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_parse_unquote,
  21     compat_urllib_parse_unquote_plus,
  22     compat_urllib_request,
  23     compat_urlparse,
  24     compat_str,
  25 )
  26 from ..utils import (
  27     clean_html,
  28     ExtractorError,
  29     float_or_none,
  30     get_element_by_attribute,
  31     get_element_by_id,
  32     int_or_none,
  33     orderedSet,
  34     str_to_int,
  35     unescapeHTML,
  36     unified_strdate,
  37     uppercase_escape,
  38     ISO3166Utils,
  39 )
  40
  41
  42 class YoutubeBaseInfoExtractor(InfoExtractor):
  43     """Provide base functions for Youtube extractors"""
  44     _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
  45     _TWOFACTOR_URL = 'https://accounts.google.com/SecondFactor'
  46     _NETRC_MACHINE = 'youtube'
  47     # If True it will raise an error if no login info is provided
  48     _LOGIN_REQUIRED = False
  49
  50     def _set_language(self):
  51         self._set_cookie(
  52             '.youtube.com', 'PREF', 'f1=50000000&hl=en',
  53             # YouTube sets the expire time to about two months
  54             expire_time=time.time() + 2 * 30 * 24 * 3600)
  55
  56     def _ids_to_results(self, ids):
  57         return [
  58             self.url_result(vid_id, 'Youtube', video_id=vid_id)
  59             for vid_id in ids]
  60
  61     def _login(self):
  62         """
  63         Attempt to log in to YouTube.
  64         True is returned if successful or skipped.
  65         False is returned if login failed.
  66
  67         If _LOGIN_REQUIRED is set and no authentication was provided, an error is raised.
  68         """
  69         (username, password) = self._get_login_info()
  70         # No authentication to be performed
  71         if username is None:
  72             if self._LOGIN_REQUIRED:
  73                 raise ExtractorError('No login info available, needed for using %s.' % self.IE_NAME, expected=True)
  74             return True
  75
  76         login_page = self._download_webpage(
  77             self._LOGIN_URL, None,
  78             note='Downloading login page',
  79             errnote='unable to fetch login page', fatal=False)
  80         if login_page is False:
  81             return
  82
  83         galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"',
  84                                   login_page, 'Login GALX parameter')
  85
  86         # Log in
  87         login_form_strs = {
  88             'continue': 'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
  89             'Email': username,
  90             'GALX': galx,
  91             'Passwd': password,
  92
  93             'PersistentCookie': 'yes',
  94             '_utf8': '霱',
  95             'bgresponse': 'js_disabled',
  96             'checkConnection': '',
  97             'checkedDomains': 'youtube',
  98             'dnConn': '',
  99             'pstMsg': '0',
 100             'rmShown': '1',
 101             'secTok': '',
 102             'signIn': 'Sign in',
 103             'timeStmp': '',
 104             'service': 'youtube',
 105             'uilel': '3',
 106             'hl': 'en_US',
 107         }
 108
 109         # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
 110         # chokes on unicode
 111         login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k, v in login_form_strs.items())
 112         login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
 113
 114         req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
 115         login_results = self._download_webpage(
 116             req, None,
 117             note='Logging in', errnote='unable to log in', fatal=False)
 118         if login_results is False:
 119             return False
 120
 121         if re.search(r'id="errormsg_0_Passwd"', login_results) is not None:
 122             raise ExtractorError('Please use your account password and a two-factor code instead of an application-specific password.', expected=True)
 123
 124         # Two-Factor
 125         # TODO add SMS and phone call support - these require making a request and then prompting the user
 126
 127         if re.search(r'(?i)<form[^>]* id="gaia_secondfactorform"', login_results) is not None:
 128             tfa_code = self._get_tfa_info()
 129
 130             if tfa_code is None:
 131                 self._downloader.report_warning('Two-factor authentication required. Provide it with --twofactor <code>')
 132                 self._downloader.report_warning('(Note that only TOTP (Google Authenticator App) codes work at this time.)')
 133                 return False
 134
 135             # Unlike the first login form, secTok and timeStmp are both required for the TFA form
 136
 137             match = re.search(r'id="secTok"\n\s+value=\'(.+)\'/>', login_results, re.M | re.U)
 138             if match is None:
 139                 self._downloader.report_warning('Failed to get secTok - did the page structure change?')
 140             secTok = match.group(1)
 141             match = re.search(r'id="timeStmp"\n\s+value=\'(.+)\'/>', login_results, re.M | re.U)
 142             if match is None:
 143                 self._downloader.report_warning('Failed to get timeStmp - did the page structure change?')
 144             timeStmp = match.group(1)
 145
 146             tfa_form_strs = {
 147                 'continue': 'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
 148                 'smsToken': '',
 149                 'smsUserPin': tfa_code,
 150                 'smsVerifyPin': 'Verify',
 151
 152                 'PersistentCookie': 'yes',
 153                 'checkConnection': '',
 154                 'checkedDomains': 'youtube',
 155                 'pstMsg': '1',
 156                 'secTok': secTok,
 157                 'timeStmp': timeStmp,
 158                 'service': 'youtube',
 159                 'hl': 'en_US',
 160             }
 161             tfa_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k, v in tfa_form_strs.items())
 162             tfa_data = compat_urllib_parse.urlencode(tfa_form).encode('ascii')
 163
 164             tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data)
 165             tfa_results = self._download_webpage(
 166                 tfa_req, None,
 167                 note='Submitting TFA code', errnote='unable to submit tfa', fatal=False)
 168
 169             if tfa_results is False:
 170                 return False
 171
 172             if re.search(r'(?i)<form[^>]* id="gaia_secondfactorform"', tfa_results) is not None:
 173                 self._downloader.report_warning('Two-factor code expired. Please try again, or use a one-use backup code instead.')
 174                 return False
 175             if re.search(r'(?i)<form[^>]* id="gaia_loginform"', tfa_results) is not None:
 176                 self._downloader.report_warning('unable to log in - did the page structure change?')
 177                 return False
 178             if re.search(r'smsauth-interstitial-reviewsettings', tfa_results) is not None:
 179                 self._downloader.report_warning('Your Google account has a security notice. Please log in on your web browser, resolve the notice, and try again.')
 180                 return False
 181
 182         if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
 183             self._downloader.report_warning('unable to log in: bad username or password')
 184             return False
 185         return True
 186
 187     def _real_initialize(self):
 188         if self._downloader is None:
 189             return
 190         self._set_language()
 191         if not self._login():
 192             return
 193
 194
 195 class YoutubeIE(YoutubeBaseInfoExtractor):
 196     IE_DESC = 'YouTube.com'
 197     _VALID_URL = r"""(?x)^
 198                      (
 199                          (?:https?://|//)                                    # http(s):// or protocol-independent URL
 200                          (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/|
 201                             (?:www\.)?deturl\.com/www\.youtube\.com/|
 202                             (?:www\.)?pwnyoutube\.com/|
 203                             (?:www\.)?yourepeat\.com/|
 204                             tube\.majestyc\.net/|
 205                             youtube\.googleapis\.com/)                        # the various hostnames, with wildcard subdomains
 206                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
 207                          (?:                                                  # the various things that can precede the ID:
 208                              (?:(?:v|embed|e)/(?!videoseries))                # v/ or embed/ or e/
 209                              |(?:                                             # or the v= param in all its forms
 210                                  (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)?  # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
 211                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
 212                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
 213                                  v=
 214                              )
 215                          ))
 216                          |youtu\.be/                                          # just youtu.be/xxxx
 217                          |(?:www\.)?cleanvideosearch\.com/media/action/yt/watch\?videoId=
 218                          )
 219                      )?                                                       # all until now is optional -> you can pass the naked ID
 220                      ([0-9A-Za-z_-]{11})                                      # here is it! the YouTube video ID
 221                      (?!.*?&list=)                                            # combined list/video URLs are handled by the playlist IE
 222                      (?(1).+)?                                                # if we found the ID, everything can follow
 223                      $"""
 224     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
 225     _formats = {
 226         '5': {'ext': 'flv', 'width': 400, 'height': 240},
 227         '6': {'ext': 'flv', 'width': 450, 'height': 270},
 228         '13': {'ext': '3gp'},
 229         '17': {'ext': '3gp', 'width': 176, 'height': 144},
 230         '18': {'ext': 'mp4', 'width': 640, 'height': 360},
 231         '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
 232         '34': {'ext': 'flv', 'width': 640, 'height': 360},
 233         '35': {'ext': 'flv', 'width': 854, 'height': 480},
 234         '36': {'ext': '3gp', 'width': 320, 'height': 240},
 235         '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
 236         '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
 237         '43': {'ext': 'webm', 'width': 640, 'height': 360},
 238         '44': {'ext': 'webm', 'width': 854, 'height': 480},
 239         '45': {'ext': 'webm', 'width': 1280, 'height': 720},
 240         '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
 241         '59': {'ext': 'mp4', 'width': 854, 'height': 480},
 242         '78': {'ext': 'mp4', 'width': 854, 'height': 480},
 243
 244
 245         # 3d videos
 246         '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20},
 247         '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20},
 248         '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20},
 249         '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20},
 250         '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20},
 251         '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20},
 252         '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20},
 253
 254         # Apple HTTP Live Streaming
 255         '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 256         '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10},
 257         '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10},
 258         '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10},
 259         '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10},
 260         '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 261         '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10},
 262
 263         # DASH mp4 video
 264         '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 265         '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 266         '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 267         '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 268         '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 269         '138': {'ext': 'mp4', 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},  # Height can vary (https://github.com/rg3/youtube-dl/issues/4559)
 270         '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 271         '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 272         '298': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'},
 273         '299': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'},
 274         '266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'h264'},
 275
 276         # Dash mp4 audio
 277         '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 48, 'preference': -50, 'container': 'm4a_dash'},
