youtube_dl/extractor/youtube.py

   1 # coding: utf-8
   2
   3 import errno
   4 import io
   5 import itertools
   6 import json
   7 import os.path
   8 import re
   9 import traceback
  10
  11 from .common import InfoExtractor, SearchInfoExtractor
  12 from .subtitles import SubtitlesInfoExtractor
  13 from ..jsinterp import JSInterpreter
  14 from ..swfinterp import SWFInterpreter
  15 from ..utils import (
  16     compat_chr,
  17     compat_parse_qs,
  18     compat_urllib_parse,
  19     compat_urllib_request,
  20     compat_urlparse,
  21     compat_str,
  22
  23     clean_html,
  24     get_cachedir,
  25     get_element_by_id,
  26     get_element_by_attribute,
  27     ExtractorError,
  28     int_or_none,
  29     PagedList,
  30     unescapeHTML,
  31     unified_strdate,
  32     orderedSet,
  33     write_json_file,
  34     uppercase_escape,
  35 )
  36
  37 class YoutubeBaseInfoExtractor(InfoExtractor):
  38     """Provide base functions for Youtube extractors"""
  39     _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
  40     _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1'
  41     _AGE_URL = 'https://www.youtube.com/verify_age?next_url=/&gl=US&hl=en'
  42     _NETRC_MACHINE = 'youtube'
  43     # If True it will raise an error if no login info is provided
  44     _LOGIN_REQUIRED = False
  45
  46     def _set_language(self):
  47         return bool(self._download_webpage(
  48             self._LANG_URL, None,
  49             note=u'Setting language', errnote='unable to set language',
  50             fatal=False))
  51
  52     def _login(self):
  53         (username, password) = self._get_login_info()
  54         # No authentication to be performed
  55         if username is None:
  56             if self._LOGIN_REQUIRED:
  57                 raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True)
  58             return False
  59
  60         login_page = self._download_webpage(
  61             self._LOGIN_URL, None,
  62             note=u'Downloading login page',
  63             errnote=u'unable to fetch login page', fatal=False)
  64         if login_page is False:
  65             return
  66
  67         galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"',
  68                                   login_page, u'Login GALX parameter')
  69
  70         # Log in
  71         login_form_strs = {
  72                 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
  73                 u'Email': username,
  74                 u'GALX': galx,
  75                 u'Passwd': password,
  76                 u'PersistentCookie': u'yes',
  77                 u'_utf8': u'霱',
  78                 u'bgresponse': u'js_disabled',
  79                 u'checkConnection': u'',
  80                 u'checkedDomains': u'youtube',
  81                 u'dnConn': u'',
  82                 u'pstMsg': u'0',
  83                 u'rmShown': u'1',
  84                 u'secTok': u'',
  85                 u'signIn': u'Sign in',
  86                 u'timeStmp': u'',
  87                 u'service': u'youtube',
  88                 u'uilel': u'3',
  89                 u'hl': u'en_US',
  90         }
  91         # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
  92         # chokes on unicode
  93         login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items())
  94         login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
  95
  96         req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
  97         login_results = self._download_webpage(
  98             req, None,
  99             note=u'Logging in', errnote=u'unable to log in', fatal=False)
 100         if login_results is False:
 101             return False
 102         if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
 103             self._downloader.report_warning(u'unable to log in: bad username or password')
 104             return False
 105         return True
 106
 107     def _confirm_age(self):
 108         age_form = {
 109             'next_url': '/',
 110             'action_confirm': 'Confirm',
 111         }
 112         req = compat_urllib_request.Request(self._AGE_URL,
 113             compat_urllib_parse.urlencode(age_form).encode('ascii'))
 114
 115         self._download_webpage(
 116             req, None,
 117             note=u'Confirming age', errnote=u'Unable to confirm age')
 118         return True
 119
 120     def _real_initialize(self):
 121         if self._downloader is None:
 122             return
 123         if not self._set_language():
 124             return
 125         if not self._login():
 126             return
 127         self._confirm_age()
 128
 129
 130 class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
 131     IE_DESC = u'YouTube.com'
 132     _VALID_URL = r"""(?x)^
 133                      (
 134                          (?:https?://|//)?                                    # http(s):// or protocol-independent URL (optional)
 135                          (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/|
 136                             (?:www\.)?deturl\.com/www\.youtube\.com/|
 137                             (?:www\.)?pwnyoutube\.com/|
 138                             (?:www\.)?yourepeat\.com/|
 139                             tube\.majestyc\.net/|
 140                             youtube\.googleapis\.com/)                        # the various hostnames, with wildcard subdomains
 141                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
 142                          (?:                                                  # the various things that can precede the ID:
 143                              (?:(?:v|embed|e)/)                               # v/ or embed/ or e/
 144                              |(?:                                             # or the v= param in all its forms
 145                                  (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)?  # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
 146                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
 147                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
 148                                  v=
 149                              )
 150                          ))
 151                          |youtu\.be/                                          # just youtu.be/xxxx
 152                          |https?://(?:www\.)?cleanvideosearch\.com/media/action/yt/watch\?videoId=
 153                          )
 154                      )?                                                       # all until now is optional -> you can pass the naked ID
 155                      ([0-9A-Za-z_-]{11})                                      # here is it! the YouTube video ID
 156                      (?(1).+)?                                                # if we found the ID, everything can follow
 157                      $"""
 158     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
 159     _formats = {
 160         '5': {'ext': 'flv', 'width': 400, 'height': 240},
 161         '6': {'ext': 'flv', 'width': 450, 'height': 270},
 162         '13': {'ext': '3gp'},
 163         '17': {'ext': '3gp', 'width': 176, 'height': 144},
 164         '18': {'ext': 'mp4', 'width': 640, 'height': 360},
 165         '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
 166         '34': {'ext': 'flv', 'width': 640, 'height': 360},
 167         '35': {'ext': 'flv', 'width': 854, 'height': 480},
 168         '36': {'ext': '3gp', 'width': 320, 'height': 240},
 169         '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
 170         '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
 171         '43': {'ext': 'webm', 'width': 640, 'height': 360},
 172         '44': {'ext': 'webm', 'width': 854, 'height': 480},
 173         '45': {'ext': 'webm', 'width': 1280, 'height': 720},
 174         '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
 175
 176
 177         # 3d videos
 178         '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20},
