_ Git - youtube-dl/blob - youtube_dl/extractor/youtube.py

   1 # coding: utf-8
   2
   3 import json
   4 import netrc
   5 import re
   6 import socket
   7 import itertools
   8
   9 from .common import InfoExtractor, SearchInfoExtractor
  10 from ..utils import (
  11     compat_http_client,
  12     compat_parse_qs,
  13     compat_urllib_error,
  14     compat_urllib_parse,
  15     compat_urllib_request,
  16     compat_str,
  17
  18     clean_html,
  19     get_element_by_id,
  20     ExtractorError,
  21     unescapeHTML,
  22     unified_strdate,
  23     orderedSet,
  24 )
  25
  26 class YoutubeBaseInfoExtractor(InfoExtractor):
  27     """Provide base functions for Youtube extractors"""
  28     _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
  29     _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1'
  30     _AGE_URL = 'http://www.youtube.com/verify_age?next_url=/&gl=US&hl=en'
  31     _NETRC_MACHINE = 'youtube'
  32     # If True it will raise an error if no login info is provided
  33     _LOGIN_REQUIRED = False
  34
  35     def report_lang(self):
  36         """Report attempt to set language."""
  37         self.to_screen(u'Setting language')
  38
  39     def _set_language(self):
  40         request = compat_urllib_request.Request(self._LANG_URL)
  41         try:
  42             self.report_lang()
  43             compat_urllib_request.urlopen(request).read()
  44         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
  45             self._downloader.report_warning(u'unable to set language: %s' % compat_str(err))
  46             return False
  47         return True
  48
  49     def _login(self):
  50         (username, password) = self._get_login_info()
  51         # No authentication to be performed
  52         if username is None:
  53             if self._LOGIN_REQUIRED:
  54                 raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True)
  55             return False
  56
  57         request = compat_urllib_request.Request(self._LOGIN_URL)
  58         try:
  59             login_page = compat_urllib_request.urlopen(request).read().decode('utf-8')
  60         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
  61             self._downloader.report_warning(u'unable to fetch login page: %s' % compat_str(err))
  62             return False
  63
  64         galx = None
  65         dsh = None
  66         match = re.search(re.compile(r'<input.+?name="GALX".+?value="(.+?)"', re.DOTALL), login_page)
  67         if match:
  68           galx = match.group(1)
  69         match = re.search(re.compile(r'<input.+?name="dsh".+?value="(.+?)"', re.DOTALL), login_page)
  70         if match:
  71           dsh = match.group(1)
  72
  73         # Log in
  74         login_form_strs = {
  75                 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
  76                 u'Email': username,
  77                 u'GALX': galx,
  78                 u'Passwd': password,
  79                 u'PersistentCookie': u'yes',
  80                 u'_utf8': u'霱',
  81                 u'bgresponse': u'js_disabled',
  82                 u'checkConnection': u'',
  83                 u'checkedDomains': u'youtube',
  84                 u'dnConn': u'',
  85                 u'dsh': dsh,
  86                 u'pstMsg': u'0',
  87                 u'rmShown': u'1',
  88                 u'secTok': u'',
  89                 u'signIn': u'Sign in',
  90                 u'timeStmp': u'',
  91                 u'service': u'youtube',
  92                 u'uilel': u'3',
  93                 u'hl': u'en_US',
  94         }
  95         # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
  96         # chokes on unicode
  97         login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items())
  98         login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
  99         request = compat_urllib_request.Request(self._LOGIN_URL, login_data)
 100         try:
 101             self.report_login()
 102             login_results = compat_urllib_request.urlopen(request).read().decode('utf-8')
 103             if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
 104                 self._downloader.report_warning(u'unable to log in: bad username or password')
 105                 return False
 106         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 107             self._downloader.report_warning(u'unable to log in: %s' % compat_str(err))
 108             return False
 109         return True
 110
 111     def _confirm_age(self):
 112         age_form = {
 113                 'next_url':     '/',
 114                 'action_confirm':   'Confirm',
 115                 }
 116         request = compat_urllib_request.Request(self._AGE_URL, compat_urllib_parse.urlencode(age_form))
 117         try:
 118             self.report_age_confirmation()
 119             compat_urllib_request.urlopen(request).read().decode('utf-8')
 120         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 121             raise ExtractorError(u'Unable to confirm age: %s' % compat_str(err))
 122         return True
 123
 124     def _real_initialize(self):
 125         if self._downloader is None:
 126             return
 127         if not self._set_language():
 128             return
 129         if not self._login():
 130             return
 131         self._confirm_age()
 132
 133 class YoutubeIE(YoutubeBaseInfoExtractor):
 134     IE_DESC = u'YouTube.com'
 135     _VALID_URL = r"""^
 136                      (
 137                          (?:https?://)?                                       # http(s):// (optional)
 138                          (?:youtu\.be/|(?:\w+\.)?youtube(?:-nocookie)?\.com/|
 139                             tube\.majestyc\.net/)                             # the various hostnames, with wildcard subdomains