 278         '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 128, 'preference': -50, 'container': 'm4a_dash'},
 279         '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 256, 'preference': -50, 'container': 'm4a_dash'},
 280
 281         # Dash webm
 282         '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 283         '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 284         '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 285         '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 286         '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 287         '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 288         '278': {'ext': 'webm', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'container': 'webm', 'vcodec': 'VP9'},
 289         '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 290         '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 291         '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 292         '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 293         '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 294         '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 295         '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 296         '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 297         '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 298         '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 299         '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 300         '308': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 301         '313': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'VP9'},
 302         '315': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'VP9'},
 303
 304         # Dash webm audio
 305         '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50},
 306         '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50},
 307
 308         # Dash webm audio with opus inside
 309         '249': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 50, 'preference': -50},
 310         '250': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 70, 'preference': -50},
 311         '251': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 160, 'preference': -50},
 312
 313         # RTMP (unnamed)
 314         '_rtmp': {'protocol': 'rtmp'},
 315     }
 316
 317     IE_NAME = 'youtube'
 318     _TESTS = [
 319         {
 320             'url': 'http://www.youtube.com/watch?v=BaW_jenozKc',
 321             'info_dict': {
 322                 'id': 'BaW_jenozKc',
 323                 'ext': 'mp4',
 324                 'title': 'youtube-dl test video "\'/\\ä↭𝕐',
 325                 'uploader': 'Philipp Hagemeister',
 326                 'uploader_id': 'phihag',
 327                 'upload_date': '20121002',
 328                 '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 .',
 329                 'categories': ['Science & Technology'],
 330                 'like_count': int,
 331                 'dislike_count': int,
 332             }
 333         },
 334         {
 335             'url': 'http://www.youtube.com/watch?v=UxxajLWwzqY',
 336             'note': 'Test generic use_cipher_signature video (#897)',
 337             'info_dict': {
 338                 'id': 'UxxajLWwzqY',
 339                 'ext': 'mp4',
 340                 'upload_date': '20120506',
 341                 'title': 'Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]',
 342                 'description': 'md5:fea86fda2d5a5784273df5c7cc994d9f',
 343                 'uploader': 'Icona Pop',
 344                 'uploader_id': 'IconaPop',
 345             }
 346         },
 347         {
 348             'url': 'https://www.youtube.com/watch?v=07FYdnEawAQ',
 349             'note': 'Test VEVO video with age protection (#956)',
 350             'info_dict': {
 351                 'id': '07FYdnEawAQ',
 352                 'ext': 'mp4',
 353                 'upload_date': '20130703',
 354                 'title': 'Justin Timberlake - Tunnel Vision (Explicit)',
 355                 'description': 'md5:64249768eec3bc4276236606ea996373',
 356                 'uploader': 'justintimberlakeVEVO',
 357                 'uploader_id': 'justintimberlakeVEVO',
 358             }
 359         },
 360         {
 361             'url': '//www.YouTube.com/watch?v=yZIXLfi8CZQ',
 362             'note': 'Embed-only video (#1746)',
 363             'info_dict': {
 364                 'id': 'yZIXLfi8CZQ',
 365                 'ext': 'mp4',
 366                 'upload_date': '20120608',
 367                 'title': 'Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012',
 368                 'description': 'md5:09b78bd971f1e3e289601dfba15ca4f7',
 369                 'uploader': 'SET India',
 370                 'uploader_id': 'setindia'
 371             }
 372         },
 373         {
 374             'url': 'http://www.youtube.com/watch?v=a9LDPn-MO4I',
 375             'note': '256k DASH audio (format 141) via DASH manifest',
 376             'info_dict': {
 377                 'id': 'a9LDPn-MO4I',
 378                 'ext': 'm4a',
 379                 'upload_date': '20121002',
 380                 'uploader_id': '8KVIDEO',
 381                 'description': '',
 382                 'uploader': '8KVIDEO',
 383                 'title': 'UHDTV TEST 8K VIDEO.mp4'
 384             },
 385             'params': {
 386                 'youtube_include_dash_manifest': True,
 387                 'format': '141',
 388             },
 389         },
 390         # DASH manifest with encrypted signature
 391         {
 392             'url': 'https://www.youtube.com/watch?v=IB3lcPjvWLA',
 393             'info_dict': {
 394                 'id': 'IB3lcPjvWLA',
 395                 'ext': 'm4a',
 396                 'title': 'Afrojack, Spree Wilson - The Spark ft. Spree Wilson',
 397                 'description': 'md5:12e7067fa6735a77bdcbb58cb1187d2d',
 398                 'uploader': 'AfrojackVEVO',
 399                 'uploader_id': 'AfrojackVEVO',
 400                 'upload_date': '20131011',
 401             },
 402             'params': {
 403                 'youtube_include_dash_manifest': True,
 404                 'format': '141',
 405             },
 406         },
 407         # JS player signature function name containing $
 408         {
 409             'url': 'https://www.youtube.com/watch?v=nfWlot6h_JM',
 410             'info_dict': {
 411                 'id': 'nfWlot6h_JM',
 412                 'ext': 'm4a',
 413                 'title': 'Taylor Swift - Shake It Off',
 414                 'description': 'md5:2acfda1b285bdd478ccec22f9918199d',
 415                 'uploader': 'TaylorSwiftVEVO',
 416                 'uploader_id': 'TaylorSwiftVEVO',
 417                 'upload_date': '20140818',
 418             },
 419             'params': {
 420                 'youtube_include_dash_manifest': True,
 421                 'format': '141',
 422             },
 423         },
 424         # Controversy video
 425         {
 426             'url': 'https://www.youtube.com/watch?v=T4XJQO3qol8',
 427             'info_dict': {
 428                 'id': 'T4XJQO3qol8',
 429                 'ext': 'mp4',
 430                 'upload_date': '20100909',
 431                 'uploader': 'The Amazing Atheist',
 432                 'uploader_id': 'TheAmazingAtheist',
 433                 'title': 'Burning Everyone\'s Koran',
 434                 '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',
 435             }
 436         },
 437         # Normal age-gate video (No vevo, embed allowed)
 438         {
 439             'url': 'http://youtube.com/watch?v=HtVdAasjOgU',
 440             'info_dict': {
 441                 'id': 'HtVdAasjOgU',
 442                 'ext': 'mp4',
 443                 'title': 'The Witcher 3: Wild Hunt - The Sword Of Destiny Trailer',
 444                 'description': 're:(?s).{100,}About the Game\n.*?The Witcher 3: Wild Hunt.{100,}',
 445                 'uploader': 'The Witcher',
 446                 'uploader_id': 'WitcherGame',
 447                 'upload_date': '20140605',
 448             },
 449         },
 450         # Age-gate video with encrypted signature
 451         {
 452             'url': 'http://www.youtube.com/watch?v=6kLq3WMV1nU',
 453             'info_dict': {
 454                 'id': '6kLq3WMV1nU',
 455                 'ext': 'mp4',
 456                 'title': 'Dedication To My Ex (Miss That) (Lyric Video)',
 457                 'description': 'md5:33765bb339e1b47e7e72b5490139bb41',
 458                 'uploader': 'LloydVEVO',
 459                 'uploader_id': 'LloydVEVO',
 460                 'upload_date': '20110629',
 461             },
 462         },
 463         # video_info is None (https://github.com/rg3/youtube-dl/issues/4421)
 464         {
 465             'url': '__2ABJjxzNo',
 466             'info_dict': {
 467                 'id': '__2ABJjxzNo',
 468                 'ext': 'mp4',
 469                 'upload_date': '20100430',
 470                 'uploader_id': 'deadmau5',
 471                 'description': 'md5:12c56784b8032162bb936a5f76d55360',
 472                 'uploader': 'deadmau5',
 473                 'title': 'Deadmau5 - Some Chords (HD)',
 474             },
 475             'expected_warnings': [
 476                 'DASH manifest missing',
 477             ]
 478         },
 479         # Olympics (https://github.com/rg3/youtube-dl/issues/4431)
 480         {
 481             'url': 'lqQg6PlCWgI',
 482             'info_dict': {
 483                 'id': 'lqQg6PlCWgI',
 484                 'ext': 'mp4',
 485                 'upload_date': '20120731',
 486                 'uploader_id': 'olympic',
 487                 'description': 'HO09  - Women -  GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games',
 488                 'uploader': 'Olympics',
 489                 'title': 'Hockey - Women -  GER-AUS - London 2012 Olympic Games',
 490             },
 491             'params': {
 492                 'skip_download': 'requires avconv',
 493             }
 494         },
 495         # Non-square pixels
 496         {
 497             'url': 'https://www.youtube.com/watch?v=_b-2C3KPAM0',
 498             'info_dict': {
 499                 'id': '_b-2C3KPAM0',
 500                 'ext': 'mp4',
 501                 'stretched_ratio': 16 / 9.,
 502                 'upload_date': '20110310',
 503                 'uploader_id': 'AllenMeow',
 504                 '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',
 505                 'uploader': '孫艾倫',
 506                 'title': '[A-made] 變態妍字幕版 太妍 我就是這樣的人',
 507             },
 508         },
 509         # url_encoded_fmt_stream_map is empty string
 510         {
 511             'url': 'qEJwOuvDf7I',
 512             'info_dict': {
 513                 'id': 'qEJwOuvDf7I',
 514                 'ext': 'mp4',
 515                 'title': 'Обсуждение судебной практики по выборам 14 сентября 2014 года в Санкт-Петербурге',
 516                 'description': '',
 517                 'upload_date': '20150404',
 518                 'uploader_id': 'spbelect',
 519                 'uploader': 'Наблюдатели Петербурга',
 520             },
 521             'params': {
 522                 'skip_download': 'requires avconv',
 523             }
 524         },
 525         # Extraction from multiple DASH manifests (https://github.com/rg3/youtube-dl/pull/6097)
 526         {
 527             'url': 'https://www.youtube.com/watch?v=FIl7x6_3R5Y',
 528             'info_dict': {
 529                 'id': 'FIl7x6_3R5Y',
 530                 'ext': 'mp4',
 531                 'title': 'md5:7b81415841e02ecd4313668cde88737a',
 532                 'description': 'md5:116377fd2963b81ec4ce64b542173306',
 533                 'upload_date': '20150625',
 534                 'uploader_id': 'dorappi2000',
 535                 'uploader': 'dorappi2000',
 536                 'formats': 'mincount:33',
 537             },
 538         },
 539         # DASH manifest with segment_list
 540         {
 541             'url': 'https://www.youtube.com/embed/CsmdDsKjzN8',
 542             'md5': '8ce563a1d667b599d21064e982ab9e31',
 543             'info_dict': {
 544                 'id': 'CsmdDsKjzN8',
 545                 'ext': 'mp4',
 546                 'upload_date': '20150501',  # According to '<meta itemprop="datePublished"', but in other places it's 20150510
 547                 'uploader': 'Airtek',
 548                 'description': 'Retransmisión en directo de la XVIII media maratón de Zaragoza.',
 549                 'uploader_id': 'UCzTzUmjXxxacNnL8I3m4LnQ',
 550                 'title': 'Retransmisión XVIII Media maratón Zaragoza 2015',
 551             },
 552             'params': {
 553                 'youtube_include_dash_manifest': True,
 554                 'format': '135',  # bestvideo
 555             }
 556         },
 557     ]
 558
 559     def __init__(self, *args, **kwargs):
 560         super(YoutubeIE, self).__init__(*args, **kwargs)
 561         self._player_cache = {}
 562
 563     def report_video_info_webpage_download(self, video_id):
 564         """Report attempt to download video info webpage."""