 179         '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20},
 180         '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20},
 181         '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20},
 182         '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20},
 183         '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20},
 184         '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20},
 185
 186         # Apple HTTP Live Streaming
 187         '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 188         '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10},
 189         '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10},
 190         '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10},
 191         '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10},
 192         '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 193         '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10},
 194
 195         # DASH mp4 video
 196         '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 197         '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 198         '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 199         '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 200         '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 201         '138': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 202         '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 203         '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 204
 205         # Dash mp4 audio
 206         '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50},
 207         '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50},
 208         '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50},
 209
 210         # Dash webm
 211         '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 212         '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 213         '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 214         '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 215         '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 216         '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 217         '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 218         '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 219         '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 220         '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 221         '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 222         '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 223         '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 224         '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 225         '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 226
 227         # Dash webm audio
 228         '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 48, 'preference': -50},
 229         '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50},
 230
 231         # RTMP (unnamed)
 232         '_rtmp': {'protocol': 'rtmp'},
 233     }
 234
 235     IE_NAME = u'youtube'
 236     _TESTS = [
 237         {
 238             u"url":  u"http://www.youtube.com/watch?v=BaW_jenozKc",
 239             u"file":  u"BaW_jenozKc.mp4",
 240             u"info_dict": {
 241                 u"title": u"youtube-dl test video \"'/\\ä↭𝕐",
 242                 u"uploader": u"Philipp Hagemeister",
 243                 u"uploader_id": u"phihag",
 244                 u"upload_date": u"20121002",
 245                 u"description": u"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 .",
 246                 u"categories": [u'Science & Technology'],
 247             }
 248         },
 249         {
 250             u"url":  u"http://www.youtube.com/watch?v=UxxajLWwzqY",
 251             u"file":  u"UxxajLWwzqY.mp4",
 252             u"note": u"Test generic use_cipher_signature video (#897)",
 253             u"info_dict": {
 254                 u"upload_date": u"20120506",
 255                 u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
 256                 u"description": u"md5:fea86fda2d5a5784273df5c7cc994d9f",
 257                 u"uploader": u"Icona Pop",
 258                 u"uploader_id": u"IconaPop"
 259             }
 260         },
 261         {
 262             u"url":  u"https://www.youtube.com/watch?v=07FYdnEawAQ",
 263             u"file":  u"07FYdnEawAQ.mp4",
 264             u"note": u"Test VEVO video with age protection (#956)",
 265             u"info_dict": {
 266                 u"upload_date": u"20130703",
 267                 u"title": u"Justin Timberlake - Tunnel Vision (Explicit)",
 268                 u"description": u"md5:64249768eec3bc4276236606ea996373",
 269                 u"uploader": u"justintimberlakeVEVO",
 270                 u"uploader_id": u"justintimberlakeVEVO"
 271             }
 272         },
 273         {
 274             u"url":  u"//www.YouTube.com/watch?v=yZIXLfi8CZQ",
 275             u"file":  u"yZIXLfi8CZQ.mp4",
 276             u"note": u"Embed-only video (#1746)",
 277             u"info_dict": {
 278                 u"upload_date": u"20120608",
 279                 u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012",
 280                 u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7",
 281                 u"uploader": u"SET India",
 282                 u"uploader_id": u"setindia"
 283             }
 284         },
 285         {
 286             u"url": u"http://www.youtube.com/watch?v=a9LDPn-MO4I",
 287             u"file": u"a9LDPn-MO4I.m4a",
 288             u"note": u"256k DASH audio (format 141) via DASH manifest",
 289             u"info_dict": {
 290                 u"upload_date": "20121002",
 291                 u"uploader_id": "8KVIDEO",
 292                 u"description": "No description available.",
 293                 u"uploader": "8KVIDEO",
 294                 u"title": "UHDTV TEST 8K VIDEO.mp4"
 295             },
 296             u"params": {
 297                 u"youtube_include_dash_manifest": True,
 298                 u"format": "141",
 299             },
 300         },
 301         # DASH manifest with encrypted signature
 302         {
 303             u'url': u'https://www.youtube.com/watch?v=IB3lcPjvWLA',
 304             u'info_dict': {
 305                 u'id': u'IB3lcPjvWLA',
 306                 u'ext': u'm4a',
 307                 u'title': u'Afrojack - The Spark ft. Spree Wilson',
 308                 u'description': u'md5:9717375db5a9a3992be4668bbf3bc0a8',
 309                 u'uploader': u'AfrojackVEVO',
 310                 u'uploader_id': u'AfrojackVEVO',
 311                 u'upload_date': u'20131011',
 312             },
 313             u"params": {
 314                 u'youtube_include_dash_manifest': True,
 315                 u'format': '141',
 316             },
 317         },
 318     ]
 319
 320
 321     @classmethod
 322     def suitable(cls, url):
 323         """Receives a URL and returns True if suitable for this IE."""
 324         if YoutubePlaylistIE.suitable(url): return False
 325         return re.match(cls._VALID_URL, url) is not None
 326
 327     def __init__(self, *args, **kwargs):
 328         super(YoutubeIE, self).__init__(*args, **kwargs)
 329         self._player_cache = {}
 330
 331     def report_video_info_webpage_download(self, video_id):
 332         """Report attempt to download video info webpage."""
 333         self.to_screen(u'%s: Downloading video info webpage' % video_id)
 334
 335     def report_information_extraction(self, video_id):
 336         """Report attempt to extract video information."""
 337         self.to_screen(u'%s: Extracting video information' % video_id)
 338
 339     def report_unavailable_format(self, video_id, format):
 340         """Report extracted video URL."""
 341         self.to_screen(u'%s: Format %s not available' % (video_id, format))
 342
 343     def report_rtmp_download(self):
 344         """Indicate the download will use the RTMP protocol."""