 140                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
 141                          (?:                                                  # the various things that can precede the ID:
 142                              (?:(?:v|embed|e)/)                               # v/ or embed/ or e/
 143                              |(?:                                             # or the v= param in all its forms
 144                                  (?:watch|movie(?:_popup)?(?:\.php)?)?              # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
 145                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
 146                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
 147                                  v=
 148                              )
 149                          )?                                                   # optional -> youtube.com/xxxx is OK
 150                      )?                                                       # all until now is optional -> you can pass the naked ID
 151                      ([0-9A-Za-z_-]+)                                         # here is it! the YouTube video ID
 152                      (?(1).+)?                                                # if we found the ID, everything can follow
 153                      $"""
 154     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
 155     # Listed in order of quality
 156     _available_formats = ['38', '37', '46', '22', '45', '35', '44', '34', '18', '43', '6', '5', '17', '13']
 157     _available_formats_prefer_free = ['38', '46', '37', '45', '22', '44', '35', '43', '34', '18', '6', '5', '17', '13']
 158     _video_extensions = {
 159         '13': '3gp',
 160         '17': 'mp4',
 161         '18': 'mp4',
 162         '22': 'mp4',
 163         '37': 'mp4',
 164         '38': 'mp4',
 165         '43': 'webm',
 166         '44': 'webm',
 167         '45': 'webm',
 168         '46': 'webm',
 169     }
 170     _video_dimensions = {
 171         '5': '240x400',
 172         '6': '???',
 173         '13': '???',
 174         '17': '144x176',
 175         '18': '360x640',
 176         '22': '720x1280',
 177         '34': '360x640',
 178         '35': '480x854',
 179         '37': '1080x1920',
 180         '38': '3072x4096',
 181         '43': '360x640',
 182         '44': '480x854',
 183         '45': '720x1280',
 184         '46': '1080x1920',
 185     }
 186     IE_NAME = u'youtube'
 187     _TESTS = [
 188         {
 189             u"url":  u"http://www.youtube.com/watch?v=BaW_jenozKc",
 190             u"file":  u"BaW_jenozKc.mp4",
 191             u"info_dict": {
 192                 u"title": u"youtube-dl test video \"'/\\ä↭𝕐",
 193                 u"uploader": u"Philipp Hagemeister",
 194                 u"uploader_id": u"phihag",
 195                 u"upload_date": u"20121002",
 196                 u"description": u"test chars:  \"'/\\ä↭𝕐\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de ."
 197             }
 198         },
 199         {
 200             u"url":  u"http://www.youtube.com/watch?v=1ltcDfZMA3U",
 201             u"file":  u"1ltcDfZMA3U.flv",
 202             u"note": u"Test VEVO video (#897)",
 203             u"info_dict": {
 204                 u"upload_date": u"20070518",
 205                 u"title": u"Maps - It Will Find You",
 206                 u"description": u"Music video by Maps performing It Will Find You.",
 207                 u"uploader": u"MuteUSA",
 208                 u"uploader_id": u"MuteUSA"
 209             }
 210         },
 211         {
 212             u"url":  u"http://www.youtube.com/watch?v=UxxajLWwzqY",
 213             u"file":  u"UxxajLWwzqY.mp4",
 214             u"note": u"Test generic use_cipher_signature video (#897)",
 215             u"info_dict": {
 216                 u"upload_date": u"20120506",
 217                 u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
 218                 u"description": u"md5:b085c9804f5ab69f4adea963a2dceb3c",
 219                 u"uploader": u"IconaPop",
 220                 u"uploader_id": u"IconaPop"
 221             }
 222         },
 223         {
 224             u"url":  u"https://www.youtube.com/watch?v=07FYdnEawAQ",
 225             u"file":  u"07FYdnEawAQ.mp4",
 226             u"note": u"Test VEVO video with age protection (#956)",
 227             u"info_dict": {
 228                 u"upload_date": u"20130703",
 229                 u"title": u"Justin Timberlake - Tunnel Vision (Explicit)",
 230                 u"description": u"md5:64249768eec3bc4276236606ea996373",
 231                 u"uploader": u"justintimberlakeVEVO",
 232                 u"uploader_id": u"justintimberlakeVEVO"
 233             }
 234         },
 235     ]
 236
 237
 238     @classmethod
 239     def suitable(cls, url):
 240         """Receives a URL and returns True if suitable for this IE."""
 241         if YoutubePlaylistIE.suitable(url) or YoutubeSubscriptionsIE.suitable(url): return False
 242         return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
 243
 244     def report_video_webpage_download(self, video_id):
 245         """Report attempt to download video webpage."""
 246         self.to_screen(u'%s: Downloading video webpage' % video_id)
 247
 248     def report_video_info_webpage_download(self, video_id):
 249         """Report attempt to download video info webpage."""
 250         self.to_screen(u'%s: Downloading video info webpage' % video_id)
 251
 252     def report_video_subtitles_download(self, video_id):
 253         """Report attempt to download video info webpage."""
 254         self.to_screen(u'%s: Checking available subtitles' % video_id)
 255
 256     def report_video_subtitles_request(self, video_id, sub_lang, format):
 257         """Report attempt to download video info webpage."""
