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     _TWOFACTOR_URL = 'https://accounts.google.com/SecondFactor'
  41     _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1'
  42     _AGE_URL = 'https://www.youtube.com/verify_age?next_url=/&gl=US&hl=en'
  43     _NETRC_MACHINE = 'youtube'
  44     # If True it will raise an error if no login info is provided
  45     _LOGIN_REQUIRED = False
  46
  47     def _set_language(self):
  48         return bool(self._download_webpage(
  49             self._LANG_URL, None,
  50             note=u'Setting language', errnote='unable to set language',
  51             fatal=False))
  52
  53     def _login(self):
  54         """
  55         Attempt to log in to YouTube.
  56         True is returned if successful or skipped.
  57         False is returned if login failed.
  58
  59         If _LOGIN_REQUIRED is set and no authentication was provided, an error is raised.
  60         """
  61         (username, password) = self._get_login_info()
  62         # No authentication to be performed
  63         if username is None:
  64             if self._LOGIN_REQUIRED:
  65                 raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True)
  66             return True
  67
  68         login_page = self._download_webpage(
  69             self._LOGIN_URL, None,
  70             note=u'Downloading login page',
  71             errnote=u'unable to fetch login page', fatal=False)
  72         if login_page is False:
  73             return
  74
  75         galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"',
  76                                   login_page, u'Login GALX parameter')
  77
  78         # Log in
  79         login_form_strs = {
  80                 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
  81                 u'Email': username,
  82                 u'GALX': galx,
  83                 u'Passwd': password,
  84
  85                 u'PersistentCookie': u'yes',
  86                 u'_utf8': u'霱',
  87                 u'bgresponse': u'js_disabled',
  88                 u'checkConnection': u'',
  89                 u'checkedDomains': u'youtube',
  90                 u'dnConn': u'',
  91                 u'pstMsg': u'0',
  92                 u'rmShown': u'1',
  93                 u'secTok': u'',
  94                 u'signIn': u'Sign in',
  95                 u'timeStmp': u'',
  96                 u'service': u'youtube',
  97                 u'uilel': u'3',
  98                 u'hl': u'en_US',
  99         }
 100
 101         # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
 102         # chokes on unicode
 103         login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items())
 104         login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
 105
 106         req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
 107         login_results = self._download_webpage(
 108             req, None,
 109             note=u'Logging in', errnote=u'unable to log in', fatal=False)
 110         if login_results is False:
 111             return False
 112
 113         if re.search(r'id="errormsg_0_Passwd"', login_results) is not None:
 114             raise ExtractorError(u'Please use your account password and a two-factor code instead of an application-specific password.', expected=True)
 115
 116         # Two-Factor
 117         # TODO add SMS and phone call support - these require making a request and then prompting the user
 118
 119         if re.search(r'(?i)<form[^>]* id="gaia_secondfactorform"', login_results) is not None:
 120             tfa_code = self._get_tfa_info()
 121
 122             if tfa_code is None:
 123                 self._downloader.report_warning(u'Two-factor authentication required. Provide it with --twofactor <code>')
 124                 self._downloader.report_warning(u'(Note that only TOTP (Google Authenticator App) codes work at this time.)')
 125                 return False
 126
 127             # Unlike the first login form, secTok and timeStmp are both required for the TFA form
 128
 129             match = re.search(r'id="secTok"\n\s+value=\'(.+)\'/>', login_results, re.M | re.U)
 130             if match is None:
 131                 self._downloader.report_warning(u'Failed to get secTok - did the page structure change?')
 132             secTok = match.group(1)
 133             match = re.search(r'id="timeStmp"\n\s+value=\'(.+)\'/>', login_results, re.M | re.U)
 134             if match is None:
 135                 self._downloader.report_warning(u'Failed to get timeStmp - did the page structure change?')
 136             timeStmp = match.group(1)
 137
 138             tfa_form_strs = {
 139                 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
 140                 u'smsToken': u'',
 141                 u'smsUserPin': tfa_code,
 142                 u'smsVerifyPin': u'Verify',
 143
 144                 u'PersistentCookie': u'yes',
 145                 u'checkConnection': u'',
 146                 u'checkedDomains': u'youtube',
 147                 u'pstMsg': u'1',
 148                 u'secTok': secTok,
 149                 u'timeStmp': timeStmp,
 150                 u'service': u'youtube',
 151                 u'hl': u'en_US',
 152             }
 153             tfa_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in tfa_form_strs.items())
 154             tfa_data = compat_urllib_parse.urlencode(tfa_form).encode('ascii')
 155
 156             tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data)
 157             tfa_results = self._download_webpage(
 158                 tfa_req, None,
 159                 note=u'Submitting TFA code', errnote=u'unable to submit tfa', fatal=False)
 160
 161             if tfa_results is False:
 162                 return False
 163
 164             if re.search(r'(?i)<form[^>]* id="gaia_secondfactorform"', tfa_results) is not None:
 165                 self._downloader.report_warning(u'Two-factor code expired. Please try again, or use a one-use backup code instead.')