 565         self.to_screen('%s: Downloading video info webpage' % video_id)
 566
 567     def report_information_extraction(self, video_id):
 568         """Report attempt to extract video information."""
 569         self.to_screen('%s: Extracting video information' % video_id)
 570
 571     def report_unavailable_format(self, video_id, format):
 572         """Report extracted video URL."""
 573         self.to_screen('%s: Format %s not available' % (video_id, format))
 574
 575     def report_rtmp_download(self):
 576         """Indicate the download will use the RTMP protocol."""
 577         self.to_screen('RTMP download detected')
 578
 579     def _signature_cache_id(self, example_sig):
 580         """ Return a string representation of a signature """
 581         return '.'.join(compat_str(len(part)) for part in example_sig.split('.'))
 582
 583     def _extract_signature_function(self, video_id, player_url, example_sig):
 584         id_m = re.match(
 585             r'.*?-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.(?P<ext>[a-z]+)$',
 586             player_url)
 587         if not id_m:
 588             raise ExtractorError('Cannot identify player %r' % player_url)
 589         player_type = id_m.group('ext')
 590         player_id = id_m.group('id')
 591
 592         # Read from filesystem cache
 593         func_id = '%s_%s_%s' % (
 594             player_type, player_id, self._signature_cache_id(example_sig))
 595         assert os.path.basename(func_id) == func_id
 596
 597         cache_spec = self._downloader.cache.load('youtube-sigfuncs', func_id)
 598         if cache_spec is not None:
 599             return lambda s: ''.join(s[i] for i in cache_spec)
 600
 601         download_note = (
 602             'Downloading player %s' % player_url
 603             if self._downloader.params.get('verbose') else
 604             'Downloading %s player %s' % (player_type, player_id)
 605         )
 606         if player_type == 'js':
 607             code = self._download_webpage(
 608                 player_url, video_id,
 609                 note=download_note,
 610                 errnote='Download of %s failed' % player_url)
 611             res = self._parse_sig_js(code)
 612         elif player_type == 'swf':
 613             urlh = self._request_webpage(
 614                 player_url, video_id,
 615                 note=download_note,
 616                 errnote='Download of %s failed' % player_url)
 617             code = urlh.read()
 618             res = self._parse_sig_swf(code)
 619         else:
 620             assert False, 'Invalid player type %r' % player_type
 621
 622         test_string = ''.join(map(compat_chr, range(len(example_sig))))
 623         cache_res = res(test_string)
 624         cache_spec = [ord(c) for c in cache_res]
 625
 626         self._downloader.cache.store('youtube-sigfuncs', func_id, cache_spec)
 627         return res
 628
 629     def _print_sig_code(self, func, example_sig):
 630         def gen_sig_code(idxs):
 631             def _genslice(start, end, step):
 632                 starts = '' if start == 0 else str(start)
 633                 ends = (':%d' % (end + step)) if end + step >= 0 else ':'
 634                 steps = '' if step == 1 else (':%d' % step)
 635                 return 's[%s%s%s]' % (starts, ends, steps)
 636
 637             step = None
 638             # Quelch pyflakes warnings - start will be set when step is set
 639             start = '(Never used)'
 640             for i, prev in zip(idxs[1:], idxs[:-1]):
 641                 if step is not None:
 642                     if i - prev == step:
 643                         continue
 644                     yield _genslice(start, prev, step)
 645                     step = None
 646                     continue
 647                 if i - prev in [-1, 1]:
 648                     step = i - prev
 649                     start = prev
 650                     continue
 651                 else:
 652                     yield 's[%d]' % prev
 653             if step is None:
 654                 yield 's[%d]' % i
 655             else:
 656                 yield _genslice(start, i, step)
 657
 658         test_string = ''.join(map(compat_chr, range(len(example_sig))))
 659         cache_res = func(test_string)
 660         cache_spec = [ord(c) for c in cache_res]
 661         expr_code = ' + '.join(gen_sig_code(cache_spec))
 662         signature_id_tuple = '(%s)' % (
 663             ', '.join(compat_str(len(p)) for p in example_sig.split('.')))
 664         code = ('if tuple(len(p) for p in s.split(\'.\')) == %s:\n'
 665                 '    return %s\n') % (signature_id_tuple, expr_code)
 666         self.to_screen('Extracted signature function:\n' + code)
 667
 668     def _parse_sig_js(self, jscode):
 669         funcname = self._search_regex(
 670             r'\.sig\|\|([a-zA-Z0-9$]+)\(', jscode,
 671             'Initial JS player signature function name')
 672
 673         jsi = JSInterpreter(jscode)
 674         initial_function = jsi.extract_function(funcname)
 675         return lambda s: initial_function([s])
 676
 677     def _parse_sig_swf(self, file_contents):
 678         swfi = SWFInterpreter(file_contents)
 679         TARGET_CLASSNAME = 'SignatureDecipher'
 680         searched_class = swfi.extract_class(TARGET_CLASSNAME)
 681         initial_function = swfi.extract_function(searched_class, 'decipher')
 682         return lambda s: initial_function([s])
 683
 684     def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
 685         """Turn the encrypted s field into a working signature"""
 686
 687         if player_url is None:
 688             raise ExtractorError('Cannot decrypt signature without player_url')
 689
 690         if player_url.startswith('//'):
 691             player_url = 'https:' + player_url
 692         try:
 693             player_id = (player_url, self._signature_cache_id(s))
 694             if player_id not in self._player_cache:
 695                 func = self._extract_signature_function(
 696                     video_id, player_url, s
 697                 )
 698                 self._player_cache[player_id] = func
 699             func = self._player_cache[player_id]
 700             if self._downloader.params.get('youtube_print_sig_code'):
 701                 self._print_sig_code(func, s)
 702             return func(s)
 703         except Exception as e:
 704             tb = traceback.format_exc()
 705             raise ExtractorError(
 706                 'Signature extraction failed: ' + tb, cause=e)
 707
 708     def _get_subtitles(self, video_id, webpage):
 709         try:
 710             subs_doc = self._download_xml(
 711                 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
 712                 video_id, note=False)
 713         except ExtractorError as err:
 714             self._downloader.report_warning('unable to download video subtitles: %s' % compat_str(err))
 715             return {}
 716
 717         sub_lang_list = {}
 718         for track in subs_doc.findall('track'):
 719             lang = track.attrib['lang_code']
 720             if lang in sub_lang_list:
 721                 continue
 722             sub_formats = []
 723             for ext in ['sbv', 'vtt', 'srt']:
 724                 params = compat_urllib_parse.urlencode({
 725                     'lang': lang,
 726                     'v': video_id,
 727                     'fmt': ext,
 728                     'name': track.attrib['name'].encode('utf-8'),
 729                 })
 730                 sub_formats.append({
 731                     'url': 'https://www.youtube.com/api/timedtext?' + params,
 732                     'ext': ext,
 733                 })
 734             sub_lang_list[lang] = sub_formats
 735         if not sub_lang_list:
 736             self._downloader.report_warning('video doesn\'t have subtitles')
 737             return {}
 738         return sub_lang_list
 739
 740     def _get_automatic_captions(self, video_id, webpage):
 741         """We need the webpage for getting the captions url, pass it as an
 742            argument to speed up the process."""