 258         self.to_screen(u'%s: Downloading video subtitles for %s.%s' % (video_id, sub_lang, format))
 259
 260     def report_video_subtitles_available(self, video_id, sub_lang_list):
 261         """Report available subtitles."""
 262         sub_lang = ",".join(list(sub_lang_list.keys()))
 263         self.to_screen(u'%s: Available subtitles for video: %s' % (video_id, sub_lang))
 264
 265     def report_information_extraction(self, video_id):
 266         """Report attempt to extract video information."""
 267         self.to_screen(u'%s: Extracting video information' % video_id)
 268
 269     def report_unavailable_format(self, video_id, format):
 270         """Report extracted video URL."""
 271         self.to_screen(u'%s: Format %s not available' % (video_id, format))
 272
 273     def report_rtmp_download(self):
 274         """Indicate the download will use the RTMP protocol."""
 275         self.to_screen(u'RTMP download detected')
 276
 277     def _decrypt_signature(self, s):
 278         """Turn the encrypted s field into a working signature"""
 279
 280         if len(s) == 92:
 281             return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83]
 282         elif len(s) == 90:
 283             return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81]
 284         elif len(s) == 88:
 285             return s[48] + s[81:67:-1] + s[82] + s[66:62:-1] + s[85] + s[61:48:-1] + s[67] + s[47:12:-1] + s[3] + s[11:3:-1] + s[2] + s[12]
 286         elif len(s) == 87:
 287             return s[4:23] + s[86] + s[24:85]
 288         elif len(s) == 86:
 289             return s[2:63] + s[82] + s[64:82] + s[63]
 290         elif len(s) == 85:
 291             return s[2:8] + s[0] + s[9:21] + s[65] + s[22:65] + s[84] + s[66:82] + s[21]
 292         elif len(s) == 84:
 293             return s[83:36:-1] + s[2] + s[35:26:-1] + s[3] + s[25:3:-1] + s[26]
 294         elif len(s) == 83:
 295             return s[6] + s[3:6] + s[33] + s[7:24] + s[0] + s[25:33] + s[53] + s[34:53] + s[24] + s[54:]
 296         elif len(s) == 82:
 297             return s[36] + s[79:67:-1] + s[81] + s[66:40:-1] + s[33] + s[39:36:-1] + s[40] + s[35] + s[0] + s[67] + s[32:0:-1] + s[34]
 298         elif len(s) == 81:
 299             return s[56] + s[79:56:-1] + s[41] + s[55:41:-1] + s[80] + s[40:34:-1] + s[0] + s[33:29:-1] + s[34] + s[28:9:-1] + s[29] + s[8:0:-1] + s[9]
 300         elif len(s) == 79:
 301             return s[54] + s[77:54:-1] + s[39] + s[53:39:-1] + s[78] + s[38:34:-1] + s[0] + s[33:29:-1] + s[34] + s[28:9:-1] + s[29] + s[8:0:-1] + s[9]
 302
 303         else:
 304             raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s)))
 305
 306     def _get_available_subtitles(self, video_id):
 307         self.report_video_subtitles_download(video_id)
 308         request = compat_urllib_request.Request('http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id)
 309         try:
 310             sub_list = compat_urllib_request.urlopen(request).read().decode('utf-8')
 311         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 312             return (u'unable to download video subtitles: %s' % compat_str(err), None)
 313         sub_lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
 314         sub_lang_list = dict((l[1], l[0]) for l in sub_lang_list)
 315         if not sub_lang_list:
 316             return (u'video doesn\'t have subtitles', None)
 317         return sub_lang_list
 318
 319     def _list_available_subtitles(self, video_id):
 320         sub_lang_list = self._get_available_subtitles(video_id)
 321         self.report_video_subtitles_available(video_id, sub_lang_list)
 322
 323     def _request_subtitle(self, sub_lang, sub_name, video_id, format):
 324         """
 325         Return tuple:
 326         (error_message, sub_lang, sub)
 327         """
 328         self.report_video_subtitles_request(video_id, sub_lang, format)
 329         params = compat_urllib_parse.urlencode({
 330             'lang': sub_lang,
 331             'name': sub_name,
 332             'v': video_id,
 333             'fmt': format,
 334         })
 335         url = 'http://www.youtube.com/api/timedtext?' + params
 336         try:
 337             sub = compat_urllib_request.urlopen(url).read().decode('utf-8')
 338         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 339             return (u'unable to download video subtitles: %s' % compat_str(err), None, None)
 340         if not sub:
 341             return (u'Did not fetch video subtitles', None, None)
 342         return (None, sub_lang, sub)
 343
 344     def _request_automatic_caption(self, video_id, webpage):
 345         """We need the webpage for getting the captions url, pass it as an
 346            argument to speed up the process."""