 345         self.to_screen(u'RTMP download detected')
 346
 347     def _extract_signature_function(self, video_id, player_url, slen):
 348         id_m = re.match(
 349             r'.*-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3)?\.(?P<ext>[a-z]+)$',
 350             player_url)
 351         player_type = id_m.group('ext')
 352         player_id = id_m.group('id')
 353
 354         # Read from filesystem cache
 355         func_id = '%s_%s_%d' % (player_type, player_id, slen)
 356         assert os.path.basename(func_id) == func_id
 357         cache_dir = get_cachedir(self._downloader.params)
 358
 359         cache_enabled = cache_dir is not None
 360         if cache_enabled:
 361             cache_fn = os.path.join(os.path.expanduser(cache_dir),
 362                                     u'youtube-sigfuncs',
 363                                     func_id + '.json')
 364             try:
 365                 with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
 366                     cache_spec = json.load(cachef)
 367                 return lambda s: u''.join(s[i] for i in cache_spec)
 368             except IOError:
 369                 pass  # No cache available
 370
 371         if player_type == 'js':
 372             code = self._download_webpage(
 373                 player_url, video_id,
 374                 note=u'Downloading %s player %s' % (player_type, player_id),
 375                 errnote=u'Download of %s failed' % player_url)
 376             res = self._parse_sig_js(code)
 377         elif player_type == 'swf':
 378             urlh = self._request_webpage(
 379                 player_url, video_id,
 380                 note=u'Downloading %s player %s' % (player_type, player_id),
 381                 errnote=u'Download of %s failed' % player_url)
 382             code = urlh.read()
 383             res = self._parse_sig_swf(code)
 384         else:
 385             assert False, 'Invalid player type %r' % player_type
 386
 387         if cache_enabled:
 388             try:
 389                 test_string = u''.join(map(compat_chr, range(slen)))
 390                 cache_res = res(test_string)
 391                 cache_spec = [ord(c) for c in cache_res]
 392                 try:
 393                     os.makedirs(os.path.dirname(cache_fn))
 394                 except OSError as ose:
 395                     if ose.errno != errno.EEXIST:
 396                         raise
 397                 write_json_file(cache_spec, cache_fn)
 398             except Exception:
 399                 tb = traceback.format_exc()
 400                 self._downloader.report_warning(
 401                     u'Writing cache to %r failed: %s' % (cache_fn, tb))
 402
 403         return res
 404
 405     def _print_sig_code(self, func, slen):
 406         def gen_sig_code(idxs):
 407             def _genslice(start, end, step):
 408                 starts = u'' if start == 0 else str(start)
 409                 ends = (u':%d' % (end+step)) if end + step >= 0 else u':'
 410                 steps = u'' if step == 1 else (u':%d' % step)
 411                 return u's[%s%s%s]' % (starts, ends, steps)
 412
 413             step = None
 414             start = '(Never used)'  # Quelch pyflakes warnings - start will be
 415                                     # set as soon as step is set
 416             for i, prev in zip(idxs[1:], idxs[:-1]):
 417                 if step is not None:
 418                     if i - prev == step:
 419                         continue
 420                     yield _genslice(start, prev, step)
 421                     step = None
 422                     continue
 423                 if i - prev in [-1, 1]:
 424                     step = i - prev
 425                     start = prev
 426                     continue
 427                 else:
 428                     yield u's[%d]' % prev
 429             if step is None:
 430                 yield u's[%d]' % i
 431             else:
 432                 yield _genslice(start, i, step)
 433
 434         test_string = u''.join(map(compat_chr, range(slen)))
 435         cache_res = func(test_string)
 436         cache_spec = [ord(c) for c in cache_res]
 437         expr_code = u' + '.join(gen_sig_code(cache_spec))
 438         code = u'if len(s) == %d:\n    return %s\n' % (slen, expr_code)
 439         self.to_screen(u'Extracted signature function:\n' + code)
 440
 441     def _parse_sig_js(self, jscode):
 442         funcname = self._search_regex(
 443             r'signature=([$a-zA-Z]+)', jscode,
 444              u'Initial JS player signature function name')
 445
 446         jsi = JSInterpreter(jscode)
 447         initial_function = jsi.extract_function(funcname)
 448         return lambda s: initial_function([s])
 449
 450     def _parse_sig_swf(self, file_contents):
 451         swfi = SWFInterpreter(file_contents)
 452         TARGET_CLASSNAME = u'SignatureDecipher'
 453         searched_class = swfi.extract_class(TARGET_CLASSNAME)
 454         initial_function = swfi.extract_function(searched_class, u'decipher')
 455         return lambda s: initial_function([s])
 456
 457     def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
 458         """Turn the encrypted s field into a working signature"""
 459
 460         if player_url is None:
 461             raise ExtractorError(u'Cannot decrypt signature without player_url')
 462
 463         if player_url.startswith(u'//'):
 464             player_url = u'https:' + player_url
 465         try:
 466             player_id = (player_url, len(s))
 467             if player_id not in self._player_cache:
 468                 func = self._extract_signature_function(
 469                     video_id, player_url, len(s)
 470                 )
 471                 self._player_cache[player_id] = func
 472             func = self._player_cache[player_id]
 473             if self._downloader.params.get('youtube_print_sig_code'):
 474                 self._print_sig_code(func, len(s))
 475             return func(s)
 476         except Exception as e:
 477             tb = traceback.format_exc()
 478             raise ExtractorError(
 479                 u'Automatic signature extraction failed: ' + tb, cause=e)
 480
 481     def _get_available_subtitles(self, video_id, webpage):
 482         try:
 483             sub_list = self._download_webpage(
 484                 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
 485                 video_id, note=False)
 486         except ExtractorError as err:
 487             self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
 488             return {}
 489         lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
 490
 491         sub_lang_list = {}
 492         for l in lang_list:
 493             lang = l[1]
 494             params = compat_urllib_parse.urlencode({
 495                 'lang': lang,
 496                 'v': video_id,
 497                 'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
 498                 'name': unescapeHTML(l[0]).encode('utf-8'),
 499             })
 500             url = u'https://www.youtube.com/api/timedtext?' + params
 501             sub_lang_list[lang] = url
 502         if not sub_lang_list:
 503             self._downloader.report_warning(u'video doesn\'t have subtitles')
 504             return {}
 505         return sub_lang_list
 506
 507     def _get_available_automatic_caption(self, video_id, webpage):
 508         """We need the webpage for getting the captions url, pass it as an
 509            argument to speed up the process."""