 743         self.to_screen('%s: Looking for automatic captions' % video_id)
 744         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
 745         err_msg = 'Couldn\'t find automatic captions for %s' % video_id
 746         if mobj is None:
 747             self._downloader.report_warning(err_msg)
 748             return {}
 749         player_config = json.loads(mobj.group(1))
 750         try:
 751             args = player_config['args']
 752             caption_url = args['ttsurl']
 753             timestamp = args['timestamp']
 754             # We get the available subtitles
 755             list_params = compat_urllib_parse.urlencode({
 756                 'type': 'list',
 757                 'tlangs': 1,
 758                 'asrs': 1,
 759             })
 760             list_url = caption_url + '&' + list_params
 761             caption_list = self._download_xml(list_url, video_id)
 762             original_lang_node = caption_list.find('track')
 763             if original_lang_node is None:
 764                 self._downloader.report_warning('Video doesn\'t have automatic captions')
 765                 return {}
 766             original_lang = original_lang_node.attrib['lang_code']
 767             caption_kind = original_lang_node.attrib.get('kind', '')
 768
 769             sub_lang_list = {}
 770             for lang_node in caption_list.findall('target'):
 771                 sub_lang = lang_node.attrib['lang_code']
 772                 sub_formats = []
 773                 for ext in ['sbv', 'vtt', 'srt']:
 774                     params = compat_urllib_parse.urlencode({
 775                         'lang': original_lang,
 776                         'tlang': sub_lang,
 777                         'fmt': ext,
 778                         'ts': timestamp,
 779                         'kind': caption_kind,
 780                     })
 781                     sub_formats.append({
 782                         'url': caption_url + '&' + params,
 783                         'ext': ext,
 784                     })
 785                 sub_lang_list[sub_lang] = sub_formats
 786             return sub_lang_list
 787         # An extractor error can be raise by the download process if there are
 788         # no automatic captions but there are subtitles
 789         except (KeyError, ExtractorError):
 790             self._downloader.report_warning(err_msg)
 791             return {}
 792
 793     @classmethod
 794     def extract_id(cls, url):
 795         mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
 796         if mobj is None:
 797             raise ExtractorError('Invalid URL: %s' % url)
 798         video_id = mobj.group(2)
 799         return video_id
 800
 801     def _extract_from_m3u8(self, manifest_url, video_id):
 802         url_map = {}
 803
 804         def _get_urls(_manifest):
 805             lines = _manifest.split('\n')
 806             urls = filter(lambda l: l and not l.startswith('#'),
 807                           lines)
 808             return urls
 809         manifest = self._download_webpage(manifest_url, video_id, 'Downloading formats manifest')
 810         formats_urls = _get_urls(manifest)
 811         for format_url in formats_urls:
 812             itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
 813             url_map[itag] = format_url
 814         return url_map
 815
 816     def _extract_annotations(self, video_id):
 817         url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
 818         return self._download_webpage(url, video_id, note='Searching for annotations.', errnote='Unable to download video annotations.')
 819
 820     def _parse_dash_manifest(
 821             self, video_id, dash_manifest_url, player_url, age_gate, fatal=True):
 822         def decrypt_sig(mobj):
 823             s = mobj.group(1)
 824             dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
 825             return '/signature/%s' % dec_s
 826         dash_manifest_url = re.sub(r'/s/([a-fA-F0-9\.]+)', decrypt_sig, dash_manifest_url)
 827         dash_doc = self._download_xml(
 828             dash_manifest_url, video_id,
 829             note='Downloading DASH manifest',
 830             errnote='Could not download DASH manifest',
 831             fatal=fatal)
 832
 833         if dash_doc is False:
 834             return []
 835
 836         formats = []
 837         for a in dash_doc.findall('.//{urn:mpeg:DASH:schema:MPD:2011}AdaptationSet'):
 838             mime_type = a.attrib.get('mimeType')
 839             for r in a.findall('{urn:mpeg:DASH:schema:MPD:2011}Representation'):
 840                 url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL')
 841                 if url_el is None:
 842                     continue
 843                 if mime_type == 'text/vtt':
 844                     # TODO implement WebVTT downloading
 845                     pass
 846                 elif mime_type.startswith('audio/') or mime_type.startswith('video/'):
 847                     segment_list = r.find('{urn:mpeg:DASH:schema:MPD:2011}SegmentList')
 848                     format_id = r.attrib['id']
 849                     video_url = url_el.text
 850                     filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
 851                     f = {
 852                         'format_id': format_id,
 853                         'url': video_url,
 854                         'width': int_or_none(r.attrib.get('width')),
 855                         'height': int_or_none(r.attrib.get('height')),
 856                         'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
 857                         'asr': int_or_none(r.attrib.get('audioSamplingRate')),
 858                         'filesize': filesize,
 859                         'fps': int_or_none(r.attrib.get('frameRate')),
 860                     }
 861                     if segment_list is not None:
 862                         f.update({
 863                             'initialization_url': segment_list.find('{urn:mpeg:DASH:schema:MPD:2011}Initialization').attrib['sourceURL'],
 864                             'segment_urls': [segment.attrib.get('media') for segment in segment_list.findall('{urn:mpeg:DASH:schema:MPD:2011}SegmentURL')],
 865                             'protocol': 'http_dash_segments',
 866                         })
 867                     try:
 868                         existing_format = next(
 869                             fo for fo in formats
 870                             if fo['format_id'] == format_id)
 871                     except StopIteration:
 872                         full_info = self._formats.get(format_id, {}).copy()
 873                         full_info.update(f)
 874                         codecs = r.attrib.get('codecs')
 875                         if codecs:
 876                             if full_info.get('acodec') == 'none' and 'vcodec' not in full_info:
 877                                 full_info['vcodec'] = codecs
 878                             elif full_info.get('vcodec') == 'none' and 'acodec' not in full_info:
 879                                 full_info['acodec'] = codecs
 880                         formats.append(full_info)
 881                     else:
 882                         existing_format.update(f)
 883                 else:
 884                     self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
 885         return formats
 886
 887     def _real_extract(self, url):
 888         proto = (
 889             'http' if self._downloader.params.get('prefer_insecure', False)
 890             else 'https')
 891
 892         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
 893         mobj = re.search(self._NEXT_URL_RE, url)
 894         if mobj:
 895             url = proto + '://www.youtube.com/' + compat_urllib_parse_unquote(mobj.group(1)).lstrip('/')
 896         video_id = self.extract_id(url)
 897
 898         # Get video webpage
 899         url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1&bpctr=9999999999' % video_id
 900         video_webpage = self._download_webpage(url, video_id)
 901
 902         # Attempt to extract SWF player URL
 903         mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
 904         if mobj is not None:
 905             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
 906         else:
 907             player_url = None
 908
 909         dash_mpds = []
 910
 911         def add_dash_mpd(video_info):
 912             dash_mpd = video_info.get('dashmpd')
 913             if dash_mpd and dash_mpd[0] not in dash_mpds:
 914                 dash_mpds.append(dash_mpd[0])
 915
 916         # Get video info
 917         embed_webpage = None
 918         if re.search(r'player-age-gate-content">', video_webpage) is not None:
 919             age_gate = True
 920             # We simulate the access to the video from www.youtube.com/v/{video_id}
 921             # this can be viewed without login into Youtube
 922             url = proto + '://www.youtube.com/embed/%s' % video_id
 923             embed_webpage = self._download_webpage(url, video_id, 'Downloading embed webpage')
 924             data = compat_urllib_parse.urlencode({
 925                 'video_id': video_id,
 926                 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
 927                 'sts': self._search_regex(
 928                     r'"sts"\s*:\s*(\d+)', embed_webpage, 'sts', default=''),
 929             })
 930             video_info_url = proto + '://www.youtube.com/get_video_info?' + data
 931             video_info_webpage = self._download_webpage(
 932                 video_info_url, video_id,
 933                 note='Refetching age-gated info webpage',
 934                 errnote='unable to download video info webpage')
 935             video_info = compat_parse_qs(video_info_webpage)
 936             add_dash_mpd(video_info)
 937         else:
 938             age_gate = False
 939             video_info = None
 940             # Try looking directly into the video webpage
 941             mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage)
 942             if mobj:
 943                 json_code = uppercase_escape(mobj.group(1))
 944                 ytplayer_config = json.loads(json_code)
 945                 args = ytplayer_config['args']
 946                 if args.get('url_encoded_fmt_stream_map'):
 947                     # Convert to the same format returned by compat_parse_qs
 948                     video_info = dict((k, [v]) for k, v in args.items())
 949                     add_dash_mpd(video_info)
 950             if not video_info or self._downloader.params.get('youtube_include_dash_manifest', True):
 951                 # We also try looking in get_video_info since it may contain different dashmpd
 952                 # URL that points to a DASH manifest with possibly different itag set (some itags
 953                 # are missing from DASH manifest pointed by webpage's dashmpd, some - from DASH
 954                 # manifest pointed by get_video_info's dashmpd).
