youtube_dl/extractor/youtube.py

   1 # coding: utf-8
   2
   3 from __future__ import unicode_literals
   4
   5
   6 import itertools
   7 import json
   8 import os.path
   9 import re
  10 import time
  11 import traceback
  12
  13 from .common import InfoExtractor, SearchInfoExtractor
  14 from ..jsinterp import JSInterpreter
  15 from ..swfinterp import SWFInterpreter
  16 from ..compat import (
  17     compat_chr,
  18     compat_parse_qs,
  19     compat_urllib_parse,
  20     compat_urllib_request,
  21     compat_urlparse,
  22     compat_str,
  23 )
  24 from ..utils import (
  25     clean_html,
  26     ExtractorError,
  27     float_or_none,
  28     get_element_by_attribute,
  29     get_element_by_id,
  30     int_or_none,
  31     orderedSet,
  32     str_to_int,
  33     unescapeHTML,
  34     unified_strdate,
  35     uppercase_escape,
  36     ISO3166Utils,
  37 )
  38
  39
  40 class YoutubeBaseInfoExtractor(InfoExtractor):
  41     """Provide base functions for Youtube extractors"""
  42     _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
  43     _TWOFACTOR_URL = 'https://accounts.google.com/SecondFactor'
  44     _NETRC_MACHINE = 'youtube'
  45     # If True it will raise an error if no login info is provided
  46     _LOGIN_REQUIRED = False
  47
  48     def _set_language(self):
  49         self._set_cookie(
  50             '.youtube.com', 'PREF', 'f1=50000000&hl=en',
  51             # YouTube sets the expire time to about two months
  52             expire_time=time.time() + 2 * 30 * 24 * 3600)
  53
  54     def _ids_to_results(self, ids):
  55         return [
  56             self.url_result(vid_id, 'Youtube', video_id=vid_id)
  57             for vid_id in ids]
  58
  59     def _login(self):
  60         """
  61         Attempt to log in to YouTube.
  62         True is returned if successful or skipped.
  63         False is returned if login failed.
  64
  65         If _LOGIN_REQUIRED is set and no authentication was provided, an error is raised.
  66         """
  67         (username, password) = self._get_login_info()
  68         # No authentication to be performed
  69         if username is None:
  70             if self._LOGIN_REQUIRED:
  71                 raise ExtractorError('No login info available, needed for using %s.' % self.IE_NAME, expected=True)
  72             return True
  73
  74         login_page = self._download_webpage(
  75             self._LOGIN_URL, None,
  76             note='Downloading login page',
  77             errnote='unable to fetch login page', fatal=False)
  78         if login_page is False:
  79             return
  80
  81         galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"',
  82                                   login_page, 'Login GALX parameter')
  83
  84         # Log in
  85         login_form_strs = {
  86             'continue': 'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
  87             'Email': username,
  88             'GALX': galx,
  89             'Passwd': password,
  90
  91             'PersistentCookie': 'yes',
  92             '_utf8': '霱',
  93             'bgresponse': 'js_disabled',
  94             'checkConnection': '',
  95             'checkedDomains': 'youtube',
  96             'dnConn': '',
  97             'pstMsg': '0',
  98             'rmShown': '1',
  99             'secTok': '',
 100             'signIn': 'Sign in',
 101             'timeStmp': '',
 102             'service': 'youtube',
 103             'uilel': '3',
 104             'hl': 'en_US',
 105         }
 106
 107         # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
 108         # chokes on unicode
 109         login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k, v in login_form_strs.items())
 110         login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
 111
 112         req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
 113         login_results = self._download_webpage(
 114             req, None,
 115             note='Logging in', errnote='unable to log in', fatal=False)
 116         if login_results is False:
 117             return False
 118
 119         if re.search(r'id="errormsg_0_Passwd"', login_results) is not None:
 120             raise ExtractorError('Please use your account password and a two-factor code instead of an application-specific password.', expected=True)
 121
 122         # Two-Factor
 123         # TODO add SMS and phone call support - these require making a request and then prompting the user
 124
 125         if re.search(r'(?i)<form[^>]* id="gaia_secondfactorform"', login_results) is not None:
 126             tfa_code = self._get_tfa_info()
 127
 128             if tfa_code is None:
 129                 self._downloader.report_warning('Two-factor authentication required. Provide it with --twofactor <code>')
 130                 self._downloader.report_warning('(Note that only TOTP (Google Authenticator App) codes work at this time.)')
 131                 return False
 132
 133             # Unlike the first login form, secTok and timeStmp are both required for the TFA form
 134
 135             match = re.search(r'id="secTok"\n\s+value=\'(.+)\'/>', login_results, re.M | re.U)
 136             if match is None:
 137                 self._downloader.report_warning('Failed to get secTok - did the page structure change?')
 138             secTok = match.group(1)
 139             match = re.search(r'id="timeStmp"\n\s+value=\'(.+)\'/>', login_results, re.M | re.U)
 140             if match is None:
 141                 self._downloader.report_warning('Failed to get timeStmp - did the page structure change?')
 142             timeStmp = match.group(1)
 143
 144             tfa_form_strs = {
 145                 'continue': 'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
 146                 'smsToken': '',
 147                 'smsUserPin': tfa_code,
 148                 'smsVerifyPin': 'Verify',
 149
 150                 'PersistentCookie': 'yes',
 151                 'checkConnection': '',
 152                 'checkedDomains': 'youtube',
 153                 'pstMsg': '1',
 154                 'secTok': secTok,
 155                 'timeStmp': timeStmp,
 156                 'service': 'youtube',
 157                 'hl': 'en_US',
 158             }
 159             tfa_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k, v in tfa_form_strs.items())
 160             tfa_data = compat_urllib_parse.urlencode(tfa_form).encode('ascii')
 161
 162             tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data)
 163             tfa_results = self._download_webpage(
 164                 tfa_req, None,
 165                 note='Submitting TFA code', errnote='unable to submit tfa', fatal=False)
 166
 167             if tfa_results is False:
 168                 return False
 169
 170             if re.search(r'(?i)<form[^>]* id="gaia_secondfactorform"', tfa_results) is not None:
 171                 self._downloader.report_warning('Two-factor code expired. Please try again, or use a one-use backup code instead.')
 172                 return False
 173             if re.search(r'(?i)<form[^>]* id="gaia_loginform"', tfa_results) is not None:
 174                 self._downloader.report_warning('unable to log in - did the page structure change?')
 175                 return False
 176             if re.search(r'smsauth-interstitial-reviewsettings', tfa_results) is not None:
 177                 self._downloader.report_warning('Your Google account has a security notice. Please log in on your web browser, resolve the notice, and try again.')