 510         sub_format = self._downloader.params.get('subtitlesformat', 'srt')
 511         self.to_screen(u'%s: Looking for automatic captions' % video_id)
 512         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
 513         err_msg = u'Couldn\'t find automatic captions for %s' % video_id
 514         if mobj is None:
 515             self._downloader.report_warning(err_msg)
 516             return {}
 517         player_config = json.loads(mobj.group(1))
 518         try:
 519             args = player_config[u'args']
 520             caption_url = args[u'ttsurl']
 521             timestamp = args[u'timestamp']
 522             # We get the available subtitles
 523             list_params = compat_urllib_parse.urlencode({
 524                 'type': 'list',
 525                 'tlangs': 1,
 526                 'asrs': 1,
 527             })
 528             list_url = caption_url + '&' + list_params
 529             caption_list = self._download_xml(list_url, video_id)
 530             original_lang_node = caption_list.find('track')
 531             if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' :
 532                 self._downloader.report_warning(u'Video doesn\'t have automatic captions')
 533                 return {}
 534             original_lang = original_lang_node.attrib['lang_code']
 535
 536             sub_lang_list = {}
 537             for lang_node in caption_list.findall('target'):
 538                 sub_lang = lang_node.attrib['lang_code']
 539                 params = compat_urllib_parse.urlencode({
 540                     'lang': original_lang,
 541                     'tlang': sub_lang,
 542                     'fmt': sub_format,
 543                     'ts': timestamp,
 544                     'kind': 'asr',
 545                 })
 546                 sub_lang_list[sub_lang] = caption_url + '&' + params
 547             return sub_lang_list
 548         # An extractor error can be raise by the download process if there are
 549         # no automatic captions but there are subtitles
 550         except (KeyError, ExtractorError):
 551             self._downloader.report_warning(err_msg)
 552             return {}
 553
 554     @classmethod
 555     def extract_id(cls, url):
 556         mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
 557         if mobj is None:
 558             raise ExtractorError(u'Invalid URL: %s' % url)
 559         video_id = mobj.group(2)
 560         return video_id
 561
 562     def _extract_from_m3u8(self, manifest_url, video_id):
 563         url_map = {}
 564         def _get_urls(_manifest):
 565             lines = _manifest.split('\n')
 566             urls = filter(lambda l: l and not l.startswith('#'),
 567                             lines)
 568             return urls
 569         manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest')
 570         formats_urls = _get_urls(manifest)
 571         for format_url in formats_urls:
 572             itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
 573             url_map[itag] = format_url
 574         return url_map
 575
 576     def _extract_annotations(self, video_id):
 577         url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
 578         return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.')
 579
 580     def _real_extract(self, url):
 581         proto = (
 582             u'http' if self._downloader.params.get('prefer_insecure', False)
 583             else u'https')
 584
 585         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
 586         mobj = re.search(self._NEXT_URL_RE, url)
 587         if mobj:
 588             url = proto + '://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
 589         video_id = self.extract_id(url)
 590
 591         # Get video webpage
 592         url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
 593         video_webpage = self._download_webpage(url, video_id)
 594
 595         # Attempt to extract SWF player URL
 596         mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
 597         if mobj is not None:
 598             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
 599         else:
 600             player_url = None
 601
 602         # Get video info
 603         self.report_video_info_webpage_download(video_id)
 604         if re.search(r'player-age-gate-content">', video_webpage) is not None:
 605             self.report_age_confirmation()
 606             age_gate = True
 607             # We simulate the access to the video from www.youtube.com/v/{video_id}
 608             # this can be viewed without login into Youtube
 609             data = compat_urllib_parse.urlencode({
 610                 'video_id': video_id,
 611                 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
 612                 'sts':'16268',
 613             })
 614             video_info_url = proto + '://www.youtube.com/get_video_info?' + data
 615             video_info_webpage = self._download_webpage(video_info_url, video_id,
 616                                     note=False,
 617                                     errnote='unable to download video info webpage')
 618             video_info = compat_parse_qs(video_info_webpage)
 619         else:
 620             age_gate = False
 621             for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
 622                 video_info_url = (proto + '://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
 623                         % (video_id, el_type))
 624                 video_info_webpage = self._download_webpage(video_info_url, video_id,
 625                                         note=False,
 626                                         errnote='unable to download video info webpage')
 627                 video_info = compat_parse_qs(video_info_webpage)
 628                 if 'token' in video_info:
 629                     break
 630         if 'token' not in video_info:
 631             if 'reason' in video_info:
 632                 raise ExtractorError(
 633                     u'YouTube said: %s' % video_info['reason'][0],
 634                     expected=True, video_id=video_id)
 635             else:
 636                 raise ExtractorError(
 637                     u'"token" parameter not in video info for unknown reason',
 638                     video_id=video_id)
 639
 640         if 'view_count' in video_info:
 641             view_count = int(video_info['view_count'][0])
 642         else:
 643             view_count = None
 644
 645         # Check for "rental" videos
 646         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
 647             raise ExtractorError(u'"rental" videos not supported')
 648
 649         # Start extracting information
 650         self.report_information_extraction(video_id)
 651
 652         # uploader
 653         if 'author' not in video_info:
 654             raise ExtractorError(u'Unable to extract uploader name')
 655         video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
 656
 657         # uploader_id
 658         video_uploader_id = None
 659         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
 660         if mobj is not None:
 661             video_uploader_id = mobj.group(1)
 662         else:
 663             self._downloader.report_warning(u'unable to extract uploader nickname')
 664
 665         # title
 666         if 'title' in video_info:
 667             video_title = video_info['title'][0]
 668         else:
 669             self._downloader.report_warning(u'Unable to extract video title')
 670             video_title = u'_'
 671
 672         # thumbnail image
 673         # We try first to get a high quality image:
 674         m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
 675                             video_webpage, re.DOTALL)
 676         if m_thumb is not None:
 677             video_thumbnail = m_thumb.group(1)
 678         elif 'thumbnail_url' not in video_info:
 679             self._downloader.report_warning(u'unable to extract video thumbnail')
 680             video_thumbnail = None
 681         else:   # don't panic if we can't find it
 682             video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
 683
 684         # upload date
 685         upload_date = None
 686         mobj = re.search(r'(?s)id="eow-date.*?>(.*?)</span>', video_webpage)
 687         if mobj is None:
 688             mobj = re.search(
 689                 r'(?s)id="watch-uploader-info".*?>.*?(?:Published|Uploaded|Streamed live) on (.*?)</strong>',
 690                 video_webpage)
 691         if mobj is not None:
 692             upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
 693             upload_date = unified_strdate(upload_date)
 694
 695         m_cat_container = get_element_by_id("eow-category", video_webpage)
 696         if m_cat_container:
 697             category = self._html_search_regex(