 166                 return False
 167             if re.search(r'(?i)<form[^>]* id="gaia_loginform"', tfa_results) is not None:
 168                 self._downloader.report_warning(u'unable to log in - did the page structure change?')
 169                 return False
 170             if re.search(r'smsauth-interstitial-reviewsettings', tfa_results) is not None:
 171                 self._downloader.report_warning(u'Your Google account has a security notice. Please log in on your web browser, resolve the notice, and try again.')
 172                 return False
 173
 174         if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
 175             self._downloader.report_warning(u'unable to log in: bad username or password')
 176             return False
 177         return True
 178
 179     def _confirm_age(self):
 180         age_form = {
 181             'next_url': '/',
 182             'action_confirm': 'Confirm',
 183         }
 184         req = compat_urllib_request.Request(self._AGE_URL,
 185             compat_urllib_parse.urlencode(age_form).encode('ascii'))
 186
 187         self._download_webpage(
 188             req, None,
 189             note=u'Confirming age', errnote=u'Unable to confirm age')
 190         return True
 191
 192     def _real_initialize(self):
 193         if self._downloader is None:
 194             return
 195         if not self._set_language():
 196             return
 197         if not self._login():
 198             return
 199         self._confirm_age()
 200
 201
 202 class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
 203     IE_DESC = u'YouTube.com'
 204     _VALID_URL = r"""(?x)^
 205                      (
 206                          (?:https?://|//)?                                    # http(s):// or protocol-independent URL (optional)
 207                          (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/|
 208                             (?:www\.)?deturl\.com/www\.youtube\.com/|
 209                             (?:www\.)?pwnyoutube\.com/|
 210                             (?:www\.)?yourepeat\.com/|
 211                             tube\.majestyc\.net/|
 212                             youtube\.googleapis\.com/)                        # the various hostnames, with wildcard subdomains
 213                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
 214                          (?:                                                  # the various things that can precede the ID:
 215                              (?:(?:v|embed|e)/)                               # v/ or embed/ or e/
 216                              |(?:                                             # or the v= param in all its forms
 217                                  (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)?  # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
 218                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
 219                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
 220                                  v=
 221                              )
 222                          ))
 223                          |youtu\.be/                                          # just youtu.be/xxxx
 224                          |https?://(?:www\.)?cleanvideosearch\.com/media/action/yt/watch\?videoId=
 225                          )
 226                      )?                                                       # all until now is optional -> you can pass the naked ID
 227                      ([0-9A-Za-z_-]{11})                                      # here is it! the YouTube video ID
 228                      (?(1).+)?                                                # if we found the ID, everything can follow
 229                      $"""
 230     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
 231     _formats = {
 232         '5': {'ext': 'flv', 'width': 400, 'height': 240},
 233         '6': {'ext': 'flv', 'width': 450, 'height': 270},
 234         '13': {'ext': '3gp'},
 235         '17': {'ext': '3gp', 'width': 176, 'height': 144},
 236         '18': {'ext': 'mp4', 'width': 640, 'height': 360},
 237         '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
 238         '34': {'ext': 'flv', 'width': 640, 'height': 360},
 239         '35': {'ext': 'flv', 'width': 854, 'height': 480},
 240         '36': {'ext': '3gp', 'width': 320, 'height': 240},
 241         '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
 242         '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
 243         '43': {'ext': 'webm', 'width': 640, 'height': 360},
 244         '44': {'ext': 'webm', 'width': 854, 'height': 480},
 245         '45': {'ext': 'webm', 'width': 1280, 'height': 720},
 246         '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
 247
 248
 249         # 3d videos
 250         '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20},
 251         '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20},
 252         '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20},
 253         '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20},
 254         '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20},
 255         '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20},
 256         '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20},
 257
 258         # Apple HTTP Live Streaming
 259         '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 260         '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10},
 261         '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10},
 262         '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10},
 263         '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10},
 264         '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10},
 265         '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10},
 266
 267         # DASH mp4 video
 268         '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 269         '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 270         '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 271         '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 272         '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 273         '138': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 274         '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 275         '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