 178                 return False
 179
 180         if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
 181             self._downloader.report_warning('unable to log in: bad username or password')
 182             return False
 183         return True
 184
 185     def _real_initialize(self):
 186         if self._downloader is None:
 187             return
 188         self._set_language()
 189         if not self._login():
 190             return
 191
 192
 193 class YoutubeIE(YoutubeBaseInfoExtractor):
 194     IE_DESC = 'YouTube.com'
 195     _VALID_URL = r"""(?x)^
 196                      (
 197                          (?:https?://|//)                                    # http(s):// or protocol-independent URL
 198                          (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/|
 199                             (?:www\.)?deturl\.com/www\.youtube\.com/|
 200                             (?:www\.)?pwnyoutube\.com/|
 201                             (?:www\.)?yourepeat\.com/|
 202                             tube\.majestyc\.net/|
 203                             youtube\.googleapis\.com/)                        # the various hostnames, with wildcard subdomains
 204                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
 205                          (?:                                                  # the various things that can precede the ID:
 206                              (?:(?:v|embed|e)/(?!videoseries))                # v/ or embed/ or e/
 207                              |(?:                                             # or the v= param in all its forms
 208                                  (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)?  # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
 209                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
 210                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
 211                                  v=
 212                              )
 213                          ))
 214                          |youtu\.be/                                          # just youtu.be/xxxx
 215                          |(?:www\.)?cleanvideosearch\.com/media/action/yt/watch\?videoId=
 216                          )
 217                      )?                                                       # all until now is optional -> you can pass the naked ID
 218                      ([0-9A-Za-z_-]{11})                                      # here is it! the YouTube video ID
 219                      (?!.*?&list=)                                            # combined list/video URLs are handled by the playlist IE
 220                      (?(1).+)?                                                # if we found the ID, everything can follow
 221                      $"""
 222     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
 223     _formats = {
 224         '5': {'ext': 'flv', 'width': 400, 'height': 240},
 225         '6': {'ext': 'flv', 'width': 450, 'height': 270},
 226         '13': {'ext': '3gp'},
 227         '17': {'ext': '3gp', 'width': 176, 'height': 144},
 228         '18': {'ext': 'mp4', 'width': 640, 'height': 360},
 229         '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
 230         '34': {'ext': 'flv', 'width': 640, 'height': 360},
 231         '35': {'ext': 'flv', 'width': 854, 'height': 480},
 232         '36': {'ext': '3gp', 'width': 320, 'height': 240},
 233         '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
 234         '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
 235         '43': {'ext': 'webm', 'width': 640, 'height': 360},
 236         '44': {'ext': 'webm', 'width': 854, 'height': 480},
 237         '45': {'ext': 'webm', 'width': 1280, 'height': 720},
 238         '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
 239         '59': {'ext': 'mp4', 'width': 854, 'height': 480},
 240         '78': {'ext': 'mp4', 'width': 854, 'height': 480},
 241
 242
 243         # 3d videos
 244         '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20},
 245         '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20},
 246         '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20},
 247         '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20},
 248         '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20},
 249         '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20},
 250         '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20},
 251
 252         # Apple HTTP Live Streaming
 253         '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 254         '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10},
 255         '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10},
 256         '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10},
 257         '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10},
 258         '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 259         '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10},
 260
 261         # DASH mp4 video
 262         '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 263         '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 264         '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 265         '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 266         '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 267         '138': {'ext': 'mp4', 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},  # Height can vary (https://github.com/rg3/youtube-dl/issues/4559)
 268         '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 269         '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 270         '298': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'},
 271         '299': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'},
 272         '266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'h264'},
 273
 274         # Dash mp4 audio
 275         '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 48, 'preference': -50, 'container': 'm4a_dash'},
 276         '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 128, 'preference': -50, 'container': 'm4a_dash'},
 277         '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 256, 'preference': -50, 'container': 'm4a_dash'},
 278
 279         # Dash webm
 280         '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 281         '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 282         '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 283         '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 284         '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 285         '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 286         '278': {'ext': 'webm', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'container': 'webm', 'vcodec': 'VP9'},
 287         '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 288         '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 289         '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 290         '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 291         '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 292         '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 293         '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 294         '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 295         '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 296         '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 297         '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 298         '308': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 299         '313': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'VP9'},
 300         '315': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 301
 302         # Dash webm audio
 303         '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50},
 304         '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50},
 305
 306         # Dash webm audio with opus inside
 307         '249': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 50, 'preference': -50},
 308         '250': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 70, 'preference': -50},
 309         '251': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 160, 'preference': -50},
 310
 311         # RTMP (unnamed)
 312         '_rtmp': {'protocol': 'rtmp'},
 313     }
 314
 315     IE_NAME = 'youtube'
 316     _TESTS = [
 317         {
 318             'url': 'http://www.youtube.com/watch?v=BaW_jenozKc',
 319             'info_dict': {
 320                 'id': 'BaW_jenozKc',
 321                 'ext': 'mp4',
 322                 'title': 'youtube-dl test video "\'/\\ä↭𝕐',
 323                 'uploader': 'Philipp Hagemeister',
 324                 'uploader_id': 'phihag',
 325                 'upload_date': '20121002',
 326                 'description': 'test chars:  "\'/\\ä↭𝕐\ntest URL: https://github.com/rg3/youtube-dl/issues/1892\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de .',
 327                 'categories': ['Science & Technology'],
 328                 'like_count': int,
 329                 'dislike_count': int,
 330             }
 331         },
 332         {
 333             'url': 'http://www.youtube.com/watch?v=UxxajLWwzqY',
 334             'note': 'Test generic use_cipher_signature video (#897)',
 335             'info_dict': {
 336                 'id': 'UxxajLWwzqY',
 337                 'ext': 'mp4',
 338                 'upload_date': '20120506',
 339                 'title': 'Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]',
 340                 'description': 'md5:fea86fda2d5a5784273df5c7cc994d9f',
 341                 'uploader': 'Icona Pop',
 342                 'uploader_id': 'IconaPop',
 343             }
 344         },
 345         {
 346             'url': 'https://www.youtube.com/watch?v=07FYdnEawAQ',
 347             'note': 'Test VEVO video with age protection (#956)',
 348             'info_dict': {
 349                 'id': '07FYdnEawAQ',
 350                 'ext': 'mp4',
 351                 'upload_date': '20130703',
 352                 'title': 'Justin Timberlake - Tunnel Vision (Explicit)',
 353                 'description': 'md5:64249768eec3bc4276236606ea996373',
 354                 'uploader': 'justintimberlakeVEVO',
 355                 'uploader_id': 'justintimberlakeVEVO',
 356             }
 357         },
 358         {
 359             'url': '//www.YouTube.com/watch?v=yZIXLfi8CZQ',
 360             'note': 'Embed-only video (#1746)',
 361             'info_dict': {
 362                 'id': 'yZIXLfi8CZQ',
 363                 'ext': 'mp4',
 364                 'upload_date': '20120608',
 365                 'title': 'Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012',
 366                 'description': 'md5:09b78bd971f1e3e289601dfba15ca4f7',
 367                 'uploader': 'SET India',
 368                 'uploader_id': 'setindia'
 369             }
 370         },
 371         {
 372             'url': 'http://www.youtube.com/watch?v=a9LDPn-MO4I',
 373             'note': '256k DASH audio (format 141) via DASH manifest',
 374             'info_dict': {
 375                 'id': 'a9LDPn-MO4I',
 376                 'ext': 'm4a',
 377                 'upload_date': '20121002',
 378                 'uploader_id': '8KVIDEO',
 379                 'description': '',
 380                 'uploader': '8KVIDEO',
 381                 'title': 'UHDTV TEST 8K VIDEO.mp4'
 382             },
 383             'params': {
 384                 'youtube_include_dash_manifest': True,
 385                 'format': '141',
 386             },
 387         },
 388         # DASH manifest with encrypted signature
 389         {
 390             'url': 'https://www.youtube.com/watch?v=IB3lcPjvWLA',
 391             'info_dict': {
 392                 'id': 'IB3lcPjvWLA',
 393                 'ext': 'm4a',
 394                 'title': 'Afrojack, Spree Wilson - The Spark ft. Spree Wilson',
 395                 'description': 'md5:12e7067fa6735a77bdcbb58cb1187d2d',
 396                 'uploader': 'AfrojackVEVO',
 397                 'uploader_id': 'AfrojackVEVO',
 398                 'upload_date': '20131011',
 399             },
 400             'params': {
 401                 'youtube_include_dash_manifest': True,
 402                 'format': '141',
 403             },
 404         },
 405         # JS player signature function name containing $
 406         {
 407             'url': 'https://www.youtube.com/watch?v=nfWlot6h_JM',
 408             'info_dict': {
 409                 'id': 'nfWlot6h_JM',
 410                 'ext': 'm4a',
 411                 'title': 'Taylor Swift - Shake It Off',
 412                 'description': 'md5:2acfda1b285bdd478ccec22f9918199d',
 413                 'uploader': 'TaylorSwiftVEVO',
 414                 'uploader_id': 'TaylorSwiftVEVO',
 415                 'upload_date': '20140818',
 416             },
 417             'params': {
 418                 'youtube_include_dash_manifest': True,
 419                 'format': '141',
 420             },
 421         },
 422         # Controversy video
 423         {
 424             'url': 'https://www.youtube.com/watch?v=T4XJQO3qol8',
 425             'info_dict': {
 426                 'id': 'T4XJQO3qol8',
 427                 'ext': 'mp4',
 428                 'upload_date': '20100909',