 698                 r'(?s)<a[^<]+>(.*?)</a>', m_cat_container, 'category',
 699                 default=None)
 700             video_categories = None if category is None else [category]
 701         else:
 702             video_categories = None
 703
 704         # description
 705         video_description = get_element_by_id("eow-description", video_webpage)
 706         if video_description:
 707             video_description = re.sub(r'''(?x)
 708                 <a\s+
 709                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
 710                     title="([^"]+)"\s+
 711                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
 712                     class="yt-uix-redirect-link"\s*>
 713                 [^<]+
 714                 </a>
 715             ''', r'\1', video_description)
 716             video_description = clean_html(video_description)
 717         else:
 718             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
 719             if fd_mobj:
 720                 video_description = unescapeHTML(fd_mobj.group(1))
 721             else:
 722                 video_description = u''
 723
 724         def _extract_count(klass):
 725             count = self._search_regex(
 726                 r'class="%s">([\d,]+)</span>' % re.escape(klass),
 727                 video_webpage, klass, default=None)
 728             if count is not None:
 729                 return int(count.replace(',', ''))
 730             return None
 731         like_count = _extract_count(u'likes-count')
 732         dislike_count = _extract_count(u'dislikes-count')
 733
 734         # subtitles
 735         video_subtitles = self.extract_subtitles(video_id, video_webpage)
 736
 737         if self._downloader.params.get('listsubtitles', False):
 738             self._list_available_subtitles(video_id, video_webpage)
 739             return
 740
 741         if 'length_seconds' not in video_info:
 742             self._downloader.report_warning(u'unable to extract video duration')
 743             video_duration = None
 744         else:
 745             video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]))
 746
 747         # annotations
 748         video_annotations = None
 749         if self._downloader.params.get('writeannotations', False):
 750                 video_annotations = self._extract_annotations(video_id)
 751
 752         # Decide which formats to download
 753         try:
 754             mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage)
 755             if not mobj:
 756                 raise ValueError('Could not find vevo ID')
 757             json_code = uppercase_escape(mobj.group(1))
 758             ytplayer_config = json.loads(json_code)
 759             args = ytplayer_config['args']
 760             # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
 761             # this signatures are encrypted
 762             if 'url_encoded_fmt_stream_map' not in args:
 763                 raise ValueError(u'No stream_map present')  # caught below
 764             re_signature = re.compile(r'[&,]s=')
 765             m_s = re_signature.search(args['url_encoded_fmt_stream_map'])
 766             if m_s is not None:
 767                 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
 768                 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
 769             m_s = re_signature.search(args.get('adaptive_fmts', u''))
 770             if m_s is not None:
 771                 if 'adaptive_fmts' in video_info:
 772                     video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts']
 773                 else:
 774                     video_info['adaptive_fmts'] = [args['adaptive_fmts']]
 775         except ValueError:
 776             pass
 777
 778         def _map_to_format_list(urlmap):
 779             formats = []
 780             for itag, video_real_url in urlmap.items():
 781                 dct = {
 782                     'format_id': itag,
 783                     'url': video_real_url,
 784                     'player_url': player_url,
 785                 }
 786                 if itag in self._formats:
 787                     dct.update(self._formats[itag])
 788                 formats.append(dct)
 789             return formats
 790
 791         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
 792             self.report_rtmp_download()
 793             formats = [{
 794                 'format_id': '_rtmp',
 795                 'protocol': 'rtmp',
 796                 'url': video_info['conn'][0],
 797                 'player_url': player_url,
 798             }]
 799         elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1:
 800             encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0]
 801             if 'rtmpe%3Dyes' in encoded_url_map:
 802                 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
 803             url_map = {}
 804             for url_data_str in encoded_url_map.split(','):
 805                 url_data = compat_parse_qs(url_data_str)
 806                 if 'itag' in url_data and 'url' in url_data:
 807                     url = url_data['url'][0]
 808                     if 'sig' in url_data:
 809                         url += '&signature=' + url_data['sig'][0]
 810                     elif 's' in url_data:
 811                         encrypted_sig = url_data['s'][0]
 812
 813                         if not age_gate:
 814                             jsplayer_url_json = self._search_regex(
 815                                 r'"assets":.+?"js":\s*("[^"]+")',
 816                                 video_webpage, u'JS player URL')
 817                             player_url = json.loads(jsplayer_url_json)
 818                         if player_url is None:
 819                             player_url_json = self._search_regex(
 820                                 r'ytplayer\.config.*?"url"\s*:\s*("[^"]+")',
 821                                 video_webpage, u'age gate player URL')
 822                             player_url = json.loads(player_url_json)
 823
 824                         if self._downloader.params.get('verbose'):
 825                             if player_url is None:
 826                                 player_version = 'unknown'
 827                                 player_desc = 'unknown'
 828                             else:
 829                                 if player_url.endswith('swf'):
 830                                     player_version = self._search_regex(
 831                                         r'-(.+?)(?:/watch_as3)?\.swf$', player_url,
 832                                         u'flash player', fatal=False)
 833                                     player_desc = 'flash player %s' % player_version
 834                                 else:
 835                                     player_version = self._search_regex(
 836                                         r'html5player-(.+?)\.js', video_webpage,
 837                                         'html5 player', fatal=False)
 838                                     player_desc = u'html5 player %s' % player_version
 839
 840                             parts_sizes = u'.'.join(compat_str(len(part)) for part in encrypted_sig.split('.'))
 841                             self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' %
 842                                 (len(encrypted_sig), parts_sizes, url_data['itag'][0], player_desc))
 843
 844                         signature = self._decrypt_signature(
 845                             encrypted_sig, video_id, player_url, age_gate)
 846                         url += '&signature=' + signature
 847                     if 'ratebypass' not in url:
 848                         url += '&ratebypass=yes'
 849                     url_map[url_data['itag'][0]] = url
 850             formats = _map_to_format_list(url_map)
 851         elif video_info.get('hlsvp'):
 852             manifest_url = video_info['hlsvp'][0]
 853             url_map = self._extract_from_m3u8(manifest_url, video_id)
 854             formats = _map_to_format_list(url_map)
 855         else:
 856             raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
 857
 858         # Look for the DASH manifest
 859         if (self._downloader.params.get('youtube_include_dash_manifest', False)):
 860             try:
 861                 # The DASH manifest used needs to be the one from the original video_webpage.
 862                 # The one found in get_video_info seems to be using different signatures.
 863                 # However, in the case of an age restriction there won't be any embedded dashmpd in the video_webpage.
 864                 # Luckily, it seems, this case uses some kind of default signature (len == 86), so the
 865                 # combination of get_video_info and the _static_decrypt_signature() decryption fallback will work here.