 347         sub_lang = self._downloader.params.get('subtitleslang') or 'en'
 348         sub_format = self._downloader.params.get('subtitlesformat')
 349         self.to_screen(u'%s: Looking for automatic captions' % video_id)
 350         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
 351         err_msg = u'Couldn\'t find automatic captions for "%s"' % sub_lang
 352         if mobj is None:
 353             return [(err_msg, None, None)]
 354         player_config = json.loads(mobj.group(1))
 355         try:
 356             args = player_config[u'args']
 357             caption_url = args[u'ttsurl']
 358             timestamp = args[u'timestamp']
 359             params = compat_urllib_parse.urlencode({
 360                 'lang': 'en',
 361                 'tlang': sub_lang,
 362                 'fmt': sub_format,
 363                 'ts': timestamp,
 364                 'kind': 'asr',
 365             })
 366             subtitles_url = caption_url + '&' + params
 367             sub = self._download_webpage(subtitles_url, video_id, u'Downloading automatic captions')
 368             return [(None, sub_lang, sub)]
 369         except KeyError:
 370             return [(err_msg, None, None)]
 371
 372     def _extract_subtitle(self, video_id):
 373         """
 374         Return a list with a tuple:
 375         [(error_message, sub_lang, sub)]
 376         """
 377         sub_lang_list = self._get_available_subtitles(video_id)
 378         sub_format = self._downloader.params.get('subtitlesformat')
 379         if  isinstance(sub_lang_list,tuple): #There was some error, it didn't get the available subtitles
 380             return [(sub_lang_list[0], None, None)]
 381         if self._downloader.params.get('subtitleslang', False):
 382             sub_lang = self._downloader.params.get('subtitleslang')
 383         elif 'en' in sub_lang_list:
 384             sub_lang = 'en'
 385         else:
 386             sub_lang = list(sub_lang_list.keys())[0]
 387         if not sub_lang in sub_lang_list:
 388             return [(u'no closed captions found in the specified language "%s"' % sub_lang, None, None)]
 389
 390         subtitle = self._request_subtitle(sub_lang, sub_lang_list[sub_lang].encode('utf-8'), video_id, sub_format)
 391         return [subtitle]
 392
 393     def _extract_all_subtitles(self, video_id):
 394         sub_lang_list = self._get_available_subtitles(video_id)
 395         sub_format = self._downloader.params.get('subtitlesformat')
 396         if  isinstance(sub_lang_list,tuple): #There was some error, it didn't get the available subtitles
 397             return [(sub_lang_list[0], None, None)]
 398         subtitles = []
 399         for sub_lang in sub_lang_list:
 400             subtitle = self._request_subtitle(sub_lang, sub_lang_list[sub_lang].encode('utf-8'), video_id, sub_format)
 401             subtitles.append(subtitle)
 402         return subtitles
 403
 404     def _print_formats(self, formats):
 405         print('Available formats:')
 406         for x in formats:
 407             print('%s\t:\t%s\t[%s]' %(x, self._video_extensions.get(x, 'flv'), self._video_dimensions.get(x, '???')))
 408
 409     def _extract_id(self, url):
 410         mobj = re.match(self._VALID_URL, url, re.VERBOSE)
 411         if mobj is None:
 412             raise ExtractorError(u'Invalid URL: %s' % url)
 413         video_id = mobj.group(2)
 414         return video_id
 415
 416     def _real_extract(self, url):
 417         if re.match(r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$', url):
 418             self._downloader.report_warning(u'Did you forget to quote the URL? Remember that & is a meta-character in most shells, so you want to put the URL in quotes, like  youtube-dl \'http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc\' (or simply  youtube-dl BaW_jenozKc  ).')
 419
 420         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
 421         mobj = re.search(self._NEXT_URL_RE, url)
 422         if mobj:
 423             url = 'https://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
 424         video_id = self._extract_id(url)
 425
 426         # Get video webpage
 427         self.report_video_webpage_download(video_id)
 428         url = 'https://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
 429         request = compat_urllib_request.Request(url)
 430         try:
 431             video_webpage_bytes = compat_urllib_request.urlopen(request).read()
 432         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 433             raise ExtractorError(u'Unable to download video webpage: %s' % compat_str(err))
 434
 435         video_webpage = video_webpage_bytes.decode('utf-8', 'ignore')
 436
 437         # Attempt to extract SWF player URL
 438         mobj = re.search(r'swfConfig.*?"(http:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
 439         if mobj is not None:
 440             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
 441         else:
 442             player_url = None
 443
 444         # Get video info
 445         self.report_video_info_webpage_download(video_id)
 446         if re.search(r'player-age-gate-content">', video_webpage) is not None:
 447             self.report_age_confirmation()
 448             age_gate = True
 449             # We simulate the access to the video from www.youtube.com/v/{video_id}
 450             # this can be viewed without login into Youtube
 451             data = compat_urllib_parse.urlencode({'video_id': video_id,
 452                                                   'el': 'embedded',
 453                                                   'gl': 'US',
 454                                                   'hl': 'en',
 455                                                   'eurl': 'https://youtube.googleapis.com/v/' + video_id,
 456                                                   'asv': 3,
 457                                                   'sts':'1588',
 458                                                   })
 459             video_info_url = 'https://www.youtube.com/get_video_info?' + data
 460             video_info_webpage = self._download_webpage(video_info_url, video_id,
 461                                     note=False,
 462                                     errnote='unable to download video info webpage')
 463             video_info = compat_parse_qs(video_info_webpage)
 464         else:
 465             age_gate = False
 466             for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
 467                 video_info_url = ('https://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
 468                         % (video_id, el_type))
 469                 video_info_webpage = self._download_webpage(video_info_url, video_id,
 470                                         note=False,
 471                                         errnote='unable to download video info webpage')
 472                 video_info = compat_parse_qs(video_info_webpage)