 955                 # The general idea is to take a union of itags of both DASH manifests (for example
 956                 # video with such 'manifest behavior' see https://github.com/rg3/youtube-dl/issues/6093)
 957                 self.report_video_info_webpage_download(video_id)
 958                 for el_type in ['&el=info', '&el=embedded', '&el=detailpage', '&el=vevo', '']:
 959                     video_info_url = (
 960                         '%s://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
 961                         % (proto, video_id, el_type))
 962                     video_info_webpage = self._download_webpage(
 963                         video_info_url,
 964                         video_id, note=False,
 965                         errnote='unable to download video info webpage')
 966                     get_video_info = compat_parse_qs(video_info_webpage)
 967                     add_dash_mpd(get_video_info)
 968                     if not video_info:
 969                         video_info = get_video_info
 970                     if 'token' in get_video_info:
 971                         break
 972         if 'token' not in video_info:
 973             if 'reason' in video_info:
 974                 if 'The uploader has not made this video available in your country.' in video_info['reason']:
 975                     regions_allowed = self._html_search_meta('regionsAllowed', video_webpage, default=None)
 976                     if regions_allowed is not None:
 977                         raise ExtractorError('YouTube said: This video is available in %s only' % (
 978                             ', '.join(map(ISO3166Utils.short2full, regions_allowed.split(',')))),
 979                             expected=True)
 980                 raise ExtractorError(
 981                     'YouTube said: %s' % video_info['reason'][0],
 982                     expected=True, video_id=video_id)
 983             else:
 984                 raise ExtractorError(
 985                     '"token" parameter not in video info for unknown reason',
 986                     video_id=video_id)
 987
 988         if 'view_count' in video_info:
 989             view_count = int(video_info['view_count'][0])
 990         else:
 991             view_count = None
 992
 993         # Check for "rental" videos
 994         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
 995             raise ExtractorError('"rental" videos not supported')
 996
 997         # Start extracting information
 998         self.report_information_extraction(video_id)
 999
1000         # uploader
1001         if 'author' not in video_info:
1002             raise ExtractorError('Unable to extract uploader name')
1003         video_uploader = compat_urllib_parse_unquote_plus(video_info['author'][0])
1004
1005         # uploader_id
1006         video_uploader_id = None
1007         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
1008         if mobj is not None:
1009             video_uploader_id = mobj.group(1)
1010         else:
1011             self._downloader.report_warning('unable to extract uploader nickname')
1012
1013         # title
1014         if 'title' in video_info:
1015             video_title = video_info['title'][0]
1016         else:
1017             self._downloader.report_warning('Unable to extract video title')
1018             video_title = '_'
1019
1020         # thumbnail image
1021         # We try first to get a high quality image:
1022         m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
1023                             video_webpage, re.DOTALL)
1024         if m_thumb is not None:
1025             video_thumbnail = m_thumb.group(1)
1026         elif 'thumbnail_url' not in video_info:
1027             self._downloader.report_warning('unable to extract video thumbnail')
1028             video_thumbnail = None
1029         else:   # don't panic if we can't find it
1030             video_thumbnail = compat_urllib_parse_unquote_plus(video_info['thumbnail_url'][0])
1031
1032         # upload date
1033         upload_date = self._html_search_meta(
1034             'datePublished', video_webpage, 'upload date', default=None)
1035         if not upload_date:
1036             upload_date = self._search_regex(
1037                 [r'(?s)id="eow-date.*?>(.*?)</span>',
1038                  r'id="watch-uploader-info".*?>.*?(?:Published|Uploaded|Streamed live|Started) on (.+?)</strong>'],
1039                 video_webpage, 'upload date', default=None)
1040             if upload_date:
1041                 upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
1042         upload_date = unified_strdate(upload_date)
1043
1044         m_cat_container = self._search_regex(
1045             r'(?s)<h4[^>]*>\s*Category\s*</h4>\s*<ul[^>]*>(.*?)</ul>',
1046             video_webpage, 'categories', default=None)
1047         if m_cat_container:
1048             category = self._html_search_regex(
1049                 r'(?s)<a[^<]+>(.*?)</a>', m_cat_container, 'category',
1050                 default=None)
1051             video_categories = None if category is None else [category]
1052         else:
1053             video_categories = None
1054
1055         # description
1056         video_description = get_element_by_id("eow-description", video_webpage)
1057         if video_description:
1058             video_description = re.sub(r'''(?x)
1059                 <a\s+
1060                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1061                     title="([^"]+)"\s+
1062                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1063                     class="yt-uix-redirect-link"\s*>
1064                 [^<]+
1065                 </a>
1066             ''', r'\1', video_description)
1067             video_description = clean_html(video_description)
1068         else:
1069             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
1070             if fd_mobj:
1071                 video_description = unescapeHTML(fd_mobj.group(1))
1072             else:
1073                 video_description = ''
1074
1075         def _extract_count(count_name):
1076             return str_to_int(self._search_regex(
1077                 r'-%s-button[^>]+><span[^>]+class="yt-uix-button-content"[^>]*>([\d,]+)</span>'
1078                 % re.escape(count_name),
1079                 video_webpage, count_name, default=None))
1080
1081         like_count = _extract_count('like')
1082         dislike_count = _extract_count('dislike')
1083
1084         # subtitles
1085         video_subtitles = self.extract_subtitles(video_id, video_webpage)
1086         automatic_captions = self.extract_automatic_captions(video_id, video_webpage)
1087
1088         if 'length_seconds' not in video_info:
1089             self._downloader.report_warning('unable to extract video duration')
1090             video_duration = None
1091         else:
1092             video_duration = int(compat_urllib_parse_unquote_plus(video_info['length_seconds'][0]))
1093
1094         # annotations
1095         video_annotations = None
1096         if self._downloader.params.get('writeannotations', False):
1097             video_annotations = self._extract_annotations(video_id)
1098
1099         def _map_to_format_list(urlmap):
1100             formats = []
1101             for itag, video_real_url in urlmap.items():
1102                 dct = {
1103                     'format_id': itag,
1104                     'url': video_real_url,
1105                     'player_url': player_url,
1106                 }
1107                 if itag in self._formats:
1108                     dct.update(self._formats[itag])
1109                 formats.append(dct)
1110             return formats
1111
1112         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
1113             self.report_rtmp_download()
1114             formats = [{
1115                 'format_id': '_rtmp',
1116                 'protocol': 'rtmp',
1117                 'url': video_info['conn'][0],
1118                 'player_url': player_url,
1119             }]
1120         elif len(video_info.get('url_encoded_fmt_stream_map', [''])[0]) >= 1 or len(video_info.get('adaptive_fmts', [''])[0]) >= 1:
1121             encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts', [''])[0]
1122             if 'rtmpe%3Dyes' in encoded_url_map:
1123                 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
1124             url_map = {}
1125             for url_data_str in encoded_url_map.split(','):
1126                 url_data = compat_parse_qs(url_data_str)
1127                 if 'itag' not in url_data or 'url' not in url_data:
1128                     continue
1129                 format_id = url_data['itag'][0]
1130                 url = url_data['url'][0]
1131
1132                 if 'sig' in url_data:
1133                     url += '&signature=' + url_data['sig'][0]
1134                 elif 's' in url_data:
1135                     encrypted_sig = url_data['s'][0]
1136                     ASSETS_RE = r'"assets":.+?"js":\s*("[^"]+")'
1137
1138                     jsplayer_url_json = self._search_regex(
1139                         ASSETS_RE,
1140                         embed_webpage if age_gate else video_webpage,
1141                         'JS player URL (1)', default=None)
1142                     if not jsplayer_url_json and not age_gate:
1143                         # We need the embed website after all
1144                         if embed_webpage is None:
1145                             embed_url = proto + '://www.youtube.com/embed/%s' % video_id
1146                             embed_webpage = self._download_webpage(
1147                                 embed_url, video_id, 'Downloading embed webpage')
1148                         jsplayer_url_json = self._search_regex(
1149                             ASSETS_RE, embed_webpage, 'JS player URL')
1150
1151                     player_url = json.loads(jsplayer_url_json)
1152                     if player_url is None:
1153                         player_url_json = self._search_regex(
1154                             r'ytplayer\.config.*?"url"\s*:\s*("[^"]+")',
1155                             video_webpage, 'age gate player URL')
1156                         player_url = json.loads(player_url_json)
1157
1158                     if self._downloader.params.get('verbose'):
1159                         if player_url is None:
1160                             player_version = 'unknown'
1161                             player_desc = 'unknown'
1162                         else:
1163                             if player_url.endswith('swf'):
1164                                 player_version = self._search_regex(
1165                                     r'-(.+?)(?:/watch_as3)?\.swf$', player_url,
1166                                     'flash player', fatal=False)
1167                                 player_desc = 'flash player %s' % player_version
1168                             else:
1169                                 player_version = self._search_regex(
1170                                     r'html5player-([^/]+?)(?:/html5player)?\.js',
1171                                     player_url,
1172                                     'html5 player', fatal=False)
1173                                 player_desc = 'html5 player %s' % player_version
1174
1175                         parts_sizes = self._signature_cache_id(encrypted_sig)
1176                         self.to_screen('{%s} signature length %s, %s' %
1177                                        (format_id, parts_sizes, player_desc))
1178
1179                     signature = self._decrypt_signature(
1180                         encrypted_sig, video_id, player_url, age_gate)
1181                     url += '&signature=' + signature
1182                 if 'ratebypass' not in url:
1183                     url += '&ratebypass=yes'
1184                 url_map[format_id] = url
1185             formats = _map_to_format_list(url_map)
1186         elif video_info.get('hlsvp'):
1187             manifest_url = video_info['hlsvp'][0]
1188             url_map = self._extract_from_m3u8(manifest_url, video_id)
1189             formats = _map_to_format_list(url_map)
1190         else:
1191             raise ExtractorError('no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
1192
1193         # Look for the DASH manifest
1194         if self._downloader.params.get('youtube_include_dash_manifest', True):
1195             dash_mpd_fatal = True
1196             for dash_manifest_url in dash_mpds:
1197                 dash_formats = {}
1198                 try:
1199                     for df in self._parse_dash_manifest(
1200                             video_id, dash_manifest_url, player_url, age_gate, dash_mpd_fatal):
1201                         # Do not overwrite DASH format found in some previous DASH manifest
1202                         if df['format_id'] not in dash_formats:
1203                             dash_formats[df['format_id']] = df
1204                         # Additional DASH manifests may end up in HTTP Error 403 therefore
1205                         # allow them to fail without bug report message if we already have
1206                         # some DASH manifest succeeded. This is temporary workaround to reduce
1207                         # burst of bug reports until we figure out the reason and whether it
1208                         # can be fixed at all.