 429                 'uploader': 'The Amazing Atheist',
 430                 'uploader_id': 'TheAmazingAtheist',
 431                 'title': 'Burning Everyone\'s Koran',
 432                 'description': 'SUBSCRIBE: http://www.youtube.com/saturninefilms\n\nEven Obama has taken a stand against freedom on this issue: http://www.huffingtonpost.com/2010/09/09/obama-gma-interview-quran_n_710282.html',
 433             }
 434         },
 435         # Normal age-gate video (No vevo, embed allowed)
 436         {
 437             'url': 'http://youtube.com/watch?v=HtVdAasjOgU',
 438             'info_dict': {
 439                 'id': 'HtVdAasjOgU',
 440                 'ext': 'mp4',
 441                 'title': 'The Witcher 3: Wild Hunt - The Sword Of Destiny Trailer',
 442                 'description': 're:(?s).{100,}About the Game\n.*?The Witcher 3: Wild Hunt.{100,}',
 443                 'uploader': 'The Witcher',
 444                 'uploader_id': 'WitcherGame',
 445                 'upload_date': '20140605',
 446             },
 447         },
 448         # Age-gate video with encrypted signature
 449         {
 450             'url': 'http://www.youtube.com/watch?v=6kLq3WMV1nU',
 451             'info_dict': {
 452                 'id': '6kLq3WMV1nU',
 453                 'ext': 'mp4',
 454                 'title': 'Dedication To My Ex (Miss That) (Lyric Video)',
 455                 'description': 'md5:33765bb339e1b47e7e72b5490139bb41',
 456                 'uploader': 'LloydVEVO',
 457                 'uploader_id': 'LloydVEVO',
 458                 'upload_date': '20110629',
 459             },
 460         },
 461         # video_info is None (https://github.com/rg3/youtube-dl/issues/4421)
 462         {
 463             'url': '__2ABJjxzNo',
 464             'info_dict': {
 465                 'id': '__2ABJjxzNo',
 466                 'ext': 'mp4',
 467                 'upload_date': '20100430',
 468                 'uploader_id': 'deadmau5',
 469                 'description': 'md5:12c56784b8032162bb936a5f76d55360',
 470                 'uploader': 'deadmau5',
 471                 'title': 'Deadmau5 - Some Chords (HD)',
 472             },
 473             'expected_warnings': [
 474                 'DASH manifest missing',
 475             ]
 476         },
 477         # Olympics (https://github.com/rg3/youtube-dl/issues/4431)
 478         {
 479             'url': 'lqQg6PlCWgI',
 480             'info_dict': {
 481                 'id': 'lqQg6PlCWgI',
 482                 'ext': 'mp4',
 483                 'upload_date': '20120731',
 484                 'uploader_id': 'olympic',
 485                 'description': 'HO09  - Women -  GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games',
 486                 'uploader': 'Olympics',
 487                 'title': 'Hockey - Women -  GER-AUS - London 2012 Olympic Games',
 488             },
 489             'params': {
 490                 'skip_download': 'requires avconv',
 491             }
 492         },
 493         # Non-square pixels
 494         {
 495             'url': 'https://www.youtube.com/watch?v=_b-2C3KPAM0',
 496             'info_dict': {
 497                 'id': '_b-2C3KPAM0',
 498                 'ext': 'mp4',
 499                 'stretched_ratio': 16 / 9.,
 500                 'upload_date': '20110310',
 501                 'uploader_id': 'AllenMeow',
 502                 'description': 'made by Wacom from Korea | 字幕&加油添醋 by TY\'s Allen | 感謝heylisa00cavey1001同學熱情提供梗及翻譯',
 503                 'uploader': '孫艾倫',
 504                 'title': '[A-made] 變態妍字幕版 太妍 我就是這樣的人',
 505             },
 506         },
 507         # url_encoded_fmt_stream_map is empty string
 508         {
 509             'url': 'qEJwOuvDf7I',
 510             'info_dict': {
 511                 'id': 'qEJwOuvDf7I',
 512                 'ext': 'mp4',
 513                 'title': 'Обсуждение судебной практики по выборам 14 сентября 2014 года в Санкт-Петербурге',
 514                 'description': '',
 515                 'upload_date': '20150404',
 516                 'uploader_id': 'spbelect',
 517                 'uploader': 'Наблюдатели Петербурга',
 518             },
 519             'params': {
 520                 'skip_download': 'requires avconv',
 521             }
 522         },
 523         # Extraction from multiple DASH manifests (https://github.com/rg3/youtube-dl/pull/6097)
 524         {
 525             'url': 'https://www.youtube.com/watch?v=FIl7x6_3R5Y',
 526             'info_dict': {
 527                 'id': 'FIl7x6_3R5Y',
 528                 'ext': 'mp4',
 529                 'title': 'md5:7b81415841e02ecd4313668cde88737a',
 530                 'description': 'md5:116377fd2963b81ec4ce64b542173306',
 531                 'upload_date': '20150625',
 532                 'uploader_id': 'dorappi2000',
 533                 'uploader': 'dorappi2000',
 534                 'formats': 'mincount:33',
 535             },
 536         }
 537     ]
 538
 539     def __init__(self, *args, **kwargs):
 540         super(YoutubeIE, self).__init__(*args, **kwargs)
 541         self._player_cache = {}
 542
 543     def report_video_info_webpage_download(self, video_id):
 544         """Report attempt to download video info webpage."""
 545         self.to_screen('%s: Downloading video info webpage' % video_id)
 546
 547     def report_information_extraction(self, video_id):
 548         """Report attempt to extract video information."""
 549         self.to_screen('%s: Extracting video information' % video_id)
 550
 551     def report_unavailable_format(self, video_id, format):
 552         """Report extracted video URL."""
 553         self.to_screen('%s: Format %s not available' % (video_id, format))
 554
 555     def report_rtmp_download(self):
 556         """Indicate the download will use the RTMP protocol."""
 557         self.to_screen('RTMP download detected')
 558
 559     def _signature_cache_id(self, example_sig):
 560         """ Return a string representation of a signature """
 561         return '.'.join(compat_str(len(part)) for part in example_sig.split('.'))
 562
 563     def _extract_signature_function(self, video_id, player_url, example_sig):
 564         id_m = re.match(
 565             r'.*?-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.(?P<ext>[a-z]+)$',
 566             player_url)
 567         if not id_m:
 568             raise ExtractorError('Cannot identify player %r' % player_url)
 569         player_type = id_m.group('ext')
 570         player_id = id_m.group('id')
 571
 572         # Read from filesystem cache
 573         func_id = '%s_%s_%s' % (
 574             player_type, player_id, self._signature_cache_id(example_sig))
 575         assert os.path.basename(func_id) == func_id
 576
 577         cache_spec = self._downloader.cache.load('youtube-sigfuncs', func_id)
 578         if cache_spec is not None:
 579             return lambda s: ''.join(s[i] for i in cache_spec)
 580
 581         download_note = (
 582             'Downloading player %s' % player_url
 583             if self._downloader.params.get('verbose') else
 584             'Downloading %s player %s' % (player_type, player_id)
 585         )
 586         if player_type == 'js':
 587             code = self._download_webpage(
 588                 player_url, video_id,
 589                 note=download_note,
 590                 errnote='Download of %s failed' % player_url)
 591             res = self._parse_sig_js(code)
 592         elif player_type == 'swf':
 593             urlh = self._request_webpage(
 594                 player_url, video_id,
 595                 note=download_note,
 596                 errnote='Download of %s failed' % player_url)
 597             code = urlh.read()
 598             res = self._parse_sig_swf(code)
 599         else:
 600             assert False, 'Invalid player type %r' % player_type
 601
 602         test_string = ''.join(map(compat_chr, range(len(example_sig))))
 603         cache_res = res(test_string)
 604         cache_spec = [ord(c) for c in cache_res]
 605
 606         self._downloader.cache.store('youtube-sigfuncs', func_id, cache_spec)
 607         return res
 608
 609     def _print_sig_code(self, func, example_sig):
 610         def gen_sig_code(idxs):
 611             def _genslice(start, end, step):
 612                 starts = '' if start == 0 else str(start)
 613                 ends = (':%d' % (end + step)) if end + step >= 0 else ':'
 614                 steps = '' if step == 1 else (':%d' % step)
 615                 return 's[%s%s%s]' % (starts, ends, steps)
 616
 617             step = None
 618             # Quelch pyflakes warnings - start will be set when step is set
 619             start = '(Never used)'
 620             for i, prev in zip(idxs[1:], idxs[:-1]):
 621                 if step is not None:
 622                     if i - prev == step:
 623                         continue
 624                     yield _genslice(start, prev, step)
 625                     step = None
 626                     continue
 627                 if i - prev in [-1, 1]:
 628                     step = i - prev
 629                     start = prev
 630                     continue
 631                 else:
 632                     yield 's[%d]' % prev
 633             if step is None:
 634                 yield 's[%d]' % i
 635             else:
 636                 yield _genslice(start, i, step)
 637
 638         test_string = ''.join(map(compat_chr, range(len(example_sig))))
 639         cache_res = func(test_string)
 640         cache_spec = [ord(c) for c in cache_res]
 641         expr_code = ' + '.join(gen_sig_code(cache_spec))
 642         signature_id_tuple = '(%s)' % (
 643             ', '.join(compat_str(len(p)) for p in example_sig.split('.')))
 644         code = ('if tuple(len(p) for p in s.split(\'.\')) == %s:\n'
 645                 '    return %s\n') % (signature_id_tuple, expr_code)
 646         self.to_screen('Extracted signature function:\n' + code)
 647
 648     def _parse_sig_js(self, jscode):
 649         funcname = self._search_regex(
 650             r'\.sig\|\|([a-zA-Z0-9$]+)\(', jscode,
 651             'Initial JS player signature function name')
 652
 653         jsi = JSInterpreter(jscode)
 654         initial_function = jsi.extract_function(funcname)
 655         return lambda s: initial_function([s])
 656
 657     def _parse_sig_swf(self, file_contents):
 658         swfi = SWFInterpreter(file_contents)
 659         TARGET_CLASSNAME = 'SignatureDecipher'
 660         searched_class = swfi.extract_class(TARGET_CLASSNAME)
 661         initial_function = swfi.extract_function(searched_class, 'decipher')
 662         return lambda s: initial_function([s])
 663
 664     def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
 665         """Turn the encrypted s field into a working signature"""
 666
 667         if player_url is None:
 668             raise ExtractorError('Cannot decrypt signature without player_url')
 669
 670         if player_url.startswith('//'):
 671             player_url = 'https:' + player_url
 672         try:
 673             player_id = (player_url, self._signature_cache_id(s))
 674             if player_id not in self._player_cache:
 675                 func = self._extract_signature_function(
 676                     video_id, player_url, s
 677                 )
 678                 self._player_cache[player_id] = func
 679             func = self._player_cache[player_id]
 680             if self._downloader.params.get('youtube_print_sig_code'):
 681                 self._print_sig_code(func, s)
 682             return func(s)
 683         except Exception as e:
 684             tb = traceback.format_exc()
 685             raise ExtractorError(
 686                 'Signature extraction failed: ' + tb, cause=e)
 687
 688     def _get_subtitles(self, video_id, webpage):
 689         try:
 690             subs_doc = self._download_xml(
 691                 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
 692                 video_id, note=False)
 693         except ExtractorError as err:
 694             self._downloader.report_warning('unable to download video subtitles: %s' % compat_str(err))
 695             return {}
 696
 697         sub_lang_list = {}
 698         for track in subs_doc.findall('track'):
 699             lang = track.attrib['lang_code']
 700             if lang in sub_lang_list:
 701                 continue
 702             sub_formats = []
 703             for ext in ['sbv', 'vtt', 'srt']:
 704                 params = compat_urllib_parse.urlencode({
 705                     'lang': lang,
 706                     'v': video_id,
 707                     'fmt': ext,
 708                     'name': track.attrib['name'].encode('utf-8'),
 709                 })
 710                 sub_formats.append({
 711                     'url': 'https://www.youtube.com/api/timedtext?' + params,
 712                     'ext': ext,
 713                 })
 714             sub_lang_list[lang] = sub_formats
 715         if not sub_lang_list:
 716             self._downloader.report_warning('video doesn\'t have subtitles')
 717             return {}
 718         return sub_lang_list
 719
 720     def _get_automatic_captions(self, video_id, webpage):
 721         """We need the webpage for getting the captions url, pass it as an
 722            argument to speed up the process."""