 276
 277         # Dash mp4 audio
 278         '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50},
 279         '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50},
 280         '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50},
 281
 282         # Dash webm
 283         '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 284         '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 285         '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 286         '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 287         '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 288         '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
 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
 299         # Dash webm audio
 300         '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50},
 301         '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50},
 302
 303         # RTMP (unnamed)
 304         '_rtmp': {'protocol': 'rtmp'},
 305     }
 306
 307     IE_NAME = u'youtube'
 308     _TESTS = [
 309         {
 310             u"url":  u"http://www.youtube.com/watch?v=BaW_jenozKc",
 311             u"file":  u"BaW_jenozKc.mp4",
 312             u"info_dict": {
 313                 u"title": u"youtube-dl test video \"'/\\ä↭𝕐",
 314                 u"uploader": u"Philipp Hagemeister",
 315                 u"uploader_id": u"phihag",
 316                 u"upload_date": u"20121002",
 317                 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 .",
 318                 u"categories": [u'Science & Technology'],
 319                 'like_count': int,
 320                 'dislike_count': int,
 321             }
 322         },
 323         {
 324             u"url":  u"http://www.youtube.com/watch?v=UxxajLWwzqY",
 325             u"file":  u"UxxajLWwzqY.mp4",
 326             u"note": u"Test generic use_cipher_signature video (#897)",
 327             u"info_dict": {
 328                 u"upload_date": u"20120506",
 329                 u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
 330                 u"description": u"md5:fea86fda2d5a5784273df5c7cc994d9f",
 331                 u"uploader": u"Icona Pop",
 332                 u"uploader_id": u"IconaPop"
 333             }
 334         },
 335         {
 336             u"url":  u"https://www.youtube.com/watch?v=07FYdnEawAQ",
 337             u"file":  u"07FYdnEawAQ.mp4",
 338             u"note": u"Test VEVO video with age protection (#956)",
 339             u"info_dict": {
 340                 u"upload_date": u"20130703",
 341                 u"title": u"Justin Timberlake - Tunnel Vision (Explicit)",
 342                 u"description": u"md5:64249768eec3bc4276236606ea996373",
 343                 u"uploader": u"justintimberlakeVEVO",
 344                 u"uploader_id": u"justintimberlakeVEVO"
 345             }
 346         },
 347         {
 348             u"url":  u"//www.YouTube.com/watch?v=yZIXLfi8CZQ",
 349             u"file":  u"yZIXLfi8CZQ.mp4",
 350             u"note": u"Embed-only video (#1746)",
 351             u"info_dict": {
 352                 u"upload_date": u"20120608",
 353                 u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012",
 354                 u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7",
 355                 u"uploader": u"SET India",
 356                 u"uploader_id": u"setindia"
 357             }
 358         },
 359         {
 360             u"url": u"http://www.youtube.com/watch?v=a9LDPn-MO4I",
 361             u"file": u"a9LDPn-MO4I.m4a",
 362             u"note": u"256k DASH audio (format 141) via DASH manifest",
 363             u"info_dict": {
 364                 u"upload_date": "20121002",
 365                 u"uploader_id": "8KVIDEO",
 366                 u"description": "No description available.",
 367                 u"uploader": "8KVIDEO",
 368                 u"title": "UHDTV TEST 8K VIDEO.mp4"
 369             },
 370             u"params": {
 371                 u"youtube_include_dash_manifest": True,
 372                 u"format": "141",
 373             },
 374         },
 375         # DASH manifest with encrypted signature
 376         {
 377             u'url': u'https://www.youtube.com/watch?v=IB3lcPjvWLA',
 378             u'info_dict': {
 379                 u'id': u'IB3lcPjvWLA',
 380                 u'ext': u'm4a',
 381                 u'title': u'Afrojack - The Spark ft. Spree Wilson',
 382                 u'description': u'md5:9717375db5a9a3992be4668bbf3bc0a8',
 383                 u'uploader': u'AfrojackVEVO',
 384                 u'uploader_id': u'AfrojackVEVO',
 385                 u'upload_date': u'20131011',
 386             },
 387             u"params": {
 388                 u'youtube_include_dash_manifest': True,
 389                 u'format': '141',
 390             },
 391         },
 392     ]
 393
 394
 395     @classmethod
 396     def suitable(cls, url):
 397         """Receives a URL and returns True if suitable for this IE."""
 398         if YoutubePlaylistIE.suitable(url): return False
 399         return re.match(cls._VALID_URL, url) is not None
 400
 401     def __init__(self, *args, **kwargs):
 402         super(YoutubeIE, self).__init__(*args, **kwargs)
 403         self._player_cache = {}
 404
 405     def report_video_info_webpage_download(self, video_id):
 406         """Report attempt to download video info webpage."""
 407         self.to_screen(u'%s: Downloading video info webpage' % video_id)
 408
 409     def report_information_extraction(self, video_id):
 410         """Report attempt to extract video information."""
 411         self.to_screen(u'%s: Extracting video information' % video_id)
 412
 413     def report_unavailable_format(self, video_id, format):
 414         """Report extracted video URL."""
 415         self.to_screen(u'%s: Format %s not available' % (video_id, format))
 416
 417     def report_rtmp_download(self):
 418         """Indicate the download will use the RTMP protocol."""
 419         self.to_screen(u'RTMP download detected')
 420
 421     def _signature_cache_id(self, example_sig):
 422         """ Return a string representation of a signature """
 423         return u'.'.join(compat_str(len(part)) for part in example_sig.split('.'))