 473                 if 'token' in video_info:
 474                     break
 475         if 'token' not in video_info:
 476             if 'reason' in video_info:
 477                 raise ExtractorError(u'YouTube said: %s' % video_info['reason'][0], expected=True)
 478             else:
 479                 raise ExtractorError(u'"token" parameter not in video info for unknown reason')
 480
 481         # Check for "rental" videos
 482         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
 483             raise ExtractorError(u'"rental" videos not supported')
 484
 485         # Start extracting information
 486         self.report_information_extraction(video_id)
 487
 488         # uploader
 489         if 'author' not in video_info:
 490             raise ExtractorError(u'Unable to extract uploader name')
 491         video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
 492
 493         # uploader_id
 494         video_uploader_id = None
 495         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
 496         if mobj is not None:
 497             video_uploader_id = mobj.group(1)
 498         else:
 499             self._downloader.report_warning(u'unable to extract uploader nickname')
 500
 501         # title
 502         if 'title' not in video_info:
 503             raise ExtractorError(u'Unable to extract video title')
 504         video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
 505
 506         # thumbnail image
 507         # We try first to get a high quality image:
 508         m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
 509                             video_webpage, re.DOTALL)
 510         if m_thumb is not None:
 511             video_thumbnail = m_thumb.group(1)
 512         elif 'thumbnail_url' not in video_info:
 513             self._downloader.report_warning(u'unable to extract video thumbnail')
 514             video_thumbnail = ''
 515         else:   # don't panic if we can't find it
 516             video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
 517
 518         # upload date
 519         upload_date = None
 520         mobj = re.search(r'id="eow-date.*?>(.*?)</span>', video_webpage, re.DOTALL)
 521         if mobj is not None:
 522             upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
 523             upload_date = unified_strdate(upload_date)
 524
 525         # description
 526         video_description = get_element_by_id("eow-description", video_webpage)
 527         if video_description:
 528             video_description = clean_html(video_description)
 529         else:
 530             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
 531             if fd_mobj:
 532                 video_description = unescapeHTML(fd_mobj.group(1))
 533             else:
 534                 video_description = u''
 535
 536         # subtitles
 537         video_subtitles = None
 538
 539         if self._downloader.params.get('writesubtitles', False):
 540             video_subtitles = self._extract_subtitle(video_id)
 541             if video_subtitles:
 542                 (sub_error, sub_lang, sub) = video_subtitles[0]
 543                 if sub_error:
 544                     self._downloader.report_warning(sub_error)
 545
 546         if self._downloader.params.get('writeautomaticsub', False):
 547             video_subtitles = self._request_automatic_caption(video_id, video_webpage)
 548             (sub_error, sub_lang, sub) = video_subtitles[0]
 549             if sub_error:
 550                 self._downloader.report_warning(sub_error)
 551
 552         if self._downloader.params.get('allsubtitles', False):
 553             video_subtitles = self._extract_all_subtitles(video_id)
 554             for video_subtitle in video_subtitles:
 555                 (sub_error, sub_lang, sub) = video_subtitle
 556                 if sub_error:
 557                     self._downloader.report_warning(sub_error)
 558
 559         if self._downloader.params.get('listsubtitles', False):
 560             self._list_available_subtitles(video_id)
 561             return
 562
 563         if 'length_seconds' not in video_info:
 564             self._downloader.report_warning(u'unable to extract video duration')
 565             video_duration = ''
 566         else:
 567             video_duration = compat_urllib_parse.unquote_plus(video_info['length_seconds'][0])
 568
 569         # Decide which formats to download
 570         req_format = self._downloader.params.get('format', None)
 571
 572         try:
 573             mobj = re.search(r';ytplayer.config = ({.*?});', video_webpage)
 574             if not mobj:
 575                 raise ValueError('Could not find vevo ID')
 576             info = json.loads(mobj.group(1))
 577             args = info['args']
 578             # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
 579             # this signatures are encrypted
 580             m_s = re.search(r'[&,]s=', args['url_encoded_fmt_stream_map'])
 581             if m_s is not None:
 582                 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
 583                 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
 584         except ValueError:
 585             pass
 586
 587         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
 588             self.report_rtmp_download()
 589             video_url_list = [(None, video_info['conn'][0])]
 590         elif 'url_encoded_fmt_stream_map' in video_info and len(video_info['url_encoded_fmt_stream_map']) >= 1:
 591             if 'rtmpe%3Dyes' in video_info['url_encoded_fmt_stream_map'][0]:
 592                 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
 593             url_map = {}
 594             for url_data_str in video_info['url_encoded_fmt_stream_map'][0].split(','):
 595                 url_data = compat_parse_qs(url_data_str)
 596                 if 'itag' in url_data and 'url' in url_data:
 597                     url = url_data['url'][0]
 598                     if 'sig' in url_data:
 599                         url += '&signature=' + url_data['sig'][0]
 600                     elif 's' in url_data:
 601                         if self._downloader.params.get('verbose'):
 602                             s = url_data['s'][0]
 603                             if age_gate:
 604                                 player_version = self._search_regex(r'ad3-(.+?)\.swf',
 605                                     video_info['ad3_module'][0], 'flash player',
 606                                     fatal=False)
 607                                 player = 'flash player %s' % player_version
 608                             else:
 609                                 player = u'html5 player %s' % self._search_regex(r'html5player-(.+?)\.js', video_webpage,
 610                                     'html5 player', fatal=False)
 611                             parts_sizes = u'.'.join(compat_str(len(part)) for part in s.split('.'))