 723         self.to_screen('%s: Looking for automatic captions' % video_id)
 724         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
 725         err_msg = 'Couldn\'t find automatic captions for %s' % video_id
 726         if mobj is None:
 727             self._downloader.report_warning(err_msg)
 728             return {}
 729         player_config = json.loads(mobj.group(1))
 730         try:
 731             args = player_config['args']
 732             caption_url = args['ttsurl']
 733             timestamp = args['timestamp']
 734             # We get the available subtitles
 735             list_params = compat_urllib_parse.urlencode({
 736                 'type': 'list',
 737                 'tlangs': 1,
 738                 'asrs': 1,
 739             })
 740             list_url = caption_url + '&' + list_params
 741             caption_list = self._download_xml(list_url, video_id)
 742             original_lang_node = caption_list.find('track')
 743             if original_lang_node is None:
 744                 self._downloader.report_warning('Video doesn\'t have automatic captions')
 745                 return {}
 746             original_lang = original_lang_node.attrib['lang_code']
 747             caption_kind = original_lang_node.attrib.get('kind', '')
 748
 749             sub_lang_list = {}
 750             for lang_node in caption_list.findall('target'):
 751                 sub_lang = lang_node.attrib['lang_code']
 752                 sub_formats = []
 753                 for ext in ['sbv', 'vtt', 'srt']:
 754                     params = compat_urllib_parse.urlencode({
 755                         'lang': original_lang,
 756                         'tlang': sub_lang,
 757                         'fmt': ext,
 758                         'ts': timestamp,
 759                         'kind': caption_kind,
 760                     })
 761                     sub_formats.append({
 762                         'url': caption_url + '&' + params,
 763                         'ext': ext,
 764                     })
 765                 sub_lang_list[sub_lang] = sub_formats
 766             return sub_lang_list
 767         # An extractor error can be raise by the download process if there are
 768         # no automatic captions but there are subtitles
 769         except (KeyError, ExtractorError):
 770             self._downloader.report_warning(err_msg)
 771             return {}
 772
 773     @classmethod
 774     def extract_id(cls, url):
 775         mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
 776         if mobj is None:
 777             raise ExtractorError('Invalid URL: %s' % url)
 778         video_id = mobj.group(2)
 779         return video_id
 780
 781     def _extract_from_m3u8(self, manifest_url, video_id):
 782         url_map = {}
 783
 784         def _get_urls(_manifest):
 785             lines = _manifest.split('\n')
 786             urls = filter(lambda l: l and not l.startswith('#'),
 787                           lines)
 788             return urls
 789         manifest = self._download_webpage(manifest_url, video_id, 'Downloading formats manifest')
 790         formats_urls = _get_urls(manifest)
 791         for format_url in formats_urls:
 792             itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
 793             url_map[itag] = format_url
 794         return url_map
 795
 796     def _extract_annotations(self, video_id):
 797         url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
 798         return self._download_webpage(url, video_id, note='Searching for annotations.', errnote='Unable to download video annotations.')
 799
 800     def _parse_dash_manifest(
 801             self, video_id, dash_manifest_url, player_url, age_gate, fatal=True):
 802         def decrypt_sig(mobj):
 803             s = mobj.group(1)
 804             dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
 805             return '/signature/%s' % dec_s
 806         dash_manifest_url = re.sub(r'/s/([a-fA-F0-9\.]+)', decrypt_sig, dash_manifest_url)
 807         dash_doc = self._download_xml(
 808             dash_manifest_url, video_id,
 809             note='Downloading DASH manifest',
 810             errnote='Could not download DASH manifest',
 811             fatal=fatal)
 812
 813         if dash_doc is False:
 814             return []
 815
 816         formats = []
 817         for a in dash_doc.findall('.//{urn:mpeg:DASH:schema:MPD:2011}AdaptationSet'):
 818             mime_type = a.attrib.get('mimeType')
 819             for r in a.findall('{urn:mpeg:DASH:schema:MPD:2011}Representation'):
 820                 url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL')
 821                 if url_el is None:
 822                     continue
 823                 if mime_type == 'text/vtt':
 824                     # TODO implement WebVTT downloading
 825                     pass
 826                 elif mime_type.startswith('audio/') or mime_type.startswith('video/'):
 827                     format_id = r.attrib['id']
 828                     video_url = url_el.text
 829                     filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
 830                     f = {
 831                         'format_id': format_id,
 832                         'url': video_url,
 833                         'width': int_or_none(r.attrib.get('width')),
 834                         'height': int_or_none(r.attrib.get('height')),
 835                         'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
 836                         'asr': int_or_none(r.attrib.get('audioSamplingRate')),
 837                         'filesize': filesize,
 838                         'fps': int_or_none(r.attrib.get('frameRate')),
 839                     }
 840                     try:
 841                         existing_format = next(
 842                             fo for fo in formats
 843                             if fo['format_id'] == format_id)
 844                     except StopIteration:
 845                         full_info = self._formats.get(format_id, {}).copy()
 846                         full_info.update(f)
 847                         codecs = r.attrib.get('codecs')
 848                         if codecs:
 849                             if full_info.get('acodec') == 'none' and 'vcodec' not in full_info:
 850                                 full_info['vcodec'] = codecs
 851                             elif full_info.get('vcodec') == 'none' and 'acodec' not in full_info:
 852                                 full_info['acodec'] = codecs
 853                         formats.append(full_info)
 854                     else:
 855                         existing_format.update(f)
 856                 else:
 857                     self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
 858         return formats
 859
 860     def _real_extract(self, url):
 861         proto = (
 862             'http' if self._downloader.params.get('prefer_insecure', False)
 863             else 'https')
 864
 865         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
 866         mobj = re.search(self._NEXT_URL_RE, url)
 867         if mobj:
 868             url = proto + '://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
 869         video_id = self.extract_id(url)
 870
 871         # Get video webpage
 872         url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1&bpctr=9999999999' % video_id
 873         video_webpage = self._download_webpage(url, video_id)
 874
 875         # Attempt to extract SWF player URL
 876         mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
 877         if mobj is not None:
 878             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
 879         else:
 880             player_url = None
 881
 882         dash_mpds = []
 883
 884         def add_dash_mpd(video_info):
 885             dash_mpd = video_info.get('dashmpd')
 886             if dash_mpd and dash_mpd[0] not in dash_mpds:
 887                 dash_mpds.append(dash_mpd[0])
 888
 889         # Get video info
 890         embed_webpage = None
 891         if re.search(r'player-age-gate-content">', video_webpage) is not None:
 892             age_gate = True
 893             # We simulate the access to the video from www.youtube.com/v/{video_id}
 894             # this can be viewed without login into Youtube
 895             url = proto + '://www.youtube.com/embed/%s' % video_id
 896             embed_webpage = self._download_webpage(url, video_id, 'Downloading embed webpage')
 897             data = compat_urllib_parse.urlencode({
 898                 'video_id': video_id,
 899                 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
 900                 'sts': self._search_regex(
 901                     r'"sts"\s*:\s*(\d+)', embed_webpage, 'sts', default=''),
 902             })
 903             video_info_url = proto + '://www.youtube.com/get_video_info?' + data
 904             video_info_webpage = self._download_webpage(
 905                 video_info_url, video_id,
 906                 note='Refetching age-gated info webpage',
 907                 errnote='unable to download video info webpage')
 908             video_info = compat_parse_qs(video_info_webpage)
 909             add_dash_mpd(video_info)
 910         else:
 911             age_gate = False
 912             video_info = None
 913             # Try looking directly into the video webpage
 914             mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage)
 915             if mobj:
 916                 json_code = uppercase_escape(mobj.group(1))
 917                 ytplayer_config = json.loads(json_code)
 918                 args = ytplayer_config['args']
 919                 if args.get('url_encoded_fmt_stream_map'):
 920                     # Convert to the same format returned by compat_parse_qs
 921                     video_info = dict((k, [v]) for k, v in args.items())
 922                     add_dash_mpd(video_info)
 923             if not video_info or self._downloader.params.get('youtube_include_dash_manifest', True):
 924                 # We also try looking in get_video_info since it may contain different dashmpd
 925                 # URL that points to a DASH manifest with possibly different itag set (some itags
 926                 # are missing from DASH manifest pointed by webpage's dashmpd, some - from DASH
 927                 # manifest pointed by get_video_info's dashmpd).