 424
 425     def _extract_signature_function(self, video_id, player_url, example_sig):
 426         id_m = re.match(
 427             r'.*-(?P<id>[a-zA-Z0-9_-]+)(?:/watch_as3|/html5player)?\.(?P<ext>[a-z]+)$',
 428             player_url)
 429         if not id_m:
 430             raise ExtractorError('Cannot identify player %r' % player_url)
 431         player_type = id_m.group('ext')
 432         player_id = id_m.group('id')
 433
 434         # Read from filesystem cache
 435         func_id = '%s_%s_%s' % (
 436             player_type, player_id, self._signature_cache_id(example_sig))
 437         assert os.path.basename(func_id) == func_id
 438         cache_dir = get_cachedir(self._downloader.params)
 439
 440         cache_enabled = cache_dir is not None
 441         if cache_enabled:
 442             cache_fn = os.path.join(os.path.expanduser(cache_dir),
 443                                     u'youtube-sigfuncs',
 444                                     func_id + '.json')
 445             try:
 446                 with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
 447                     cache_spec = json.load(cachef)
 448                 return lambda s: u''.join(s[i] for i in cache_spec)
 449             except IOError:
 450                 pass  # No cache available
 451             except ValueError:
 452                 try:
 453                     file_size = os.path.getsize(cache_fn)
 454                 except (OSError, IOError) as oe:
 455                     file_size = str(oe)
 456                 self._downloader.report_warning(
 457                     u'Cache %s failed (%s)' % (cache_fn, file_size))
 458
 459         if player_type == 'js':
 460             code = self._download_webpage(
 461                 player_url, video_id,
 462                 note=u'Downloading %s player %s' % (player_type, player_id),
 463                 errnote=u'Download of %s failed' % player_url)
 464             res = self._parse_sig_js(code)
 465         elif player_type == 'swf':
 466             urlh = self._request_webpage(
 467                 player_url, video_id,
 468                 note=u'Downloading %s player %s' % (player_type, player_id),
 469                 errnote=u'Download of %s failed' % player_url)
 470             code = urlh.read()
 471             res = self._parse_sig_swf(code)
 472         else:
 473             assert False, 'Invalid player type %r' % player_type
 474
 475         if cache_enabled:
 476             try:
 477                 test_string = u''.join(map(compat_chr, range(len(example_sig))))
 478                 cache_res = res(test_string)
 479                 cache_spec = [ord(c) for c in cache_res]
 480                 try:
 481                     os.makedirs(os.path.dirname(cache_fn))
 482                 except OSError as ose:
 483                     if ose.errno != errno.EEXIST:
 484                         raise
 485                 write_json_file(cache_spec, cache_fn)
 486             except Exception:
 487                 tb = traceback.format_exc()
 488                 self._downloader.report_warning(
 489                     u'Writing cache to %r failed: %s' % (cache_fn, tb))
 490
 491         return res
 492
 493     def _print_sig_code(self, func, example_sig):
 494         def gen_sig_code(idxs):
 495             def _genslice(start, end, step):
 496                 starts = u'' if start == 0 else str(start)
 497                 ends = (u':%d' % (end+step)) if end + step >= 0 else u':'
 498                 steps = u'' if step == 1 else (u':%d' % step)
 499                 return u's[%s%s%s]' % (starts, ends, steps)
 500
 501             step = None
 502             start = '(Never used)'  # Quelch pyflakes warnings - start will be
 503                                     # set as soon as step is set
 504             for i, prev in zip(idxs[1:], idxs[:-1]):
 505                 if step is not None:
 506                     if i - prev == step:
 507                         continue
 508                     yield _genslice(start, prev, step)
 509                     step = None
 510                     continue
 511                 if i - prev in [-1, 1]:
 512                     step = i - prev
 513                     start = prev
 514                     continue
 515                 else:
 516                     yield u's[%d]' % prev
 517             if step is None:
 518                 yield u's[%d]' % i
 519             else:
 520                 yield _genslice(start, i, step)
 521
 522         test_string = u''.join(map(compat_chr, range(len(example_sig))))
 523         cache_res = func(test_string)
 524         cache_spec = [ord(c) for c in cache_res]
 525         expr_code = u' + '.join(gen_sig_code(cache_spec))
 526         signature_id_tuple = '(%s)' % (
 527             ', '.join(compat_str(len(p)) for p in example_sig.split('.')))