1209                         dash_mpd_fatal = False
1210                 except (ExtractorError, KeyError) as e:
1211                     self.report_warning(
1212                         'Skipping DASH manifest: %r' % e, video_id)
1213                 if dash_formats:
1214                     # Remove the formats we found through non-DASH, they
1215                     # contain less info and it can be wrong, because we use
1216                     # fixed values (for example the resolution). See
1217                     # https://github.com/rg3/youtube-dl/issues/5774 for an
1218                     # example.
1219                     formats = [f for f in formats if f['format_id'] not in dash_formats.keys()]
1220                     formats.extend(dash_formats.values())
1221
1222         # Check for malformed aspect ratio
1223         stretched_m = re.search(
1224             r'<meta\s+property="og:video:tag".*?content="yt:stretch=(?P<w>[0-9]+):(?P<h>[0-9]+)">',
1225             video_webpage)
1226         if stretched_m:
1227             ratio = float(stretched_m.group('w')) / float(stretched_m.group('h'))
1228             for f in formats:
1229                 if f.get('vcodec') != 'none':
1230                     f['stretched_ratio'] = ratio
1231
1232         self._sort_formats(formats)
1233
1234         return {
1235             'id': video_id,
1236             'uploader': video_uploader,
1237             'uploader_id': video_uploader_id,
1238             'upload_date': upload_date,
1239             'title': video_title,
1240             'thumbnail': video_thumbnail,
1241             'description': video_description,
1242             'categories': video_categories,
1243             'subtitles': video_subtitles,
1244             'automatic_captions': automatic_captions,
1245             'duration': video_duration,
1246             'age_limit': 18 if age_gate else 0,
1247             'annotations': video_annotations,
1248             'webpage_url': proto + '://www.youtube.com/watch?v=%s' % video_id,
1249             'view_count': view_count,
1250             'like_count': like_count,
1251             'dislike_count': dislike_count,
1252             'average_rating': float_or_none(video_info.get('avg_rating', [None])[0]),
1253             'formats': formats,
1254         }
1255
1256
1257 class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
1258     IE_DESC = 'YouTube.com playlists'
1259     _VALID_URL = r"""(?x)(?:
1260                         (?:https?://)?
1261                         (?:\w+\.)?
1262                         youtube\.com/
1263                         (?:
1264                            (?:course|view_play_list|my_playlists|artist|playlist|watch|embed/videoseries)
1265                            \? (?:.*?&)*? (?:p|a|list)=
1266                         |  p/
1267                         )
1268                         (
1269                             (?:PL|LL|EC|UU|FL|RD|UL)?[0-9A-Za-z-_]{10,}
1270                             # Top tracks, they can also include dots
1271                             |(?:MC)[\w\.]*
1272                         )
1273                         .*
1274                      |
1275                         ((?:PL|LL|EC|UU|FL|RD|UL)[0-9A-Za-z-_]{10,})
1276                      )"""
1277     _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s'
1278     _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&amp;[^"]*?index=(?P<index>\d+)'
1279     IE_NAME = 'youtube:playlist'
1280     _TESTS = [{
1281         'url': 'https://www.youtube.com/playlist?list=PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re',
1282         'info_dict': {
1283             'title': 'ytdl test PL',
1284             'id': 'PLwiyx1dc3P2JR9N8gQaQN_BCvlSlap7re',
1285         },
1286         'playlist_count': 3,
1287     }, {
1288         'url': 'https://www.youtube.com/playlist?list=PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx',
1289         'info_dict': {
1290             'id': 'PLtPgu7CB4gbZDA7i_euNxn75ISqxwZPYx',
1291             'title': 'YDL_Empty_List',
1292         },
1293         'playlist_count': 0,
1294     }, {
1295         'note': 'Playlist with deleted videos (#651). As a bonus, the video #51 is also twice in this list.',
1296         'url': 'https://www.youtube.com/playlist?list=PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC',
1297         'info_dict': {
1298             'title': '29C3: Not my department',
1299             'id': 'PLwP_SiAcdui0KVebT0mU9Apz359a4ubsC',
1300         },
1301         'playlist_count': 95,
1302     }, {
1303         'note': 'issue #673',
1304         'url': 'PLBB231211A4F62143',
1305         'info_dict': {
1306             'title': '[OLD]Team Fortress 2 (Class-based LP)',
1307             'id': 'PLBB231211A4F62143',
1308         },
1309         'playlist_mincount': 26,
1310     }, {
1311         'note': 'Large playlist',
1312         'url': 'https://www.youtube.com/playlist?list=UUBABnxM4Ar9ten8Mdjj1j0Q',
1313         'info_dict': {
1314             'title': 'Uploads from Cauchemar',
1315             'id': 'UUBABnxM4Ar9ten8Mdjj1j0Q',
1316         },
1317         'playlist_mincount': 799,
1318     }, {
1319         'url': 'PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl',
1320         'info_dict': {
1321             'title': 'YDL_safe_search',
1322             'id': 'PLtPgu7CB4gbY9oDN3drwC3cMbJggS7dKl',
1323         },
1324         'playlist_count': 2,
1325     }, {
1326         'note': 'embedded',
1327         'url': 'http://www.youtube.com/embed/videoseries?list=PL6IaIsEjSbf96XFRuNccS_RuEXwNdsoEu',
1328         'playlist_count': 4,
1329         'info_dict': {
1330             'title': 'JODA15',
1331             'id': 'PL6IaIsEjSbf96XFRuNccS_RuEXwNdsoEu',
1332         }
1333     }, {
1334         'note': 'Embedded SWF player',
1335         'url': 'http://www.youtube.com/p/YN5VISEtHet5D4NEvfTd0zcgFk84NqFZ?hl=en_US&fs=1&rel=0',
1336         'playlist_count': 4,
1337         'info_dict': {
1338             'title': 'JODA7',
1339             'id': 'YN5VISEtHet5D4NEvfTd0zcgFk84NqFZ',
1340         }
1341     }, {
1342         'note': 'Buggy playlist: the webpage has a "Load more" button but it doesn\'t have more videos',
1343         'url': 'https://www.youtube.com/playlist?list=UUXw-G3eDE9trcvY2sBMM_aA',
1344         'info_dict': {
1345             'title': 'Uploads from Interstellar Movie',
1346             'id': 'UUXw-G3eDE9trcvY2sBMM_aA',
1347         },
1348         'playlist_mincout': 21,
1349     }]
1350
1351     def _real_initialize(self):
1352         self._login()
1353
1354     def _extract_mix(self, playlist_id):
1355         # The mixes are generated from a single video
1356         # the id of the playlist is just 'RD' + video_id
1357         url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id)
1358         webpage = self._download_webpage(
1359             url, playlist_id, 'Downloading Youtube mix')
1360         search_title = lambda class_name: get_element_by_attribute('class', class_name, webpage)
1361         title_span = (
1362             search_title('playlist-title') or
1363             search_title('title long-title') or
1364             search_title('title'))
1365         title = clean_html(title_span)
1366         ids = orderedSet(re.findall(
1367             r'''(?xs)data-video-username=".*?".*?