 928                 # The general idea is to take a union of itags of both DASH manifests (for example
 929                 # video with such 'manifest behavior' see https://github.com/rg3/youtube-dl/issues/6093)
 930                 self.report_video_info_webpage_download(video_id)
 931                 for el_type in ['&el=info', '&el=embedded', '&el=detailpage', '&el=vevo', '']:
 932                     video_info_url = (
 933                         '%s://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
 934                         % (proto, video_id, el_type))
 935                     video_info_webpage = self._download_webpage(
 936                         video_info_url,
 937                         video_id, note=False,
 938                         errnote='unable to download video info webpage')
 939                     get_video_info = compat_parse_qs(video_info_webpage)
 940                     add_dash_mpd(get_video_info)
 941                     if not video_info:
 942                         video_info = get_video_info
 943                     if 'token' in get_video_info:
 944                         break
 945         if 'token' not in video_info:
 946             if 'reason' in video_info:
 947                 if 'The uploader has not made this video available in your country.' in video_info['reason']:
 948                     regions_allowed = self._html_search_meta('regionsAllowed', video_webpage, default=None)
 949                     if regions_allowed is not None:
 950                         raise ExtractorError('YouTube said: This video is available in %s only' % (
 951                             ', '.join(map(ISO3166Utils.short2full, regions_allowed.split(',')))),
 952                             expected=True)
 953                 raise ExtractorError(
 954                     'YouTube said: %s' % video_info['reason'][0],
 955                     expected=True, video_id=video_id)
 956             else:
 957                 raise ExtractorError(
 958                     '"token" parameter not in video info for unknown reason',
 959                     video_id=video_id)
 960
 961         if 'view_count' in video_info:
 962             view_count = int(video_info['view_count'][0])
 963         else:
 964             view_count = None
 965
 966         # Check for "rental" videos
 967         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
 968             raise ExtractorError('"rental" videos not supported')
 969
 970         # Start extracting information
 971         self.report_information_extraction(video_id)
 972
 973         # uploader
 974         if 'author' not in video_info:
 975             raise ExtractorError('Unable to extract uploader name')
 976         video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
 977
 978         # uploader_id
 979         video_uploader_id = None
 980         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
 981         if mobj is not None:
 982             video_uploader_id = mobj.group(1)
 983         else:
 984             self._downloader.report_warning('unable to extract uploader nickname')
 985
 986         # title
 987         if 'title' in video_info:
 988             video_title = video_info['title'][0]
 989         else:
 990             self._downloader.report_warning('Unable to extract video title')
 991             video_title = '_'
 992
 993         # thumbnail image
 994         # We try first to get a high quality image:
 995         m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
 996                             video_webpage, re.DOTALL)
 997         if m_thumb is not None:
 998             video_thumbnail = m_thumb.group(1)
 999         elif 'thumbnail_url' not in video_info:
1000             self._downloader.report_warning('unable to extract video thumbnail')
1001             video_thumbnail = None
1002         else:   # don't panic if we can't find it
1003             video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
1004
1005         # upload date
1006         upload_date = self._html_search_meta(
1007             'datePublished', video_webpage, 'upload date', default=None)
1008         if not upload_date:
1009             upload_date = self._search_regex(
1010                 [r'(?s)id="eow-date.*?>(.*?)</span>',
1011                  r'id="watch-uploader-info".*?>.*?(?:Published|Uploaded|Streamed live|Started) on (.+?)</strong>'],
1012                 video_webpage, 'upload date', default=None)
1013             if upload_date:
1014                 upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
1015         upload_date = unified_strdate(upload_date)
1016
1017         m_cat_container = self._search_regex(
1018             r'(?s)<h4[^>]*>\s*Category\s*</h4>\s*<ul[^>]*>(.*?)</ul>',
1019             video_webpage, 'categories', default=None)
1020         if m_cat_container:
1021             category = self._html_search_regex(
1022                 r'(?s)<a[^<]+>(.*?)</a>', m_cat_container, 'category',
1023                 default=None)
1024             video_categories = None if category is None else [category]
1025         else:
1026             video_categories = None
1027
1028         # description
1029         video_description = get_element_by_id("eow-description", video_webpage)
1030         if video_description:
1031             video_description = re.sub(r'''(?x)
1032                 <a\s+
1033                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1034                     title="([^"]+)"\s+
1035                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1036                     class="yt-uix-redirect-link"\s*>
1037                 [^<]+
1038                 </a>
1039             ''', r'\1', video_description)
1040             video_description = clean_html(video_description)
1041         else:
1042             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
1043             if fd_mobj:
1044                 video_description = unescapeHTML(fd_mobj.group(1))
1045             else:
1046                 video_description = ''
1047
1048         def _extract_count(count_name):
1049             return str_to_int(self._search_regex(
1050                 r'-%s-button[^>]+><span[^>]+class="yt-uix-button-content"[^>]*>([\d,]+)</span>'
1051                 % re.escape(count_name),
1052                 video_webpage, count_name, default=None))
1053
1054         like_count = _extract_count('like')
1055         dislike_count = _extract_count('dislike')
1056
1057         # subtitles
1058         video_subtitles = self.extract_subtitles(video_id, video_webpage)
1059         automatic_captions = self.extract_automatic_captions(video_id, video_webpage)
1060
1061         if 'length_seconds' not in video_info:
1062             self._downloader.report_warning('unable to extract video duration')
1063             video_duration = None
1064         else:
1065             video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]))
1066
1067         # annotations
1068         video_annotations = None
1069         if self._downloader.params.get('writeannotations', False):
1070             video_annotations = self._extract_annotations(video_id)
1071
1072         def _map_to_format_list(urlmap):
1073             formats = []
1074             for itag, video_real_url in urlmap.items():
1075                 dct = {
1076                     'format_id': itag,
1077                     'url': video_real_url,
1078                     'player_url': player_url,
1079                 }
1080                 if itag in self._formats:
1081                     dct.update(self._formats[itag])
1082                 formats.append(dct)
1083             return formats
1084
1085         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
1086             self.report_rtmp_download()
1087             formats = [{
1088                 'format_id': '_rtmp',
1089                 'protocol': 'rtmp',
1090                 'url': video_info['conn'][0],
1091                 'player_url': player_url,
1092             }]
1093         elif len(video_info.get('url_encoded_fmt_stream_map', [''])[0]) >= 1 or len(video_info.get('adaptive_fmts', [''])[0]) >= 1:
1094             encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts', [''])[0]
1095             if 'rtmpe%3Dyes' in encoded_url_map:
1096                 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
1097             url_map = {}
1098             for url_data_str in encoded_url_map.split(','):
1099                 url_data = compat_parse_qs(url_data_str)
1100                 if 'itag' not in url_data or 'url' not in url_data:
1101                     continue
1102                 format_id = url_data['itag'][0]
1103                 url = url_data['url'][0]
1104
1105                 if 'sig' in url_data:
1106                     url += '&signature=' + url_data['sig'][0]
1107                 elif 's' in url_data:
1108                     encrypted_sig = url_data['s'][0]
1109                     ASSETS_RE = r'"assets":.+?"js":\s*("[^"]+")'
1110
1111                     jsplayer_url_json = self._search_regex(
1112                         ASSETS_RE,
1113                         embed_webpage if age_gate else video_webpage,
1114                         'JS player URL (1)', default=None)
1115                     if not jsplayer_url_json and not age_gate:
1116                         # We need the embed website after all
1117                         if embed_webpage is None:
1118                             embed_url = proto + '://www.youtube.com/embed/%s' % video_id
1119                             embed_webpage = self._download_webpage(
1120                                 embed_url, video_id, 'Downloading embed webpage')
1121                         jsplayer_url_json = self._search_regex(
1122                             ASSETS_RE, embed_webpage, 'JS player URL')
1123
1124                     player_url = json.loads(jsplayer_url_json)
1125                     if player_url is None:
1126                         player_url_json = self._search_regex(
1127                             r'ytplayer\.config.*?"url"\s*:\s*("[^"]+")',
1128                             video_webpage, 'age gate player URL')
1129                         player_url = json.loads(player_url_json)
1130
1131                     if self._downloader.params.get('verbose'):
1132                         if player_url is None:
1133                             player_version = 'unknown'
1134                             player_desc = 'unknown'
1135                         else:
1136                             if player_url.endswith('swf'):
1137                                 player_version = self._search_regex(
1138                                     r'-(.+?)(?:/watch_as3)?\.swf$', player_url,
1139                                     'flash player', fatal=False)
1140                                 player_desc = 'flash player %s' % player_version
1141                             else:
1142                                 player_version = self._search_regex(
1143                                     r'html5player-([^/]+?)(?:/html5player)?\.js',
1144                                     player_url,
1145                                     'html5 player', fatal=False)
1146                                 player_desc = 'html5 player %s' % player_version
1147
1148                         parts_sizes = self._signature_cache_id(encrypted_sig)
1149                         self.to_screen('{%s} signature length %s, %s' %
1150                                        (format_id, parts_sizes, player_desc))
1151
1152                     signature = self._decrypt_signature(
1153                         encrypted_sig, video_id, player_url, age_gate)
1154                     url += '&signature=' + signature
1155                 if 'ratebypass' not in url:
1156                     url += '&ratebypass=yes'
1157                 url_map[format_id] = url
1158             formats = _map_to_format_list(url_map)
1159         elif video_info.get('hlsvp'):
1160             manifest_url = video_info['hlsvp'][0]
1161             url_map = self._extract_from_m3u8(manifest_url, video_id)
1162             formats = _map_to_format_list(url_map)
1163         else:
1164             raise ExtractorError('no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
1165
1166         # Look for the DASH manifest
1167         if self._downloader.params.get('youtube_include_dash_manifest', True):
1168             dash_mpd_fatal = True
1169             for dash_manifest_url in dash_mpds:
1170                 dash_formats = {}
1171                 try:
1172                     for df in self._parse_dash_manifest(
1173                             video_id, dash_manifest_url, player_url, age_gate, dash_mpd_fatal):
1174                         # Do not overwrite DASH format found in some previous DASH manifest
1175                         if df['format_id'] not in dash_formats:
1176                             dash_formats[df['format_id']] = df
1177                         # Additional DASH manifests may end up in HTTP Error 403 therefore
1178                         # allow them to fail without bug report message if we already have
1179                         # some DASH manifest succeeded. This is temporary workaround to reduce
1180                         # burst of bug reports until we figure out the reason and whether it
1181                         # can be fixed at all.