 528         code = (u'if tuple(len(p) for p in s.split(\'.\')) == %s:\n'
 529                 u'    return %s\n') % (signature_id_tuple, expr_code)
 530         self.to_screen(u'Extracted signature function:\n' + code)
 531
 532     def _parse_sig_js(self, jscode):
 533         funcname = self._search_regex(
 534             r'signature=([$a-zA-Z]+)', jscode,
 535              u'Initial JS player signature function name')
 536
 537         jsi = JSInterpreter(jscode)
 538         initial_function = jsi.extract_function(funcname)
 539         return lambda s: initial_function([s])
 540
 541     def _parse_sig_swf(self, file_contents):
 542         swfi = SWFInterpreter(file_contents)
 543         TARGET_CLASSNAME = u'SignatureDecipher'
 544         searched_class = swfi.extract_class(TARGET_CLASSNAME)
 545         initial_function = swfi.extract_function(searched_class, u'decipher')
 546         return lambda s: initial_function([s])
 547
 548     def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
 549         """Turn the encrypted s field into a working signature"""
 550
 551         if player_url is None:
 552             raise ExtractorError(u'Cannot decrypt signature without player_url')
 553
 554         if player_url.startswith(u'//'):
 555             player_url = u'https:' + player_url
 556         try:
 557             player_id = (player_url, self._signature_cache_id(s))
 558             if player_id not in self._player_cache:
 559                 func = self._extract_signature_function(
 560                     video_id, player_url, s
 561                 )
 562                 self._player_cache[player_id] = func
 563             func = self._player_cache[player_id]
 564             if self._downloader.params.get('youtube_print_sig_code'):
 565                 self._print_sig_code(func, s)
 566             return func(s)
 567         except Exception as e:
 568             tb = traceback.format_exc()
 569             raise ExtractorError(
 570                 u'Signature extraction failed: ' + tb, cause=e)
 571
 572     def _get_available_subtitles(self, video_id, webpage):
 573         try:
 574             sub_list = self._download_webpage(
 575                 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
 576                 video_id, note=False)
 577         except ExtractorError as err:
 578             self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
 579             return {}
 580         lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
 581
 582         sub_lang_list = {}
 583         for l in lang_list:
 584             lang = l[1]
 585             if lang in sub_lang_list:
 586                 continue
 587             params = compat_urllib_parse.urlencode({
 588                 'lang': lang,
 589                 'v': video_id,
 590                 'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
 591                 'name': unescapeHTML(l[0]).encode('utf-8'),
 592             })
 593             url = u'https://www.youtube.com/api/timedtext?' + params
 594             sub_lang_list[lang] = url
 595         if not sub_lang_list:
 596             self._downloader.report_warning(u'video doesn\'t have subtitles')
 597             return {}
 598         return sub_lang_list
 599
 600     def _get_available_automatic_caption(self, video_id, webpage):
 601         """We need the webpage for getting the captions url, pass it as an
 602            argument to speed up the process."""
 603         sub_format = self._downloader.params.get('subtitlesformat', 'srt')
 604         self.to_screen(u'%s: Looking for automatic captions' % video_id)
 605         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
 606         err_msg = u'Couldn\'t find automatic captions for %s' % video_id
 607         if mobj is None:
 608             self._downloader.report_warning(err_msg)
 609             return {}
 610         player_config = json.loads(mobj.group(1))
 611         try:
 612             args = player_config[u'args']
 613             caption_url = args[u'ttsurl']
 614             timestamp = args[u'timestamp']
 615             # We get the available subtitles
 616             list_params = compat_urllib_parse.urlencode({
 617                 'type': 'list',
 618                 'tlangs': 1,
 619                 'asrs': 1,
 620             })
 621             list_url = caption_url + '&' + list_params
 622             caption_list = self._download_xml(list_url, video_id)
 623             original_lang_node = caption_list.find('track')
 624             if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' :
 625                 self._downloader.report_warning(u'Video doesn\'t have automatic captions')
 626                 return {}
 627             original_lang = original_lang_node.attrib['lang_code']
 628
 629             sub_lang_list = {}
 630             for lang_node in caption_list.findall('target'):
 631                 sub_lang = lang_node.attrib['lang_code']
 632                 params = compat_urllib_parse.urlencode({
 633                     'lang': original_lang,
 634                     'tlang': sub_lang,
 635                     'fmt': sub_format,
 636                     'ts': timestamp,
 637                     'kind': 'asr',
 638                 })
 639                 sub_lang_list[sub_lang] = caption_url + '&' + params
 640             return sub_lang_list
 641         # An extractor error can be raise by the download process if there are
 642         # no automatic captions but there are subtitles
 643         except (KeyError, ExtractorError):
 644             self._downloader.report_warning(err_msg)
 645             return {}
 646
 647     @classmethod
 648     def extract_id(cls, url):
 649         mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
 650         if mobj is None:
 651             raise ExtractorError(u'Invalid URL: %s' % url)
 652         video_id = mobj.group(2)
 653         return video_id
 654
 655     def _extract_from_m3u8(self, manifest_url, video_id):
 656         url_map = {}
 657         def _get_urls(_manifest):
 658             lines = _manifest.split('\n')
 659             urls = filter(lambda l: l and not l.startswith('#'),
 660                             lines)
 661             return urls
 662         manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest')
 663         formats_urls = _get_urls(manifest)
 664         for format_url in formats_urls:
 665             itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
 666             url_map[itag] = format_url
 667         return url_map
 668
 669     def _extract_annotations(self, video_id):
 670         url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
 671         return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.')
