real_extract(self, url):
 356         course_path = self._match_id(url)
 357
 358         webpage = self._download_webpage(url, course_path)
 359
 360         course_id, title = self._extract_course_info(webpage, course_path)
 361
 362         self._enroll_course(url, webpage, course_id)
 363
 364         response = self._download_json(
 365             'https://www.udemy.com/api-2.0/courses/%s/cached-subscriber-curriculum-items' % course_id,
 366             course_id, 'Downloading course curriculum', query={
 367                 'fields[chapter]': 'title,object_index',
 368                 'fields[lecture]': 'title,asset',
 369                 'page_size': '1000',
 370             })
 371
 372         entries = []
 373         chapter, chapter_number = [None] * 2
 374         for entry in response['results']:
 375             clazz = entry.get('_class')
 376             if clazz == 'lecture':
 377                 asset = entry.get('asset')
 378                 if isinstance(asset, dict):
 379                     asset_type = asset.get('asset_type') or asset.get('assetType')
 380                     if asset_type != 'Video':
 381                         continue
 382                 lecture_id = entry.get('id')
 383                 if lecture_id:
 384                     entry = {
 385                         '_type': 'url_transparent',
 386                         'url': 'https://www.udemy.com/%s/learn/v4/t/lecture/%s' % (course_path, entry['id']),
 387                         'title': entry.get('title'),
 388                         'ie_key': UdemyIE.ie_key(),
 389                     }
 390                     if chapter_number:
 391                         entry['chapter_number'] = chapter_number
 392                     if chapter:
 393                         entry['chapter'] = chapter
 394                     entries.append(entry)
 395             elif clazz == 'chapter':
 396                 chapter_number = entry.get('object_index')
 397                 chapter = entry.get('title')
 398
 399         return self.playlist_result(entries, course_id, title)