1182                         dash_mpd_fatal = False
1183                 except (ExtractorError, KeyError) as e:
1184                     self.report_warning(
1185                         'Skipping DASH manifest: %r' % e, video_id)
1186                 if dash_formats:
1187                     # Remove the formats we found through non-DASH, they
1188                     # contain less info and it can be wrong, because we use
1189                     # fixed values (for example the resolution). See
1190                     # https://github.com/rg3/youtube-dl/issues/5774 for an
1191                     # example.
1192                     formats = [f for f in formats if f['format_id'] not in dash_formats.keys()]
1193                     formats.extend(dash_formats.values())
1194
1195         # Check for malformed aspect ratio
1196         stretched_m = re.search(
1197             r'<meta\s+property="og:video:tag".*?content="yt:stretch=(?P<w>[0-9]+):(?P<h>[0-9]+)">',
1198             video_webpage)
1199         if stretched_m:
1200             ratio = float(stretched_m.group('w')) / float(stretched_m.group('h'))
1201             for f in formats:
1202                 if f.get('vcodec') != 'none':
1203                     f['stretched_ratio'] = ratio
1204
1205         self._sort_formats(formats)
1206
1207         return {
1208             'id': video_id,
1209             'uploader': video_uploader,
1210             'uploader_id': video_uploader_id,
1211             'upload_date': upload_date,
1212             'title': video_title,
1213             'thumbnail': video_thumbnail,
1214             'description': video_description,
1215             'categories': video_categories,
1216             'subtitles': video_subtitles,
1217             'automatic_captions': automatic_captions,
1218             'duration': video_duration,
1219             'age_limit': 18 if age_gate else 0,
1220             'annotations': video_annotations,
1221             'webpage_url': proto + '://www.youtube.com/watch?v=%s' % video_id,
1222             'view_count': view_count,
1223             'like_count': like_count,
1224             'dislike_count': dislike_count,
1225             'average_rating': float_or_none(video_info.get('avg_rating', [None])[0]),
1226             'formats': formats,
1227         }
1228
1229
1230 class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
1231     IE_DESC = 'YouTube.com playlists'
1232     _VALID_URL = r"""(?x)(?:
1233                         (?:https?://)?
1234                         (?:\w+\.)?
1235                         youtube\.com/
1236                         (?:
1237                            (?:course|view_play_list|my_playlists|artist|playlist|watch|embed/videoseries)
1238                            \? (?:.*?&)*? (?:p|a|list)=
1239                         |  p/
1240                         )
1241                         (
1242                             (?:PL|LL|EC|UU|FL|RD|UL)?[0-9A-Za-z-_]{10,}
1243                             # Top tracks, they can also include dots
1244                             |(?:MC)[\w\.]*
1245                         )
1246                         .*
1247                      |
1248                         ((?:PL|LL|EC|UU|FL|RD|UL)[0-9A-Za-z-_]{10,})
1249                      )"""
1250     _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s'
1251     _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&amp;[^"]*?index=(?P<index>\d+)'
1252     IE_NAME = 'youtube:playlist'
1253     _TESTS = [{
1254         'url': 'https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re',
1255         'info_dict': {
1256             'title': 'ytdl test PL',
1257             'id': 'PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re',
1258         },
1259         'playlist_count': 3,
1260     }, {
1261         'url': 'https://www.youtube.com/playlist?list=PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx',
1262         'info_dict': {
1263             'id': 'PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx',
1264             'title': 'YDL_Empty_List',
1265         },
1266         'playlist_count': 0,
1267     }, {
1268         'note': 'Playlist with deleted videos (#651). As a bonus, the video #51 is also twice in this list.',
1269         'url': 'https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC',
1270         'info_dict': {
1271             'title': '29C3: Not my department',
1272             'id': 'PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC',
1273         },
1274         'playlist_count': 95,
1275     }, {
1276         'note': 'issue #673',
1277         'url': 'PLBB231211A4F62143',
1278         'info_dict': {
1279             'title': '[OLD]Team Fortress 2 (Class-based LP)',
1280             'id': 'PLBB231211A4F62143',
1281         },
1282         'playlist_mincount': 26,
1283     }, {
1284         'note': 'Large playlist',
1285         'url': 'https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q',
1286         'info_dict': {
1287             'title': 'Uploads from Cauchemar',
1288             'id': 'UUBABnxM4Ar9ten8Mdjj1j0Q',
1289         },
1290         'playlist_mincount': 799,
1291     }, {
1292         'url': 'PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl',
1293         'info_dict': {
1294             'title': 'YDL_safe_search',
1295             'id': 'PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl',
1296         },
1297         'playlist_count': 2,
1298     }, {
1299         'note': 'embedded',
1300         'url': 'http://www.youtube.com/embed/videoseries?list=PL6IaIsEjSbf96XFRuNccS_RuEXwNdsoEu',
1301         'playlist_count': 4,
1302         'info_dict': {
1303             'title': 'JODA15',
1304             'id': 'PL6IaIsEjSbf96XFRuNccS_RuEXwNdsoEu',
1305         }
1306     }, {
1307         'note': 'Embedded SWF player',
1308         'url': 'http://www.youtube.com/p/YN5VISEtHet5D4NEvfTd0zcgFk84NqFZ?hl=en_US&fs=1&rel=0',
1309         'playlist_count': 4,
1310         'info_dict': {
1311             'title': 'JODA7',
1312             'id': 'YN5VISEtHet5D4NEvfTd0zcgFk84NqFZ',
1313         }
1314     }, {
1315         'note': 'Buggy playlist: the webpage has a "Load more" button but it doesn\'t have more videos',
1316         'url': 'https://www.youtube.com/playlist?list=UUXw-G3eDE9trcvY2sBMM_aA',
1317         'info_dict': {
1318             'title': 'Uploads from Interstellar Movie',
1319             'id': 'UUXw-G3eDE9trcvY2sBMM_aA',
1320         },
1321         'playlist_mincout': 21,
1322     }]
1323
1324     def _real_initialize(self):
1325         self._login()
1326
1327     def _extract_mix(self, playlist_id):
1328         # The mixes are generated from a single video
1329         # the id of the playlist is just 'RD' + video_id
1330         url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id)
1331         webpage = self._download_webpage(
1332             url, playlist_id, 'Downloading Youtube mix')
1333         search_title = lambda class_name: get_element_by_attribute('class', class_name, webpage)
1334         title_span = (
1335             search_title('playlist-title') or
1336             search_title('title long-title') or
1337             search_title('title'))
1338         title = clean_html(title_span)
1339         ids = orderedSet(re.findall(
1340             r'''(?xs)data-video-username=".*?".*?