 672
 673     def _real_extract(self, url):
 674         proto = (
 675             u'http' if self._downloader.params.get('prefer_insecure', False)
 676             else u'https')
 677
 678         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
 679         mobj = re.search(self._NEXT_URL_RE, url)
 680         if mobj:
 681             url = proto + '://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
 682         video_id = self.extract_id(url)
 683
 684         # Get video webpage
 685         url = proto + '://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
 686         video_webpage = self._download_webpage(url, video_id)
 687
 688         # Attempt to extract SWF player URL
 689         mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
 690         if mobj is not None:
 691             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
 692         else:
 693             player_url = None
 694
 695         # Get video info
 696         self.report_video_info_webpage_download(video_id)
 697         if re.search(r'player-age-gate-content">', video_webpage) is not None:
 698             self.report_age_confirmation()
 699             age_gate = True
 700             # We simulate the access to the video from www.youtube.com/v/{video_id}
 701             # this can be viewed without login into Youtube
 702             data = compat_urllib_parse.urlencode({
 703                 'video_id': video_id,
 704                 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
 705                 'sts': self._search_regex(
 706                     r'"sts"\s*:\s*(\d+)', video_webpage, 'sts'),
 707             })
 708             video_info_url = proto + '://www.youtube.com/get_video_info?' + data
 709             video_info_webpage = self._download_webpage(video_info_url, video_id,
 710                                     note=False,
 711                                     errnote='unable to download video info webpage')
 712             video_info = compat_parse_qs(video_info_webpage)
 713         else:
 714             age_gate = False
 715             for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
 716                 video_info_url = (proto + '://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
 717                         % (video_id, el_type))
 718                 video_info_webpage = self._download_webpage(video_info_url, video_id,
 719                                         note=False,
 720                                         errnote='unable to download video info webpage')
 721                 video_info = compat_parse_qs(video_info_webpage)
 722                 if 'token' in video_info:
 723                     break
 724         if 'token' not in video_info:
 725             if 'reason' in video_info:
 726                 raise ExtractorError(
 727                     u'YouTube said: %s' % video_info['reason'][0],
 728                     expected=True, video_id=video_id)
 729             else:
 730                 raise ExtractorError(
 731                     u'"token" parameter not in video info for unknown reason',
 732                     video_id=video_id)
 733
 734         if 'view_count' in video_info:
 735             view_count = int(video_info['view_count'][0])
 736         else:
 737             view_count = None
 738
 739         # Check for "rental" videos
 740         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
 741             raise ExtractorError(u'"rental" videos not supported')
 742
 743         # Start extracting information
 744         self.report_information_extraction(video_id)
 745
 746         # uploader
 747         if 'author' not in video_info:
 748             raise ExtractorError(u'Unable to extract uploader name')
 749         video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
 750
 751         # uploader_id
 752         video_uploader_id = None
 753         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
 754         if mobj is not None:
 755             video_uploader_id = mobj.group(1)
 756         else:
 757             self._downloader.report_warning(u'unable to extract uploader nickname')
 758
 759         # title
 760         if 'title' in video_info:
 761             video_title = video_info['title'][0]
 762         else:
 763             self._downloader.report_warning(u'Unable to extract video title')
 764             video_title = u'_'
 765
 766         # thumbnail image
 767         # We try first to get a high quality image:
 768         m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
 769                             video_webpage, re.DOTALL)
 770         if m_thumb is not None:
 771             video_thumbnail = m_thumb.group(1)
 772         elif 'thumbnail_url' not in video_info:
 773             self._downloader.report_warning(u'unable to extract video thumbnail')
 774             video_thumbnail = None
 775         else:   # don't panic if we can't find it
 776             video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
 777
 778         # upload date
 779         upload_date = None
 780         mobj = re.search(r'(?s)id="eow-date.*?>(.*?)</span>', video_webpage)
 781         if mobj is None:
 782             mobj = re.search(
 783                 r'(?s)id="watch-uploader-info".*?>.*?(?:Published|Uploaded|Streamed live) on (.*?)</strong>',
 784                 video_webpage)
 785         if mobj is not None:
 786             upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
 787             upload_date = unified_strdate(upload_date)
 788
 789         m_cat_container = get_element_by_id("eow-category", video_webpage)
 790         if m_cat_container:
 791             category = self._html_search_regex(
 792                 r'(?s)<a[^<]+>(.*?)</a>', m_cat_container, 'category',
 793                 default=None)
 794             video_categories = None if category is None else [category]
 795         else:
 796             video_categories = None
 797
 798         # description
 799         video_description = get_element_by_id("eow-description", video_webpage)
 800         if video_description:
 801             video_description = re.sub(r'''(?x)
 802                 <a\s+
 803                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
 804                     title="([^"]+)"\s+
 805                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
 806                     class="yt-uix-redirect-link"\s*>
 807                 [^<]+
 808                 </a>
 809             ''', r'\1', video_description)
 810             video_description = clean_html(video_description)
 811         else:
 812             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
 813             if fd_mobj:
 814                 video_description = unescapeHTML(fd_mobj.group(1))
 815             else:
 816                 video_description = u''
 817
 818         def _extract_count(count_name):
 819             count = self._search_regex(
 820                 r'id="watch-%s"[^>]*>.*?([\d,]+)\s*</span>' % re.escape(count_name),