 612                             self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' %
 613                                 (len(s), parts_sizes, url_data['itag'][0], player))
 614                         signature = self._decrypt_signature(url_data['s'][0])
 615                         url += '&signature=' + signature
 616                     if 'ratebypass' not in url:
 617                         url += '&ratebypass=yes'
 618                     url_map[url_data['itag'][0]] = url
 619
 620             format_limit = self._downloader.params.get('format_limit', None)
 621             available_formats = self._available_formats_prefer_free if self._downloader.params.get('prefer_free_formats', False) else self._available_formats
 622             if format_limit is not None and format_limit in available_formats:
 623                 format_list = available_formats[available_formats.index(format_limit):]
 624             else:
 625                 format_list = available_formats
 626             existing_formats = [x for x in format_list if x in url_map]
 627             if len(existing_formats) == 0:
 628                 raise ExtractorError(u'no known formats available for video')
 629             if self._downloader.params.get('listformats', None):
 630                 self._print_formats(existing_formats)
 631                 return
 632             if req_format is None or req_format == 'best':
 633                 video_url_list = [(existing_formats[0], url_map[existing_formats[0]])] # Best quality
 634             elif req_format == 'worst':
 635                 video_url_list = [(existing_formats[-1], url_map[existing_formats[-1]])] # worst quality
 636             elif req_format in ('-1', 'all'):
 637                 video_url_list = [(f, url_map[f]) for f in existing_formats] # All formats
 638             else:
 639                 # Specific formats. We pick the first in a slash-delimeted sequence.
 640                 # For example, if '1/2/3/4' is requested and '2' and '4' are available, we pick '2'.
 641                 req_formats = req_format.split('/')
 642                 video_url_list = None
 643                 for rf in req_formats:
 644                     if rf in url_map:
 645                         video_url_list = [(rf, url_map[rf])]
 646                         break
 647                 if video_url_list is None:
 648                     raise ExtractorError(u'requested format not available')
 649         else:
 650             raise ExtractorError(u'no conn or url_encoded_fmt_stream_map information found in video info')
 651
 652         results = []
 653         for format_param, video_real_url in video_url_list:
 654             # Extension
 655             video_extension = self._video_extensions.get(format_param, 'flv')
 656
 657             video_format = '{0} - {1}'.format(format_param if format_param else video_extension,
 658                                               self._video_dimensions.get(format_param, '???'))
 659
 660             results.append({
 661                 'id':       video_id,
 662                 'url':      video_real_url,
 663                 'uploader': video_uploader,
 664                 'uploader_id': video_uploader_id,
 665                 'upload_date':  upload_date,
 666                 'title':    video_title,
 667                 'ext':      video_extension,
 668                 'format':   video_format,
 669                 'thumbnail':    video_thumbnail,
 670                 'description':  video_description,
 671                 'player_url':   player_url,
 672                 'subtitles':    video_subtitles,
 673                 'duration':     video_duration
 674             })
 675         return results
 676
 677 class YoutubePlaylistIE(InfoExtractor):
 678     IE_DESC = u'YouTube.com playlists'
 679     _VALID_URL = r"""(?:
 680                         (?:https?://)?
 681                         (?:\w+\.)?
 682                         youtube\.com/
 683                         (?:
 684                            (?:course|view_play_list|my_playlists|artist|playlist|watch)
 685                            \? (?:.*?&)*? (?:p|a|list)=
 686                         |  p/
 687                         )
 688                         ((?:PL|EC|UU|FL)?[0-9A-Za-z-_]{10,})
 689                         .*
 690                      |
 691                         ((?:PL|EC|UU|FL)[0-9A-Za-z-_]{10,})
 692                      )"""
 693     _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/playlists/%s?max-results=%i&start-index=%i&v=2&alt=json&safeSearch=none'
 694     _MAX_RESULTS = 50
 695     IE_NAME = u'youtube:playlist'
 696
 697     @classmethod
 698     def suitable(cls, url):
 699         """Receives a URL and returns True if suitable for this IE."""