1341                        href="/watch\?v=([0-9A-Za-z_-]{11})&amp;[^"]*?list=%s''' % re.escape(playlist_id),
1342             webpage))
1343         url_results = self._ids_to_results(ids)
1344
1345         return self.playlist_result(url_results, playlist_id, title)
1346
1347     def _extract_playlist(self, playlist_id):
1348         url = self._TEMPLATE_URL % playlist_id
1349         page = self._download_webpage(url, playlist_id)
1350
1351         for match in re.findall(r'<div class="yt-alert-message">([^<]+)</div>', page):
1352             match = match.strip()
1353             # Check if the playlist exists or is private
1354             if re.match(r'[^<]*(The|This) playlist (does not exist|is private)[^<]*', match):
1355                 raise ExtractorError(
1356                     'The playlist doesn\'t exist or is private, use --username or '
1357                     '--netrc to access it.',
1358                     expected=True)
1359             elif re.match(r'[^<]*Invalid parameters[^<]*', match):
1360                 raise ExtractorError(
1361                     'Invalid parameters. Maybe URL is incorrect.',
1362                     expected=True)
1363             elif re.match(r'[^<]*Choose your language[^<]*', match):
1364                 continue
1365             else:
1366                 self.report_warning('Youtube gives an alert message: ' + match)
1367
1368         # Extract the video ids from the playlist pages
1369         def _entries():
1370             more_widget_html = content_html = page
1371             for page_num in itertools.count(1):
1372                 matches = re.finditer(self._VIDEO_RE, content_html)
1373                 # We remove the duplicates and the link with index 0
1374                 # (it's not the first video of the playlist)
1375                 new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
1376                 for vid_id in new_ids:
1377                     yield self.url_result(vid_id, 'Youtube', video_id=vid_id)
1378
1379                 mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
1380                 if not mobj:
1381                     break
1382
1383                 more = self._download_json(
1384                     'https://youtube.com/%s' % mobj.group('more'), playlist_id,
1385                     'Downloading page #%s' % page_num,
1386                     transform_source=uppercase_escape)
1387                 content_html = more['content_html']
1388                 if not content_html.strip():
1389                     # Some webpages show a "Load more" button but they don't
1390                     # have more videos
1391                     break
1392                 more_widget_html = more['load_more_widget_html']
1393
1394         playlist_title = self._html_search_regex(
1395             r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>',
1396             page, 'title')
1397
1398         return self.playlist_result(_entries(), playlist_id, playlist_title)
1399
1400     def _real_extract(self, url):
1401         # Extract playlist id
1402         mobj = re.match(self._VALID_URL, url)
1403         if mobj is None:
1404             raise ExtractorError('Invalid URL: %s' % url)
1405         playlist_id = mobj.group(1) or mobj.group(2)
1406
1407         # Check if it's a video-specific URL
1408         query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1409         if 'v' in query_dict:
1410             video_id = query_dict['v'][0]
1411             if self._downloader.params.get('noplaylist'):
1412                 self.to_screen('Downloading just video %s because of --no-playlist' % video_id)
1413                 return self.url_result(video_id, 'Youtube', video_id=video_id)
1414             else:
1415                 self.to_screen('Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1416
1417         if playlist_id.startswith('RD') or playlist_id.startswith('UL'):
1418             # Mixes require a custom extraction process
1419             return self._extract_mix(playlist_id)
1420
1421         return self._extract_playlist(playlist_id)
1422
1423
1424 class YoutubeChannelIE(InfoExtractor):
1425     IE_DESC = 'YouTube.com channels'
1426     _VALID_URL = r'https?://(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/(?P<id>[0-9A-Za-z_-]+)'
1427     _TEMPLATE_URL = 'https://www.youtube.com/channel/%s/videos'
1428     IE_NAME = 'youtube:channel'
1429     _TESTS = [{
1430         'note': 'paginated channel',
1431         'url': 'https://www.youtube.com/channel/UCKfVa3S1e4PHvxWcwyMMg8w',
1432         'playlist_mincount': 91,
1433         'info_dict': {
1434             'id': 'UCKfVa3S1e4PHvxWcwyMMg8w',
1435         }
1436     }]
1437
1438     @staticmethod
1439     def extract_videos_from_page(page):
1440         ids_in_page = []
1441         titles_in_page = []
1442         for mobj in re.finditer(r'(?:title="(?P<title>[^"]+)"[^>]+)?href="/watch\?v=(?P<id>[0-9A-Za-z_-]+)&?', page):
1443             video_id = mobj.group('id')
1444             video_title = unescapeHTML(mobj.group('title'))
1445             try:
1446                 idx = ids_in_page.index(video_id)
1447                 if video_title and not titles_in_page[idx]:
1448                     titles_in_page[idx] = video_title
1449             except ValueError:
1450                 ids_in_page.append(video_id)
1451                 titles_in_page.append(video_title)
1452         return zip(ids_in_page, titles_in_page)
1453
1454     def _real_extract(self, url):
1455         channel_id = self._match_id(url)
1456
1457         url = self._TEMPLATE_URL % channel_id
1458
1459         # Channel by page listing is restricted to 35 pages of 30 items, i.e. 1050 videos total (see #5778)
1460         # Workaround by extracting as a playlist if managed to obtain channel playlist URL
1461         # otherwise fallback on channel by page extraction
1462         channel_page = self._download_webpage(
1463             url + '?view=57', channel_id,
1464             'Downloading channel page', fatal=False)
1465         channel_playlist_id = self._html_search_meta(
1466             'channelId', channel_page, 'channel id', default=None)
1467         if not channel_playlist_id:
1468             channel_playlist_id = self._search_regex(
1469                 r'data-channel-external-id="([^"]+)"',
1470                 channel_page, 'channel id', default=None)
1471         if channel_playlist_id and channel_playlist_id.startswith('UC'):
1472             playlist_id = 'UU' + channel_playlist_id[2:]
1473             return self.url_result(
1474                 compat_urlparse.urljoin(url, '/playlist?list=%s' % playlist_id), 'YoutubePlaylist')
1475
1476         channel_page = self._download_webpage(url, channel_id, 'Downloading page #1')
1477         autogenerated = re.search(r'''(?x)
1478                 class="[^"]*?(?:
1479                     channel-header-autogenerated-label|
1480                     yt-channel-title-autogenerated
1481                 )[^"]*"''', channel_page) is not None
1482
1483         if autogenerated:
1484             # The videos are contained in a single page
1485             # the ajax pages can't be used, they are empty
1486             entries = [
1487                 self.url_result(
1488                     video_id, 'Youtube', video_id=video_id,
1489                     video_title=video_title)
1490                 for video_id, video_title in self.extract_videos_from_page(channel_page)]
1491             return self.playlist_result(entries, channel_id)
1492
1493         def _entries():
1494             more_widget_html = content_html = channel_page
1495             for pagenum in itertools.count(1):
1496
1497                 for video_id, video_title in self.extract_videos_from_page(content_html):
1498                     yield self.url_result(
1499                         video_id, 'Youtube', video_id=video_id,
1500                         video_title=video_title)
1501
1502                 mobj = re.search(
1503                     r'data-uix-load-more-href="/?(?P<more>[^"]+)"',
1504                     more_widget_html)
1505                 if not mobj:
1506                     break
1507
1508                 more = self._download_json(
1509                     'https://youtube.com/%s' % mobj.group('more'), channel_id,
1510                     'Downloading page #%s' % (pagenum + 1),
1511                     transform_source=uppercase_escape)
1512                 content_html = more['content_html']
1513                 more_widget_html = more['load_more_widget_html']
1514
1515         return self.playlist_result(_entries(), channel_id)
1516
1517
1518 class YoutubeUserIE(YoutubeChannelIE):
1519     IE_DESC = 'YouTube.com user videos (URL or "ytuser" keyword)'
1520     _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)(?P<id>[A-Za-z0-9_-]+)'
1521     _TEMPLATE_URL = 'https://www.youtube.com/user/%s/videos'
1522     IE_NAME = 'youtube:user'
1523
1524     _TESTS = [{
1525         'url': 'https://www.youtube.com/user/TheLinuxFoundation',
1526         'playlist_mincount': 320,
1527         'info_dict': {
1528             'title': 'TheLinuxFoundation',
1529         }
1530     }, {
1531         'url': 'ytuser:phihag',
1532         'only_matching': True,
1533     }]
1534
1535     @classmethod
1536     def suitable(cls, url):
1537         # Don't return True if the url can be extracted with other youtube
1538         # extractor, the regex would is too permissive and it would match.