 821                 video_webpage, count_name, default=None)
 822             if count is not None:
 823                 return int(count.replace(',', ''))
 824             return None
 825         like_count = _extract_count(u'like')
 826         dislike_count = _extract_count(u'dislike')
 827
 828         # subtitles
 829         video_subtitles = self.extract_subtitles(video_id, video_webpage)
 830
 831         if self._downloader.params.get('listsubtitles', False):
 832             self._list_available_subtitles(video_id, video_webpage)
 833             return
 834
 835         if 'length_seconds' not in video_info:
 836             self._downloader.report_warning(u'unable to extract video duration')
 837             video_duration = None
 838         else:
 839             video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]))
 840
 841         # annotations
 842         video_annotations = None
 843         if self._downloader.params.get('writeannotations', False):
 844                 video_annotations = self._extract_annotations(video_id)
 845
 846         # Decide which formats to download
 847         try:
 848             mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});', video_webpage)
 849             if not mobj:
 850                 raise ValueError('Could not find vevo ID')
 851             json_code = uppercase_escape(mobj.group(1))
 852             ytplayer_config = json.loads(json_code)
 853             args = ytplayer_config['args']
 854             # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
 855             # this signatures are encrypted
 856             if 'url_encoded_fmt_stream_map' not in args:
 857                 raise ValueError(u'No stream_map present')  # caught below
 858             re_signature = re.compile(r'[&,]s=')
 859             m_s = re_signature.search(args['url_encoded_fmt_stream_map'])
 860             if m_s is not None:
 861                 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
 862                 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
 863             m_s = re_signature.search(args.get('adaptive_fmts', u''))
 864             if m_s is not None:
 865                 if 'adaptive_fmts' in video_info:
 866                     video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts']
 867                 else:
 868                     video_info['adaptive_fmts'] = [args['adaptive_fmts']]
 869         except ValueError:
 870             pass
 871
 872         def _map_to_format_list(urlmap):
 873             formats = []
 874             for itag, video_real_url in urlmap.items():
 875                 dct = {
 876                     'format_id': itag,
 877                     'url': video_real_url,
 878                     'player_url': player_url,
 879                 }
 880                 if itag in self._formats:
 881                     dct.update(self._formats[itag])
 882                 formats.append(dct)
 883             return formats
 884
 885         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
 886             self.report_rtmp_download()
 887             formats = [{
 888                 'format_id': '_rtmp',
 889                 'protocol': 'rtmp',
 890                 'url': video_info['conn'][0],
 891                 'player_url': player_url,
 892             }]
 893         elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1:
 894             encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0]
 895             if 'rtmpe%3Dyes' in encoded_url_map:
 896                 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
 897             url_map = {}
 898             for url_data_str in encoded_url_map.split(','):
 899                 url_data = compat_parse_qs(url_data_str)
 900                 if 'itag' not in url_data or 'url' not in url_data:
 901                     continue
 902                 format_id = url_data['itag'][0]
 903                 url = url_data['url'][0]
 904
 905                 if 'sig' in url_data:
 906                     url += '&signature=' + url_data['sig'][0]
 907                 elif 's' in url_data:
 908                     encrypted_sig = url_data['s'][0]
 909
 910                     if not age_gate:
 911                         jsplayer_url_json = self._search_regex(
 912                             r'"assets":.+?"js":\s*("[^"]+")',
 913                             video_webpage, u'JS player URL')
 914                         player_url = json.loads(jsplayer_url_json)
 915                     if player_url is None:
 916                         player_url_json = self._search_regex(
 917                             r'ytplayer\.config.*?"url"\s*:\s*("[^"]+")',
 918                             video_webpage, u'age gate player URL')
 919                         player_url = json.loads(player_url_json)
 920
 921                     if self._downloader.params.get('verbose'):
 922                         if player_url is None:
 923                             player_version = 'unknown'
 924                             player_desc = 'unknown'
 925                         else:
 926                             if player_url.endswith('swf'):
 927                                 player_version = self._search_regex(
 928                                     r'-(.+?)(?:/watch_as3)?\.swf$', player_url,
 929                                     u'flash player', fatal=False)
 930                                 player_desc = 'flash player %s' % player_version
 931                             else:
 932                                 player_version = self._search_regex(
 933                                     r'html5player-([^/]+?)(?:/html5player)?\.js',
 934                                     player_url,
 935                                     'html5 player', fatal=False)
 936                                 player_desc = u'html5 player %s' % player_version
 937
 938                         parts_sizes = self._signature_cache_id(encrypted_sig)
 939                         self.to_screen(u'{%s} signature length %s, %s' %
 940                             (format_id, parts_sizes, player_desc))
 941
 942                     signature = self._decrypt_signature(
 943                         encrypted_sig, video_id, player_url, age_gate)
 944                     url += '&signature=' + signature
 945                 if 'ratebypass' not in url:
 946                     url += '&ratebypass=yes'
 947                 url_map[format_id] = url
 948             formats = _map_to_format_list(url_map)
 949         elif video_info.get('hlsvp'):
 950             manifest_url = video_info['hlsvp'][0]
 951             url_map = self._extract_from_m3u8(manifest_url, video_id)
 952             formats = _map_to_format_list(url_map)
 953         else:
 954             raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
 955
 956         # Look for the DASH manifest
 957         if (self._downloader.params.get('youtube_include_dash_manifest', False)):
 958             try:
 959                 # The DASH manifest used needs to be the one from the original video_webpage.