 700         return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
 701
 702     def _real_extract(self, url):
 703         # Extract playlist id
 704         mobj = re.match(self._VALID_URL, url, re.VERBOSE)
 705         if mobj is None:
 706             raise ExtractorError(u'Invalid URL: %s' % url)
 707
 708         # Download playlist videos from API
 709         playlist_id = mobj.group(1) or mobj.group(2)
 710         videos = []
 711
 712         for page_num in itertools.count(1):
 713             start_index = self._MAX_RESULTS * (page_num - 1) + 1
 714             if start_index >= 1000:
 715                 self._downloader.report_warning(u'Max number of results reached')
 716                 break
 717             url = self._TEMPLATE_URL % (playlist_id, self._MAX_RESULTS, start_index)
 718             page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
 719
 720             try:
 721                 response = json.loads(page)
 722             except ValueError as err:
 723                 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
 724
 725             if 'feed' not in response:
 726                 raise ExtractorError(u'Got a malformed response from YouTube API')
 727             playlist_title = response['feed']['title']['$t']
 728             if 'entry' not in response['feed']:
 729                 # Number of videos is a multiple of self._MAX_RESULTS
 730                 break
 731
 732             for entry in response['feed']['entry']:
 733                 index = entry['yt$position']['$t']
 734                 if 'media$group' in entry and 'media$player' in entry['media$group']:
 735                     videos.append((index, entry['media$group']['media$player']['url']))
 736
 737         videos = [v[1] for v in sorted(videos)]
 738
 739         url_results = [self.url_result(vurl, 'Youtube') for vurl in videos]
 740         return [self.playlist_result(url_results, playlist_id, playlist_title)]
 741
 742
 743 class YoutubeChannelIE(InfoExtractor):
 744     IE_DESC = u'YouTube.com channels'
 745     _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
 746     _TEMPLATE_URL = 'http://www.youtube.com/channel/%s/videos?sort=da&flow=list&view=0&page=%s&gl=US&hl=en'
 747     _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
 748     _MORE_PAGES_URL = 'http://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s'
 749     IE_NAME = u'youtube:channel'
 750
 751     def extract_videos_from_page(self, page):
 752         ids_in_page = []
 753         for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
 754             if mobj.group(1) not in ids_in_page:
 755                 ids_in_page.append(mobj.group(1))
 756         return ids_in_page
 757
 758     def _real_extract(self, url):
 759         # Extract channel id
 760         mobj = re.match(self._VALID_URL, url)
 761         if mobj is None:
 762             raise ExtractorError(u'Invalid URL: %s' % url)
 763
 764         # Download channel page
 765         channel_id = mobj.group(1)
 766         video_ids = []
 767         pagenum = 1
 768
 769         url = self._TEMPLATE_URL % (channel_id, pagenum)
 770         page = self._download_webpage(url, channel_id,
 771                                       u'Downloading page #%s' % pagenum)
 772
 773         # Extract video identifiers
 774         ids_in_page = self.extract_videos_from_page(page)
 775         video_ids.extend(ids_in_page)
 776
 777         # Download any subsequent channel pages using the json-based channel_ajax query
 778         if self._MORE_PAGES_INDICATOR in page:
 779             for pagenum in itertools.count(1):
 780                 url = self._MORE_PAGES_URL % (pagenum, channel_id)
 781                 page = self._download_webpage(url, channel_id,
 782                                               u'Downloading page #%s' % pagenum)
 783
 784                 page = json.loads(page)
 785
 786                 ids_in_page = self.extract_videos_from_page(page['content_html'])
 787                 video_ids.extend(ids_in_page)
 788
 789                 if self._MORE_PAGES_INDICATOR  not in page['load_more_widget_html']:
 790                     break
 791
 792         self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
 793
 794         urls = ['http://www.youtube.com/watch?v=%s' % id for id in video_ids]
 795         url_entries = [self.url_result(eurl, 'Youtube') for eurl in urls]
 796         return [self.playlist_result(url_entries, channel_id)]
 797
 798
 799 class YoutubeUserIE(InfoExtractor):
 800     IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
 801     _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/user/)|ytuser:)([A-Za-z0-9_-]+)'
 802     _TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s'
 803     _GDATA_PAGE_SIZE = 50
 804     _GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d'
 805     _VIDEO_INDICATOR = r'/watch\?v=(.+?)[\<&]'
 806     IE_NAME = u'youtube:user'
 807
 808     def _real_extract(self, url):
 809         # Extract username
 810         mobj = re.match(self._VALID_URL, url)
 811         if mobj is None:
 812             raise ExtractorError(u'Invalid URL: %s' % url)
 813
 814         username = mobj.group(1)
 815
 816         # Download video ids using YouTube Data API. Result size per
 817         # query is limited (currently to 50 videos) so we need to query
 818         # page by page until there are no video ids - it means we got
 819         # all of them.
 820
 821         video_ids = []
 822
 823         for pagenum in itertools.count(0):
 824             start_index = pagenum * self._GDATA_PAGE_SIZE + 1
 825
 826             gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
 827             page = self._download_webpage(gdata_url, username,
 828                                           u'Downloading video ids from %d to %d' % (start_index, start_index + self._GDATA_PAGE_SIZE))
 829
 830             # Extract video identifiers
 831             ids_in_page = []
 832
 833             for mobj in re.finditer(self._VIDEO_INDICATOR, page):
 834                 if mobj.group(1) not in ids_in_page:
 835                     ids_in_page.append(mobj.group(1))
 836
 837             video_ids.extend(ids_in_page)
 838
 839             # A little optimization - if current page is not
 840             # "full", ie. does not contain PAGE_SIZE video ids then
 841             # we can assume that this page is the last one - there
 842             # are no more ids on further pages - no need to query
 843             # again.