1368                        href="/watch\?v=([0-9A-Za-z_-]{11})&amp;[^"]*?list=%s''' % re.escape(playlist_id),
1369             webpage))
1370         url_results = self._ids_to_results(ids)
1371
1372         return self.playlist_result(url_results, playlist_id, title)
1373
1374     def _extract_playlist(self, playlist_id):
1375         url = self._TEMPLATE_URL % playlist_id
1376         page = self._download_webpage(url, playlist_id)
1377
1378         for match in re.findall(r'<div class="yt-alert-message">([^<]+)</div>', page):
1379             match = match.strip()
1380             # Check if the playlist exists or is private
1381             if re.match(r'[^<]*(The|This) playlist (does not exist|is private)[^<]*', match):
1382                 raise ExtractorError(
1383                     'The playlist doesn\'t exist or is private, use --username or '
1384                     '--netrc to access it.',
1385                     expected=True)
1386             elif re.match(r'[^<]*Invalid parameters[^<]*', match):
1387                 raise ExtractorError(
1388                     'Invalid parameters. Maybe URL is incorrect.',
1389                     expected=True)
1390             elif re.match(r'[^<]*Choose your language[^<]*', match):
1391                 continue
1392             else:
1393                 self.report_warning('Youtube gives an alert message: ' + match)
1394
1395         # Extract the video ids from the playlist pages
1396         def _entries():
1397             more_widget_html = content_html = page
1398             for page_num in itertools.count(1):
1399                 matches = re.finditer(self._VIDEO_RE, content_html)
1400                 # We remove the duplicates and the link with index 0
1401                 # (it's not the first video of the playlist)
1402                 new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
1403                 for vid_id in new_ids:
1404                     yield self.url_result(vid_id, 'Youtube', video_id=vid_id)
1405
1406                 mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
1407                 if not mobj:
1408                     break
1409
1410                 more = self._download_json(
1411                     'https://youtube.com/%s' % mobj.group('more'), playlist_id,
1412                     'Downloading page #%s' % page_num,
1413                     transform_source=uppercase_escape)
1414                 content_html = more['content_html']
1415                 if not content_html.strip():
1416                     # Some webpages show a "Load more" button but they don't
1417                     # have more videos
1418                     break
1419                 more_widget_html = more['load_more_widget_html']
1420
1421         playlist_title = self._html_search_regex(
1422             r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>',
1423             page, 'title')
1424
1425         return self.playlist_result(_entries(), playlist_id, playlist_title)
1426
1427     def _real_extract(self, url):
1428         # Extract playlist id
1429         mobj = re.match(self._VALID_URL, url)
1430         if mobj is None:
1431             raise ExtractorError('Invalid URL: %s' % url)
1432         playlist_id = mobj.group(1) or mobj.group(2)
1433
1434         # Check if it's a video-specific URL
1435         query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1436         if 'v' in query_dict:
1437             video_id = query_dict['v'][0]
1438             if self._downloader.params.get('noplaylist'):
1439                 self.to_screen('Downloading just video %s because of --no-playlist' % video_id)
1440                 return self.url_result(video_id, 'Youtube', video_id=video_id)
1441             else:
1442                 self.to_screen('Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1443
1444         if playlist_id.startswith('RD') or playlist_id.startswith('UL'):
1445             # Mixes require a custom extraction process
1446             return self._extract_mix(playlist_id)
1447
1448         return self._extract_playlist(playlist_id)
1449
1450
1451 class YoutubeChannelIE(InfoExtractor):
1452     IE_DESC = 'YouTube.com channels'
1453     _VALID_URL = r'https?://(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/(?P<id>[0-9A-Za-z_-]+)'
1454     _TEMPLATE_URL = 'https://www.youtube.com/channel/%s/videos'
1455     IE_NAME = 'youtube:channel'
1456     _TESTS = [{
1457         'note': 'paginated channel',
1458         'url': 'https://www.youtube.com/channel/UCKfVa3S1e4PHvxWcwyMMg8w',
1459         'playlist_mincount': 91,
1460         'info_dict': {
1461             'id': 'UCKfVa3S1e4PHvxWcwyMMg8w',
1462         }
1463     }]
1464
1465     @staticmethod
1466     def extract_videos_from_page(page):
1467         ids_in_page = []
1468         titles_in_page = []
1469         for mobj in re.finditer(r'(?:title="(?P<title>[^"]+)"[^>]+)?href="/watch\?v=(?P<id>[0-9A-Za-z_-]+)&?', page):
1470             video_id = mobj.group('id')
1471             video_title = unescapeHTML(mobj.group('title'))
1472             try:
1473                 idx = ids_in_page.index(video_id)
1474                 if video_title and not titles_in_page[idx]:
1475                     titles_in_page[idx] = video_title
1476             except ValueError:
1477                 ids_in_page.append(video_id)
1478                 titles_in_page.append(video_title)
1479         return zip(ids_in_page, titles_in_page)
1480
1481     def _real_extract(self, url):
1482         channel_id = self._match_id(url)
1483
1484         url = self._TEMPLATE_URL % channel_id
1485
1486         # Channel by page listing is restricted to 35 pages of 30 items, i.e. 1050 videos total (see #5778)
1487         # Workaround by extracting as a playlist if managed to obtain channel playlist URL
1488         # otherwise fallback on channel by page extraction
1489         channel_page = self._download_webpage(
1490             url + '?view=57', channel_id,
1491             'Downloading channel page', fatal=False)
1492         channel_playlist_id = self._html_search_meta(
1493             'channelId', channel_page, 'channel id', default=None)
1494         if not channel_playlist_id:
1495             channel_playlist_id = self._search_regex(
1496                 r'data-channel-external-id="([^"]+)"',
1497                 channel_page, 'channel id', default=None)
1498         if channel_playlist_id and channel_playlist_id.startswith('UC'):
1499             playlist_id = 'UU' + channel_playlist_id[2:]
1500             return self.url_result(
1501                 compat_urlparse.urljoin(url, '/playlist?list=%s' % playlist_id), 'YoutubePlaylist')
1502
1503         channel_page = self._download_webpage(url, channel_id, 'Downloading page #1')
1504         autogenerated = re.search(r'''(?x)
1505                 class="[^"]*?(?:
1506                     channel-header-autogenerated-label|
1507                     yt-channel-title-autogenerated
1508                 )[^"]*"''', channel_page) is not None
1509
1510         if autogenerated:
1511             # The videos are contained in a single page
1512             # the ajax pages can't be used, they are empty
1513             entries = [
1514                 self.url_result(
1515                     video_id, 'Youtube', video_id=video_id,
1516                     video_title=video_title)
1517                 for video_id, video_title in self.extract_videos_from_page(channel_page)]
1518             return self.playlist_result(entries, channel_id)
1519
1520         def _entries():
1521             more_widget_html = content_html = channel_page
1522             for pagenum in itertools.count(1):
1523
1524                 for video_id, video_title in self.extract_videos_from_page(content_html):
1525                     yield self.url_result(
1526                         video_id, 'Youtube', video_id=video_id,
1527                         video_title=video_title)
1528
1529                 mobj = re.search(
1530                     r'data-uix-load-more-href="/?(?P<more>[^"]+)"',
1531                     more_widget_html)
1532                 if not mobj:
1533                     break
1534
1535                 more = self._download_json(
1536                     'https://youtube.com/%s' % mobj.group('more'), channel_id,
1537                     'Downloading page #%s' % (pagenum + 1),
1538                     transform_source=uppercase_escape)
1539                 content_html = more['content_html']
1540                 more_widget_html = more['load_more_widget_html']
1541
1542         return self.playlist_result(_entries(), channel_id)
1543
1544
1545 class YoutubeUserIE(YoutubeChannelIE):
1546     IE_DESC = 'YouTube.com user videos (URL or "ytuser" keyword)'
1547     _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_-]+)'
1548     _TEMPLATE_URL = 'https://www.youtube.com/user/%s/videos'
1549     IE_NAME = 'youtube:user'
1550
1551     _TESTS = [{
1552         'url': 'https://www.youtube.com/user/TheLinuxFoundation',
1553         'playlist_mincount': 320,
1554         'info_dict': {
1555             'title': 'TheLinuxFoundation',
1556         }
1557     }, {
1558         'url': 'ytuser:phihag',
1559         'only_matching': True,
1560     }]
1561
1562     @classmethod
1563     def suitable(cls, url):
1564         # Don't return True if the url can be extracted with other youtube
1565         # extractor, the regex would is too permissive and it would match.