 866                 if age_gate:
 867                     dash_manifest_url = video_info.get('dashmpd')[0]
 868                 else:
 869                     dash_manifest_url = ytplayer_config['args']['dashmpd']
 870                 def decrypt_sig(mobj):
 871                     s = mobj.group(1)
 872                     dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
 873                     return '/signature/%s' % dec_s
 874                 dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url)
 875                 dash_doc = self._download_xml(
 876                     dash_manifest_url, video_id,
 877                     note=u'Downloading DASH manifest',
 878                     errnote=u'Could not download DASH manifest')
 879                 for r in dash_doc.findall(u'.//{urn:mpeg:DASH:schema:MPD:2011}Representation'):
 880                     url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL')
 881                     if url_el is None:
 882                         continue
 883                     format_id = r.attrib['id']
 884                     video_url = url_el.text
 885                     filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
 886                     f = {
 887                         'format_id': format_id,
 888                         'url': video_url,
 889                         'width': int_or_none(r.attrib.get('width')),
 890                         'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
 891                         'asr': int_or_none(r.attrib.get('audioSamplingRate')),
 892                         'filesize': filesize,
 893                     }
 894                     try:
 895                         existing_format = next(
 896                             fo for fo in formats
 897                             if fo['format_id'] == format_id)
 898                     except StopIteration:
 899                         f.update(self._formats.get(format_id, {}))
 900                         formats.append(f)
 901                     else:
 902                         existing_format.update(f)
 903
 904             except (ExtractorError, KeyError) as e:
 905                 self.report_warning(u'Skipping DASH manifest: %s' % e, video_id)
 906
 907         self._sort_formats(formats)
 908
 909         return {
 910             'id':           video_id,
 911             'uploader':     video_uploader,
 912             'uploader_id':  video_uploader_id,
 913             'upload_date':  upload_date,
 914             'title':        video_title,
 915             'thumbnail':    video_thumbnail,
 916             'description':  video_description,
 917             'categories':   video_categories,
 918             'subtitles':    video_subtitles,
 919             'duration':     video_duration,
 920             'age_limit':    18 if age_gate else 0,
 921             'annotations':  video_annotations,
 922             'webpage_url': proto + '://www.youtube.com/watch?v=%s' % video_id,
 923             'view_count':   view_count,
 924             'like_count': like_count,
 925             'dislike_count': dislike_count,
 926             'formats':      formats,
 927         }
 928
 929 class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
 930     IE_DESC = u'YouTube.com playlists'
 931     _VALID_URL = r"""(?x)(?:
 932                         (?:https?://)?
 933                         (?:\w+\.)?
 934                         youtube\.com/
 935                         (?:
 936                            (?:course|view_play_list|my_playlists|artist|playlist|watch)
 937                            \? (?:.*?&)*? (?:p|a|list)=
 938                         |  p/
 939                         )
 940                         (
 941                             (?:PL|LL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,}
 942                             # Top tracks, they can also include dots
 943                             |(?:MC)[\w\.]*
 944                         )
 945                         .*
 946                      |
 947                         ((?:PL|LL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,})
 948                      )"""
 949     _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s'
 950     _MORE_PAGES_INDICATOR = r'data-link-type="next"'
 951     _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&amp;[^"]*?index=(?P<index>\d+)'
 952     IE_NAME = u'youtube:playlist'
 953
 954     def _real_initialize(self):
 955         self._login()
 956
 957     def _ids_to_results(self, ids):
 958         return [self.url_result(vid_id, 'Youtube', video_id=vid_id)
 959                        for vid_id in ids]
 960
 961     def _extract_mix(self, playlist_id):
 962         # The mixes are generated from a a single video
 963         # the id of the playlist is just 'RD' + video_id
 964         url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id)
 965         webpage = self._download_webpage(url, playlist_id, u'Downloading Youtube mix')
 966         search_title = lambda class_name: get_element_by_attribute('class', class_name, webpage)
 967         title_span = (search_title('playlist-title') or
 968             search_title('title long-title') or search_title('title'))
 969         title = clean_html(title_span)
 970         video_re = r'''(?x)data-video-username=".*?".*?
 971                        href="/watch\?v=([0-9A-Za-z_-]{11})&amp;[^"]*?list=%s''' % re.escape(playlist_id)
 972         ids = orderedSet(re.findall(video_re, webpage, flags=re.DOTALL))
 973         url_results = self._ids_to_results(ids)
 974
 975         return self.playlist_result(url_results, playlist_id, title)
 976
 977     def _real_extract(self, url):
 978         # Extract playlist id
 979         mobj = re.match(self._VALID_URL, url)
 980         if mobj is None:
 981             raise ExtractorError(u'Invalid URL: %s' % url)
 982         playlist_id = mobj.group(1) or mobj.group(2)
 983
 984         # Check if it's a video-specific URL
 985         query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
 986         if 'v' in query_dict:
 987             video_id = query_dict['v'][0]
 988             if self._downloader.params.get('noplaylist'):
 989                 self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
 990                 return self.url_result(video_id, 'Youtube', video_id=video_id)
 991             else:
 992                 self.to_screen(u'Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id))
 993
 994         if playlist_id.startswith('RD'):
 995             # Mixes require a custom extraction process
 996             return self._extract_mix(playlist_id)
 997         if playlist_id.startswith('TL'):
 998             raise ExtractorError(u'For downloading YouTube.com top lists, use '
 999                 u'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True)
1000
1001         url = self._TEMPLATE_URL % playlist_id
1002         page = self._download_webpage(url, playlist_id)
1003         more_widget_html = content_html = page
1004
1005         # Check if the playlist exists or is private
1006         if re.search(r'<div class="yt-alert-message">[^<]*?(The|This) playlist (does not exist|is private)[^<]*?</div>', page) is not None:
1007             raise ExtractorError(
1008                 u'The playlist doesn\'t exist or is private, use --username or '
1009                 '--netrc to access it.',
1010                 expected=True)
1011
1012         # Extract the video ids from the playlist pages
1013         ids = []
1014
1015         for page_num in itertools.count(1):
1016             matches = re.finditer(self._VIDEO_RE, content_html)
1017             # We remove the duplicates and the link with index 0
1018             # (it's not the first video of the playlist)
1019             new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
1020             ids.extend(new_ids)
1021
1022             mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
1023             if not mobj:
1024                 break
1025
1026             more = self._download_json(
1027                 'https://youtube.com/%s' % mobj.group('more'), playlist_id,