1539         other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1540         if any(ie.suitable(url) for ie in other_ies):
1541             return False
1542         else:
1543             return super(YoutubeUserIE, cls).suitable(url)
1544
1545
1546 class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE):
1547     IE_DESC = 'YouTube.com searches'
1548     # there doesn't appear to be a real limit, for example if you search for
1549     # 'python' you get more than 8.000.000 results
1550     _MAX_RESULTS = float('inf')
1551     IE_NAME = 'youtube:search'
1552     _SEARCH_KEY = 'ytsearch'
1553     _EXTRA_QUERY_ARGS = {}
1554     _TESTS = []
1555
1556     def _get_n_results(self, query, n):
1557         """Get a specified number of results for a query"""
1558
1559         videos = []
1560         limit = n
1561
1562         for pagenum in itertools.count(1):
1563             url_query = {
1564                 'search_query': query.encode('utf-8'),
1565                 'page': pagenum,
1566                 'spf': 'navigate',
1567             }
1568             url_query.update(self._EXTRA_QUERY_ARGS)
1569             result_url = 'https://www.youtube.com/results?' + compat_urllib_parse.urlencode(url_query)
1570             data = self._download_json(
1571                 result_url, video_id='query "%s"' % query,
1572                 note='Downloading page %s' % pagenum,
1573                 errnote='Unable to download API page')
1574             html_content = data[1]['body']['content']
1575
1576             if 'class="search-message' in html_content:
1577                 raise ExtractorError(
1578                     '[youtube] No video results', expected=True)
1579
1580             new_videos = self._ids_to_results(orderedSet(re.findall(
1581                 r'href="/watch\?v=(.{11})', html_content)))
1582             videos += new_videos
1583             if not new_videos or len(videos) > limit:
1584                 break
1585
1586         if len(videos) > n:
1587             videos = videos[:n]
1588         return self.playlist_result(videos, query)
1589
1590
1591 class YoutubeSearchDateIE(YoutubeSearchIE):
1592     IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
1593     _SEARCH_KEY = 'ytsearchdate'
1594     IE_DESC = 'YouTube.com searches, newest videos first'
1595     _EXTRA_QUERY_ARGS = {'search_sort': 'video_date_uploaded'}
1596
1597
1598 class YoutubeSearchURLIE(InfoExtractor):
1599     IE_DESC = 'YouTube.com search URLs'
1600     IE_NAME = 'youtube:search_url'
1601     _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?search_query=(?P<query>[^&]+)(?:[&]|$)'
1602     _TESTS = [{
1603         'url': 'https://www.youtube.com/results?baz=bar&search_query=youtube-dl+test+video&filters=video&lclk=video',
1604         'playlist_mincount': 5,
1605         'info_dict': {
1606             'title': 'youtube-dl test video',
1607         }
1608     }]
1609
1610     def _real_extract(self, url):
1611         mobj = re.match(self._VALID_URL, url)
1612         query = compat_urllib_parse.unquote_plus(mobj.group('query'))
1613
1614         webpage = self._download_webpage(url, query)
1615         result_code = self._search_regex(
1616             r'(?s)<ol[^>]+class="item-section"(.*?)</ol>', webpage, 'result HTML')
1617
1618         part_codes = re.findall(
1619             r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code)
1620         entries = []
1621         for part_code in part_codes:
1622             part_title = self._html_search_regex(
1623                 [r'(?s)title="([^"]+)"', r'>([^<]+)</a>'], part_code, 'item title', fatal=False)
1624             part_url_snippet = self._html_search_regex(
1625                 r'(?s)href="([^"]+)"', part_code, 'item URL')
1626             part_url = compat_urlparse.urljoin(
1627                 'https://www.youtube.com/', part_url_snippet)
1628             entries.append({
1629                 '_type': 'url',
1630                 'url': part_url,
1631                 'title': part_title,
1632             })
1633
1634         return {
1635             '_type': 'playlist',
1636             'entries': entries,
1637             'title': query,
1638         }
1639
1640
1641 class YoutubeShowIE(InfoExtractor):
1642     IE_DESC = 'YouTube.com (multi-season) shows'
1643     _VALID_URL = r'https?://www\.youtube\.com/show/(?P<id>[^?#]*)'
1644     IE_NAME = 'youtube:show'
1645     _TESTS = [{
1646         'url': 'http://www.youtube.com/show/airdisasters',
1647         'playlist_mincount': 3,
1648         'info_dict': {
1649             'id': 'airdisasters',
1650             'title': 'Air Disasters',
1651         }
1652     }]
1653
1654     def _real_extract(self, url):
1655         mobj = re.match(self._VALID_URL, url)
1656         playlist_id = mobj.group('id')
1657         webpage = self._download_webpage(
1658             url, playlist_id, 'Downloading show webpage')
1659         # There's one playlist for each season of the show
1660         m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1661         self.to_screen('%s: Found %s seasons' % (playlist_id, len(m_seasons)))
1662         entries = [
1663             self.url_result(
1664                 'https://www.youtube.com' + season.group(1), 'YoutubePlaylist')
1665             for season in m_seasons
1666         ]
1667         title = self._og_search_title(webpage, fatal=False)
1668
1669         return {
1670             '_type': 'playlist',
1671             'id': playlist_id,
1672             'title': title,
1673             'entries': entries,
1674         }
1675
1676
1677 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1678     """
1679     Base class for feed extractors
1680     Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1681     """
1682     _LOGIN_REQUIRED = True
1683
1684     @property
1685     def IE_NAME(self):
1686         return 'youtube:%s' % self._FEED_NAME
1687
1688     def _real_initialize(self):
1689         self._login()
1690
1691     def _real_extract(self, url):
1692         page = self._download_webpage(
1693             'https://www.youtube.com/feed/%s' % self._FEED_NAME, self._PLAYLIST_TITLE)
1694
1695         # The extraction process is the same as for playlists, but the regex
1696         # for the video ids doesn't contain an index
1697         ids = []
1698         more_widget_html = content_html = page
1699         for page_num in itertools.count(1):
1700             matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html)
1701
1702             # 'recommended' feed has infinite 'load more' and each new portion spins
1703             # the same videos in (sometimes) slightly different order, so we'll check
1704             # for unicity and break when portion has no new videos
1705             new_ids = filter(lambda video_id: video_id not in ids, orderedSet(matches))
1706             if not new_ids:
1707                 break
1708
1709             ids.extend(new_ids)
1710
1711             mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
1712             if not mobj:
1713                 break
1714
1715             more = self._download_json(
1716                 'https://youtube.com/%s' % mobj.group('more'), self._PLAYLIST_TITLE,
1717                 'Downloading page #%s' % page_num,
1718                 transform_source=uppercase_escape)
1719             content_html = more['content_html']
1720             more_widget_html = more['load_more_widget_html']
1721
1722         return self.playlist_result(
1723             self._ids_to_results(ids), playlist_title=self._PLAYLIST_TITLE)
1724
1725
1726 class YoutubeWatchLaterIE(YoutubePlaylistIE):
1727     IE_NAME = 'youtube:watchlater'
1728     IE_DESC = 'Youtube watch later list, ":ytwatchlater" for short (requires authentication)'
1729     _VALID_URL = r'https?://www\.youtube\.com/(?:feed/watch_later|playlist\?list=WL)|:ytwatchlater'
1730
1731     _TESTS = []  # override PlaylistIE tests
1732
1733     def _real_extract(self, url):
1734         return self._extract_playlist('WL')
1735
1736
1737 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1738     IE_NAME = 'youtube:favorites'
1739     IE_DESC = 'YouTube.com favourite videos, ":ytfav" for short (requires authentication)'
1740     _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1741     _LOGIN_REQUIRED = True
1742
1743     def _real_extract(self, url):
1744         webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1745         playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, 'favourites playlist id')
1746         return self.url_result(playlist_id, 'YoutubePlaylist')
1747
1748
1749 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1750     IE_DESC = 'YouTube.com recommended videos, ":ytrec" for short (requires authentication)'
1751     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1752     _FEED_NAME = 'recommended'
1753     _PLAYLIST_TITLE = 'Youtube Recommended videos'
1754
1755
1756 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1757     IE_DESC = 'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
1758     _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1759     _FEED_NAME = 'subscriptions'
1760     _PLAYLIST_TITLE = 'Youtube Subscriptions'
1761
1762
1763 class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
1764     IE_DESC = 'Youtube watch history, ":ythistory" for short (requires authentication)'
1765     _VALID_URL = 'https?://www\.youtube\.com/feed/history|:ythistory'
1766     _FEED_NAME = 'history'
1767     _PLAYLIST_TITLE = 'Youtube History'
1768
1769
1770 class YoutubeTruncatedURLIE(InfoExtractor):
1771     IE_NAME = 'youtube:truncated_url'
1772     IE_DESC = False  # Do not list
1773     _VALID_URL = r'''(?x)
1774         (?:https?://)?
1775         (?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/
1776         (?:watch\?(?:
1777             feature=[a-z_]+|
1778             annotation_id=annotation_[^&]+|
1779             x-yt-cl=[0-9]+|
1780             hl=[^&]*|
1781         )?
1782         |
1783             attribution_link\?a=[^&]+
1784         )
1785         $
1786     '''
1787
1788     _TESTS = [{
1789         'url': 'http://www.youtube.com/watch?annotation_id=annotation_3951667041',
1790         'only_matching': True,
1791     }, {
1792         'url': 'http://www.youtube.com/watch?',
1793         'only_matching': True,
1794     }, {
1795         'url': 'https://www.youtube.com/watch?x-yt-cl=84503534',
1796         'only_matching': True,
1797     }, {
1798         'url': 'https://www.youtube.com/watch?feature=foo',
1799         'only_matching': True,
1800     }, {
1801         'url': 'https://www.youtube.com/watch?hl=en-GB',
1802         'only_matching': True,
1803     }]
1804
1805     def _real_extract(self, url):
1806         raise ExtractorError(
1807             'Did you forget to quote the URL? Remember that & is a meta '
1808             'character in most shells, so you want to put the URL in quotes, '
1809             'like  youtube-dl '
1810             '"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
1811             ' or simply  youtube-dl BaW_jenozKc  .',
1812             expected=True)
1813
1814
1815 class YoutubeTruncatedIDIE(InfoExtractor):
1816     IE_NAME = 'youtube:truncated_id'
1817     IE_DESC = False  # Do not list
1818     _VALID_URL = r'https?://(?:www\.)?youtube\.com/watch\?v=(?P<id>[0-9A-Za-z_-]{1,10})$'
1819
1820     _TESTS = [{
1821         'url': 'https://www.youtube.com/watch?v=N_708QY7Ob',
1822         'only_matching': True,
1823     }]
1824
1825     def _real_extract(self, url):
1826         video_id = self._match_id(url)
1827         raise ExtractorError(
1828             'Incomplete YouTube ID %s. URL %s looks truncated.' % (video_id, url),
1829             expected=True)