1566         other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1567         if any(ie.suitable(url) for ie in other_ies):
1568             return False
1569         else:
1570             return super(YoutubeUserIE, cls).suitable(url)
1571
1572
1573 class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE):
1574     IE_DESC = 'YouTube.com searches'
1575     # there doesn't appear to be a real limit, for example if you search for
1576     # 'python' you get more than 8.000.000 results
1577     _MAX_RESULTS = float('inf')
1578     IE_NAME = 'youtube:search'
1579     _SEARCH_KEY = 'ytsearch'
1580     _EXTRA_QUERY_ARGS = {}
1581     _TESTS = []
1582
1583     def _get_n_results(self, query, n):
1584         """Get a specified number of results for a query"""
1585
1586         videos = []
1587         limit = n
1588
1589         for pagenum in itertools.count(1):
1590             url_query = {
1591                 'search_query': query.encode('utf-8'),
1592                 'page': pagenum,
1593                 'spf': 'navigate',
1594             }
1595             url_query.update(self._EXTRA_QUERY_ARGS)
1596             result_url = 'https://www.youtube.com/results?' + compat_urllib_parse.urlencode(url_query)
1597             data = self._download_json(
1598                 result_url, video_id='query "%s"' % query,
1599                 note='Downloading page %s' % pagenum,
1600                 errnote='Unable to download API page')
1601             html_content = data[1]['body']['content']
1602
1603             if 'class="search-message' in html_content:
1604                 raise ExtractorError(
1605                     '[youtube] No video results', expected=True)
1606
1607             new_videos = self._ids_to_results(orderedSet(re.findall(
1608                 r'href="/watch\?v=(.{11})', html_content)))
1609             videos += new_videos
1610             if not new_videos or len(videos) > limit:
1611                 break
1612
1613         if len(videos) > n:
1614             videos = videos[:n]
1615         return self.playlist_result(videos, query)
1616
1617
1618 class YoutubeSearchDateIE(YoutubeSearchIE):
1619     IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
1620     _SEARCH_KEY = 'ytsearchdate'
1621     IE_DESC = 'YouTube.com searches, newest videos first'
1622     _EXTRA_QUERY_ARGS = {'search_sort': 'video_date_uploaded'}
1623
1624
1625 class YoutubeSearchURLIE(InfoExtractor):
1626     IE_DESC = 'YouTube.com search URLs'
1627     IE_NAME = 'youtube:search_url'
1628     _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?search_query=(?P<query>[^&]+)(?:[&]|$)'
1629     _TESTS = [{
1630         'url': 'https://www.youtube.com/results?baz=bar&search_query=youtube-dl+test+video&filters=video&lclk=video',
1631         'playlist_mincount': 5,
1632         'info_dict': {
1633             'title': 'youtube-dl test video',
1634         }
1635     }]
1636
1637     def _real_extract(self, url):
1638         mobj = re.match(self._VALID_URL, url)
1639         query = compat_urllib_parse_unquote_plus(mobj.group('query'))
1640
1641         webpage = self._download_webpage(url, query)
1642         result_code = self._search_regex(
1643             r'(?s)<ol[^>]+class="item-section"(.*?)</ol>', webpage, 'result HTML')
1644
1645         part_codes = re.findall(
1646             r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code)
1647         entries = []
1648         for part_code in part_codes:
1649             part_title = self._html_search_regex(
1650                 [r'(?s)title="([^"]+)"', r'>([^<]+)</a>'], part_code, 'item title', fatal=False)
1651             part_url_snippet = self._html_search_regex(
1652                 r'(?s)href="([^"]+)"', part_code, 'item URL')
1653             part_url = compat_urlparse.urljoin(
1654                 'https://www.youtube.com/', part_url_snippet)
1655             entries.append({
1656                 '_type': 'url',
1657                 'url': part_url,
1658                 'title': part_title,
1659             })
1660
1661         return {
1662             '_type': 'playlist',
1663             'entries': entries,
1664             'title': query,
1665         }
1666
1667
1668 class YoutubeShowIE(InfoExtractor):
1669     IE_DESC = 'YouTube.com (multi-season) shows'
1670     _VALID_URL = r'https?://www\.youtube\.com/show/(?P<id>[^?#]*)'
1671     IE_NAME = 'youtube:show'
1672     _TESTS = [{
1673         'url': 'http://www.youtube.com/show/airdisasters',
1674         'playlist_mincount': 3,
1675         'info_dict': {
1676             'id': 'airdisasters',
1677             'title': 'Air Disasters',
1678         }
1679     }]
1680
1681     def _real_extract(self, url):
1682         mobj = re.match(self._VALID_URL, url)
1683         playlist_id = mobj.group('id')
1684         webpage = self._download_webpage(
1685             url, playlist_id, 'Downloading show webpage')
1686         # There's one playlist for each season of the show
1687         m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1688         self.to_screen('%s: Found %s seasons' % (playlist_id, len(m_seasons)))
1689         entries = [
1690             self.url_result(
1691                 'https://www.youtube.com' + season.group(1), 'YoutubePlaylist')
1692             for season in m_seasons
1693         ]
1694         title = self._og_search_title(webpage, fatal=False)
1695
1696         return {
1697             '_type': 'playlist',
1698             'id': playlist_id,
1699             'title': title,
1700             'entries': entries,
1701         }
1702
1703
1704 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1705     """
1706     Base class for feed extractors
1707     Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1708     """
1709     _LOGIN_REQUIRED = True
1710
1711     @property
1712     def IE_NAME(self):
1713         return 'youtube:%s' % self._FEED_NAME
1714
1715     def _real_initialize(self):
1716         self._login()
1717
1718     def _real_extract(self, url):
1719         page = self._download_webpage(
1720             'https://www.youtube.com/feed/%s' % self._FEED_NAME, self._PLAYLIST_TITLE)
1721
1722         # The extraction process is the same as for playlists, but the regex
1723         # for the video ids doesn't contain an index
1724         ids = []
1725         more_widget_html = content_html = page
1726         for page_num in itertools.count(1):
1727             matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html)
1728
1729             # 'recommended' feed has infinite 'load more' and each new portion spins
1730             # the same videos in (sometimes) slightly different order, so we'll check
1731             # for unicity and break when portion has no new videos
1732             new_ids = filter(lambda video_id: video_id not in ids, orderedSet(matches))
1733             if not new_ids:
1734                 break
1735
1736             ids.extend(new_ids)
1737
1738             mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
1739             if not mobj:
1740                 break
1741
1742             more = self._download_json(
1743                 'https://youtube.com/%s' % mobj.group('more'), self._PLAYLIST_TITLE,
1744                 'Downloading page #%s' % page_num,
1745                 transform_source=uppercase_escape)
1746             content_html = more['content_html']
1747             more_widget_html = more['load_more_widget_html']
1748
1749         return self.playlist_result(
1750             self._ids_to_results(ids), playlist_title=self._PLAYLIST_TITLE)
1751
1752
1753 class YoutubeWatchLaterIE(YoutubePlaylistIE):
1754     IE_NAME = 'youtube:watchlater'
1755     IE_DESC = 'Youtube watch later list, ":ytwatchlater" for short (requires authentication)'
1756     _VALID_URL = r'https?://www\.youtube\.com/(?:feed/watch_later|playlist\?list=WL)|:ytwatchlater'
1757
1758     _TESTS = []  # override PlaylistIE tests
1759
1760     def _real_extract(self, url):
1761         return self._extract_playlist('WL')
1762
1763
1764 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1765     IE_NAME = 'youtube:favorites'
1766     IE_DESC = 'YouTube.com favourite videos, ":ytfav" for short (requires authentication)'
1767     _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1768     _LOGIN_REQUIRED = True
1769
1770     def _real_extract(self, url):
1771         webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1772         playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, 'favourites playlist id')
1773         return self.url_result(playlist_id, 'YoutubePlaylist')
1774
1775
1776 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1777     IE_DESC = 'YouTube.com recommended videos, ":ytrec" for short (requires authentication)'
1778     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1779     _FEED_NAME = 'recommended'
1780     _PLAYLIST_TITLE = 'Youtube Recommended videos'
1781
1782
1783 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1784     IE_DESC = 'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
1785     _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1786     _FEED_NAME = 'subscriptions'
1787     _PLAYLIST_TITLE = 'Youtube Subscriptions'
1788
1789
1790 class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
1791     IE_DESC = 'Youtube watch history, ":ythistory" for short (requires authentication)'
1792     _VALID_URL = 'https?://www\.youtube\.com/feed/history|:ythistory'
1793     _FEED_NAME = 'history'
1794     _PLAYLIST_TITLE = 'Youtube History'
1795
1796
1797 class YoutubeTruncatedURLIE(InfoExtractor):
1798     IE_NAME = 'youtube:truncated_url'
1799     IE_DESC = False  # Do not list
1800     _VALID_URL = r'''(?x)
1801         (?:https?://)?
1802         (?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/
1803         (?:watch\?(?:
1804             feature=[a-z_]+|
1805             annotation_id=annotation_[^&]+|
1806             x-yt-cl=[0-9]+|
1807             hl=[^&]*|
1808         )?
1809         |
1810             attribution_link\?a=[^&]+
1811         )
1812         $
1813     '''
1814
1815     _TESTS = [{
1816         'url': 'http://www.youtube.com/watch?annotation_id=annotation_3951667041',
1817         'only_matching': True,
1818     }, {
1819         'url': 'http://www.youtube.com/watch?',
1820         'only_matching': True,
1821     }, {
1822         'url': 'https://www.youtube.com/watch?x-yt-cl=84503534',
1823         'only_matching': True,
1824     }, {
1825         'url': 'https://www.youtube.com/watch?feature=foo',
1826         'only_matching': True,
1827     }, {
1828         'url': 'https://www.youtube.com/watch?hl=en-GB',
1829         'only_matching': True,
1830     }]
1831
1832     def _real_extract(self, url):
1833         raise ExtractorError(
1834             'Did you forget to quote the URL? Remember that & is a meta '
1835             'character in most shells, so you want to put the URL in quotes, '
1836             'like  youtube-dl '
1837             '"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
1838             ' or simply  youtube-dl BaW_jenozKc  .',
1839             expected=True)
1840
1841
1842 class YoutubeTruncatedIDIE(InfoExtractor):
1843     IE_NAME = 'youtube:truncated_id'
1844     IE_DESC = False  # Do not list
1845     _VALID_URL = r'https?://(?:www\.)?youtube\.com/watch\?v=(?P<id>[0-9A-Za-z_-]{1,10})$'
1846
1847     _TESTS = [{
1848         'url': 'https://www.youtube.com/watch?v=N_708QY7Ob',
1849         'only_matching': True,
1850     }]
1851
1852     def _real_extract(self, url):
1853         video_id = self._match_id(url)
1854         raise ExtractorError(
1855             'Incomplete YouTube ID %s. URL %s looks truncated.' % (video_id, url),
1856             expected=True)