 844
 845             if len(ids_in_page) < self._GDATA_PAGE_SIZE:
 846                 break
 847
 848         urls = ['http://www.youtube.com/watch?v=%s' % video_id for video_id in video_ids]
 849         url_results = [self.url_result(rurl, 'Youtube') for rurl in urls]
 850         return [self.playlist_result(url_results, playlist_title = username)]
 851
 852 class YoutubeSearchIE(SearchInfoExtractor):
 853     IE_DESC = u'YouTube.com searches'
 854     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
 855     _MAX_RESULTS = 1000
 856     IE_NAME = u'youtube:search'
 857     _SEARCH_KEY = 'ytsearch'
 858
 859     def report_download_page(self, query, pagenum):
 860         """Report attempt to download search page with given number."""
 861         self._downloader.to_screen(u'[youtube] query "%s": Downloading page %s' % (query, pagenum))
 862
 863     def _get_n_results(self, query, n):
 864         """Get a specified number of results for a query"""
 865
 866         video_ids = []
 867         pagenum = 0
 868         limit = n
 869
 870         while (50 * pagenum) < limit:
 871             self.report_download_page(query, pagenum+1)
 872             result_url = self._API_URL % (compat_urllib_parse.quote_plus(query), (50*pagenum)+1)
 873             request = compat_urllib_request.Request(result_url)
 874             try:
 875                 data = compat_urllib_request.urlopen(request).read().decode('utf-8')
 876             except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 877                 raise ExtractorError(u'Unable to download API page: %s' % compat_str(err))
 878             api_response = json.loads(data)['data']
 879
 880             if not 'items' in api_response:
 881                 raise ExtractorError(u'[youtube] No video results')
 882
 883             new_ids = list(video['id'] for video in api_response['items'])
 884             video_ids += new_ids
 885
 886             limit = min(n, api_response['totalItems'])
 887             pagenum += 1
 888
 889         if len(video_ids) > n:
 890             video_ids = video_ids[:n]
 891         videos = [self.url_result('http://www.youtube.com/watch?v=%s' % id, 'Youtube') for id in video_ids]
 892         return self.playlist_result(videos, query)
 893
 894
 895 class YoutubeShowIE(InfoExtractor):
 896     IE_DESC = u'YouTube.com (multi-season) shows'
 897     _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
 898     IE_NAME = u'youtube:show'
 899
 900     def _real_extract(self, url):
 901         mobj = re.match(self._VALID_URL, url)
 902         show_name = mobj.group(1)
 903         webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
 904         # There's one playlist for each season of the show
 905         m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
 906         self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
 907         return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
 908
 909
 910 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
 911     """
 912     Base class for extractors that fetch info from
 913     http://www.youtube.com/feed_ajax
 914     Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
 915     """
 916     _LOGIN_REQUIRED = True
 917     _PAGING_STEP = 30
 918     # use action_load_personal_feed instead of action_load_system_feed
 919     _PERSONAL_FEED = False
 920
 921     @property
 922     def _FEED_TEMPLATE(self):
 923         action = 'action_load_system_feed'
 924         if self._PERSONAL_FEED:
 925             action = 'action_load_personal_feed'
 926         return 'http://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
 927
 928     @property
 929     def IE_NAME(self):
 930         return u'youtube:%s' % self._FEED_NAME
 931
 932     def _real_initialize(self):
 933         self._login()
 934
 935     def _real_extract(self, url):
 936         feed_entries = []
 937         # The step argument is available only in 2.7 or higher
 938         for i in itertools.count(0):
 939             paging = i*self._PAGING_STEP
 940             info = self._download_webpage(self._FEED_TEMPLATE % paging,
 941                                           u'%s feed' % self._FEED_NAME,
 942                                           u'Downloading page %s' % i)
 943             info = json.loads(info)
 944             feed_html = info['feed_html']
 945             m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
 946             ids = orderedSet(m.group(1) for m in m_ids)
 947             feed_entries.extend(self.url_result(id, 'Youtube') for id in ids)
 948             if info['paging'] is None:
 949                 break
 950         return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
 951
 952 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
 953     IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword(requires authentication)'
 954     _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
 955     _FEED_NAME = 'subscriptions'
 956     _PLAYLIST_TITLE = u'Youtube Subscriptions'
 957
 958 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
 959     IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
 960     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
 961     _FEED_NAME = 'recommended'
 962     _PLAYLIST_TITLE = u'Youtube Recommended videos'
 963
 964 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
 965     IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
 966     _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
 967     _FEED_NAME = 'watch_later'
 968     _PLAYLIST_TITLE = u'Youtube Watch Later'
 969     _PAGING_STEP = 100
 970     _PERSONAL_FEED = True
 971
 972 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
 973     IE_NAME = u'youtube:favorites'
 974     IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
 975     _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:o?rites)?'
 976     _LOGIN_REQUIRED = True
 977
 978     def _real_extract(self, url):
 979         webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
 980         playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
 981         return self.url_result(playlist_id, 'YoutubePlaylist')