1028                 'Downloading page #%s' % page_num,
1029                 transform_source=uppercase_escape)
1030             content_html = more['content_html']
1031             more_widget_html = more['load_more_widget_html']
1032
1033         playlist_title = self._html_search_regex(
1034             r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>',
1035             page, u'title')
1036
1037         url_results = self._ids_to_results(ids)
1038         return self.playlist_result(url_results, playlist_id, playlist_title)
1039
1040
1041 class YoutubeTopListIE(YoutubePlaylistIE):
1042     IE_NAME = u'youtube:toplist'
1043     IE_DESC = (u'YouTube.com top lists, "yttoplist:{channel}:{list title}"'
1044         u' (Example: "yttoplist:music:Top Tracks")')
1045     _VALID_URL = r'yttoplist:(?P<chann>.*?):(?P<title>.*?)$'
1046
1047     def _real_extract(self, url):
1048         mobj = re.match(self._VALID_URL, url)
1049         channel = mobj.group('chann')
1050         title = mobj.group('title')
1051         query = compat_urllib_parse.urlencode({'title': title})
1052         playlist_re = 'href="([^"]+?%s.*?)"' % re.escape(query)
1053         channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title)
1054         link = self._html_search_regex(playlist_re, channel_page, u'list')
1055         url = compat_urlparse.urljoin('https://www.youtube.com/', link)
1056
1057         video_re = r'data-index="\d+".*?data-video-id="([0-9A-Za-z_-]{11})"'
1058         ids = []
1059         # sometimes the webpage doesn't contain the videos
1060         # retry until we get them
1061         for i in itertools.count(0):
1062             msg = u'Downloading Youtube mix'
1063             if i > 0:
1064                 msg += ', retry #%d' % i
1065             webpage = self._download_webpage(url, title, msg)
1066             ids = orderedSet(re.findall(video_re, webpage))
1067             if ids:
1068                 break
1069         url_results = self._ids_to_results(ids)
1070         return self.playlist_result(url_results, playlist_title=title)
1071
1072
1073 class YoutubeChannelIE(InfoExtractor):
1074     IE_DESC = u'YouTube.com channels'
1075     _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
1076     _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
1077     _MORE_PAGES_URL = 'https://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s'
1078     IE_NAME = u'youtube:channel'
1079
1080     def extract_videos_from_page(self, page):
1081         ids_in_page = []
1082         for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
1083             if mobj.group(1) not in ids_in_page:
1084                 ids_in_page.append(mobj.group(1))
1085         return ids_in_page
1086
1087     def _real_extract(self, url):
1088         # Extract channel id
1089         mobj = re.match(self._VALID_URL, url)
1090         if mobj is None:
1091             raise ExtractorError(u'Invalid URL: %s' % url)
1092
1093         # Download channel page
1094         channel_id = mobj.group(1)
1095         video_ids = []
1096         url = 'https://www.youtube.com/channel/%s/videos' % channel_id
1097         channel_page = self._download_webpage(url, channel_id)
1098         autogenerated = re.search(r'''(?x)
1099                 class="[^"]*?(?:
1100                     channel-header-autogenerated-label|
1101                     yt-channel-title-autogenerated
1102                 )[^"]*"''', channel_page) is not None
1103
1104         if autogenerated:
1105             # The videos are contained in a single page
1106             # the ajax pages can't be used, they are empty
1107             video_ids = self.extract_videos_from_page(channel_page)
1108         else:
1109             # Download all channel pages using the json-based channel_ajax query
1110             for pagenum in itertools.count(1):
1111                 url = self._MORE_PAGES_URL % (pagenum, channel_id)
1112                 page = self._download_json(
1113                     url, channel_id, note=u'Downloading page #%s' % pagenum,
1114                     transform_source=uppercase_escape)
1115
1116                 ids_in_page = self.extract_videos_from_page(page['content_html'])
1117                 video_ids.extend(ids_in_page)
1118
1119                 if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
1120                     break
1121
1122         self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
1123
1124         url_entries = [self.url_result(video_id, 'Youtube', video_id=video_id)
1125                        for video_id in video_ids]
1126         return self.playlist_result(url_entries, channel_id)
1127
1128
1129 class YoutubeUserIE(InfoExtractor):
1130     IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
1131     _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
1132     _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s'
1133     _GDATA_PAGE_SIZE = 50
1134     _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
1135     IE_NAME = u'youtube:user'
1136
1137     @classmethod
1138     def suitable(cls, url):
1139         # Don't return True if the url can be extracted with other youtube
1140         # extractor, the regex would is too permissive and it would match.
1141         other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1142         if any(ie.suitable(url) for ie in other_ies): return False
1143         else: return super(YoutubeUserIE, cls).suitable(url)
1144
1145     def _real_extract(self, url):
1146         # Extract username
1147         mobj = re.match(self._VALID_URL, url)
1148         if mobj is None:
1149             raise ExtractorError(u'Invalid URL: %s' % url)
1150
1151         username = mobj.group(1)
1152
1153         # Download video ids using YouTube Data API. Result size per
1154         # query is limited (currently to 50 videos) so we need to query
1155         # page by page until there are no video ids - it means we got
1156         # all of them.
1157
1158         def download_page(pagenum):
1159             start_index = pagenum * self._GDATA_PAGE_SIZE + 1
1160
1161             gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
1162             page = self._download_webpage(
1163                 gdata_url, username,
1164                 u'Downloading video ids from %d to %d' % (
1165                     start_index, start_index + self._GDATA_PAGE_SIZE))
1166
1167             try:
1168                 response = json.loads(page)
1169             except ValueError as err:
1170                 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1171             if 'entry' not in response['feed']:
1172                 return
1173
1174             # Extract video identifiers
1175             entries = response['feed']['entry']
1176             for entry in entries:
1177                 title = entry['title']['$t']
1178                 video_id = entry['id']['$t'].split('/')[-1]
1179                 yield {
1180                     '_type': 'url',
1181                     'url': video_id,
1182                     'ie_key': 'Youtube',
1183                     'id': video_id,
1184                     'title': title,
1185                 }
1186         url_results = PagedList(download_page, self._GDATA_PAGE_SIZE)
1187
1188         return self.playlist_result(url_results, playlist_title=username)
1189
1190
1191 class YoutubeSearchIE(SearchInfoExtractor):
1192     IE_DESC = u'YouTube.com searches'
1193     _API_URL = u'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
1194     _MAX_RESULTS = 1000
1195     IE_NAME = u'youtube:search'
1196     _SEARCH_KEY = 'ytsearch'
1197
1198     def _get_n_results(self, query, n):
1199         """Get a specified number of results for a query"""
1200
1201         video_ids = []
1202         pagenum = 0
1203         limit = n
1204         PAGE_SIZE = 50
1205
1206         while (PAGE_SIZE * pagenum) < limit:
1207             result_url = self._API_URL % (
1208                 compat_urllib_parse.quote_plus(query.encode('utf-8')),
1209                 (PAGE_SIZE * pagenum) + 1)
1210             data_json = self._download_webpage(
1211                 result_url, video_id=u'query "%s"' % query,
1212                 note=u'Downloading page %s' % (pagenum + 1),
1213                 errnote=u'Unable to download API page')
1214             data = json.loads(data_json)
1215             api_response = data['data']
1216
1217             if 'items' not in api_response:
1218                 raise ExtractorError(
1219                     u'[youtube] No video results', expected=True)
1220
1221             new_ids = list(video['id'] for video in api_response['items'])
1222             video_ids += new_ids
1223
1224             limit = min(n, api_response['totalItems'])
1225             pagenum += 1
1226
1227         if len(video_ids) > n:
1228             video_ids = video_ids[:n]
1229         videos = [self.url_result(video_id, 'Youtube', video_id=video_id)
1230                   for video_id in video_ids]
1231         return self.playlist_result(videos, query)
1232
1233
1234 class YoutubeSearchDateIE(YoutubeSearchIE):
1235     IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
1236     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published'
1237     _SEARCH_KEY = 'ytsearchdate'
1238     IE_DESC = u'YouTube.com searches, newest videos first'
1239
1240
1241 class YoutubeSearchURLIE(InfoExtractor):
1242     IE_DESC = u'YouTube.com search URLs'
1243     IE_NAME = u'youtube:search_url'
1244     _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?search_query=(?P<query>[^&]+)(?:[&]|$)'
1245
1246     def _real_extract(self, url):
1247         mobj = re.match(self._VALID_URL, url)
1248         query = compat_urllib_parse.unquote_plus(mobj.group('query'))
1249
1250         webpage = self._download_webpage(url, query)
1251         result_code = self._search_regex(
1252             r'(?s)<ol class="item-section"(.*?)</ol>', webpage, u'result HTML')
1253
1254         part_codes = re.findall(
1255             r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code)
1256         entries = []
1257         for part_code in part_codes:
1258             part_title = self._html_search_regex(
1259                 [r'(?s)title="([^"]+)"', r'>([^<]+)</a>'], part_code, 'item title', fatal=False)
1260             part_url_snippet = self._html_search_regex(
1261                 r'(?s)href="([^"]+)"', part_code, 'item URL')
1262             part_url = compat_urlparse.urljoin(
1263                 'https://www.youtube.com/', part_url_snippet)
1264             entries.append({
1265                 '_type': 'url',
1266                 'url': part_url,
1267                 'title': part_title,
1268             })
1269
1270         return {
1271             '_type': 'playlist',
1272             'entries': entries,
1273             'title': query,
1274         }
1275
1276
1277 class YoutubeShowIE(InfoExtractor):
1278     IE_DESC = u'YouTube.com (multi-season) shows'
1279     _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
1280     IE_NAME = u'youtube:show'
1281
1282     def _real_extract(self, url):
1283         mobj = re.match(self._VALID_URL, url)
1284         show_name = mobj.group(1)
1285         webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
1286         # There's one playlist for each season of the show
1287         m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1288         self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
1289         return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
1290
1291
1292 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1293     """
1294     Base class for extractors that fetch info from
1295     http://www.youtube.com/feed_ajax
1296     Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1297     """
1298     _LOGIN_REQUIRED = True
1299     # use action_load_personal_feed instead of action_load_system_feed
1300     _PERSONAL_FEED = False
1301
1302     @property
1303     def _FEED_TEMPLATE(self):
1304         action = 'action_load_system_feed'
1305         if self._PERSONAL_FEED:
1306             action = 'action_load_personal_feed'
1307         return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
1308
1309     @property
1310     def IE_NAME(self):
1311         return u'youtube:%s' % self._FEED_NAME
1312
1313     def _real_initialize(self):
1314         self._login()
1315
1316     def _real_extract(self, url):
1317         feed_entries = []
1318         paging = 0
1319         for i in itertools.count(1):
1320             info = self._download_json(self._FEED_TEMPLATE % paging,
1321                                           u'%s feed' % self._FEED_NAME,
1322                                           u'Downloading page %s' % i)
1323             feed_html = info.get('feed_html') or info.get('content_html')
1324             m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
1325             ids = orderedSet(m.group(1) for m in m_ids)
1326             feed_entries.extend(
1327                 self.url_result(video_id, 'Youtube', video_id=video_id)
1328                 for video_id in ids)
1329             mobj = re.search(
1330                 r'data-uix-load-more-href="/?[^"]+paging=(?P<paging>\d+)',
1331                 feed_html)
1332             if mobj is None:
1333                 break
1334             paging = mobj.group('paging')
1335         return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
1336
1337 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1338     IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
1339     _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1340     _FEED_NAME = 'subscriptions'
1341     _PLAYLIST_TITLE = u'Youtube Subscriptions'
1342
1343 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1344     IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
1345     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1346     _FEED_NAME = 'recommended'
1347     _PLAYLIST_TITLE = u'Youtube Recommended videos'
1348
1349 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
1350     IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
1351     _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
1352     _FEED_NAME = 'watch_later'
1353     _PLAYLIST_TITLE = u'Youtube Watch Later'
1354     _PERSONAL_FEED = True
1355
1356 class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
1357     IE_DESC = u'Youtube watch history, "ythistory" keyword (requires authentication)'
1358     _VALID_URL = u'https?://www\.youtube\.com/feed/history|:ythistory'
1359     _FEED_NAME = 'history'
1360     _PERSONAL_FEED = True
1361     _PLAYLIST_TITLE = u'Youtube Watch History'
1362
1363 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1364     IE_NAME = u'youtube:favorites'
1365     IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
1366     _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1367     _LOGIN_REQUIRED = True
1368
1369     def _real_extract(self, url):
1370         webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1371         playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
1372         return self.url_result(playlist_id, 'YoutubePlaylist')
1373
1374
1375 class YoutubeTruncatedURLIE(InfoExtractor):
1376     IE_NAME = 'youtube:truncated_url'
1377     IE_DESC = False  # Do not list
1378     _VALID_URL = r'''(?x)
1379         (?:https?://)?[^/]+/watch\?(?:
1380             feature=[a-z_]+|
1381             annotation_id=annotation_[^&]+
1382         )?$|
1383         (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$
1384     '''
1385
1386     _TESTS = [{
1387         'url': 'http://www.youtube.com/watch?annotation_id=annotation_3951667041',
1388         'only_matching': True,
1389     }, {
1390         'url': 'http://www.youtube.com/watch?',
1391         'only_matching': True,
1392     }]
1393
1394     def _real_extract(self, url):
1395         raise ExtractorError(
1396             u'Did you forget to quote the URL? Remember that & is a meta '
1397             u'character in most shells, so you want to put the URL in quotes, '
1398             u'like  youtube-dl '
1399             u'"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
1400             u' or simply  youtube-dl BaW_jenozKc  .',
1401             expected=True)