 960                 # The one found in get_video_info seems to be using different signatures.
 961                 # However, in the case of an age restriction there won't be any embedded dashmpd in the video_webpage.
 962                 # Luckily, it seems, this case uses some kind of default signature (len == 86), so the
 963                 # combination of get_video_info and the _static_decrypt_signature() decryption fallback will work here.
 964                 if age_gate:
 965                     dash_manifest_url = video_info.get('dashmpd')[0]
 966                 else:
 967                     dash_manifest_url = ytplayer_config['args']['dashmpd']
 968                 def decrypt_sig(mobj):
 969                     s = mobj.group(1)
 970                     dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
 971                     return '/signature/%s' % dec_s
 972                 dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url)
 973                 dash_doc = self._download_xml(
 974                     dash_manifest_url, video_id,
 975                     note=u'Downloading DASH manifest',
 976                     errnote=u'Could not download DASH manifest')
 977                 for r in dash_doc.findall(u'.//{urn:mpeg:DASH:schema:MPD:2011}Representation'):
 978                     url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL')
 979                     if url_el is None:
 980                         continue
 981                     format_id = r.attrib['id']
 982                     video_url = url_el.text
 983                     filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
 984                     f = {
 985                         'format_id': format_id,
 986                         'url': video_url,
 987                         'width': int_or_none(r.attrib.get('width')),
 988                         'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
 989                         'asr': int_or_none(r.attrib.get('audioSamplingRate')),
 990                         'filesize': filesize,
 991                     }
 992                     try:
 993                         existing_format = next(
 994                             fo for fo in formats
 995                             if fo['format_id'] == format_id)
 996                     except StopIteration:
 997                         f.update(self._formats.get(format_id, {}))
 998                         formats.append(f)
 999                     else:
1000                         existing_format.update(f)
1001
1002             except (ExtractorError, KeyError) as e:
1003                 self.report_warning(u'Skipping DASH manifest: %s' % e, video_id)
1004
1005         self._sort_formats(formats)
1006
1007         return {
1008             'id':           video_id,
1009             'uploader':     video_uploader,
1010             'uploader_id':  video_uploader_id,
1011             'upload_date':  upload_date,
1012             'title':        video_title,
1013             'thumbnail':    video_thumbnail,
1014             'description':  video_description,
1015             'categories':   video_categories,
1016             'subtitles':    video_subtitles,
1017             'duration':     video_duration,
1018             'age_limit':    18 if age_gate else 0,
1019             'annotations':  video_annotations,
1020             'webpage_url': proto + '://www.youtube.com/watch?v=%s' % video_id,
1021             'view_count':   view_count,
1022             'like_count': like_count,
1023             'dislike_count': dislike_count,
1024             'formats':      formats,
1025         }
1026
1027 class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
1028     IE_DESC = u'YouTube.com playlists'
1029     _VALID_URL = r"""(?x)(?:
1030                         (?:https?://)?
1031                         (?:\w+\.)?
1032                         youtube\.com/
1033                         (?:
1034                            (?:course|view_play_list|my_playlists|artist|playlist|watch)
1035                            \? (?:.*?&)*? (?:p|a|list)=
1036                         |  p/
1037                         )
1038                         (
1039                             (?:PL|LL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,}
1040                             # Top tracks, they can also include dots
1041                             |(?:MC)[\w\.]*
1042                         )
1043                         .*
1044                      |
1045                         ((?:PL|LL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,})
1046                      )"""
1047     _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s'
1048     _MORE_PAGES_INDICATOR = r'data-link-type="next"'
1049     _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&amp;[^"]*?index=(?P<index>\d+)'
1050     IE_NAME = u'youtube:playlist'
1051
1052     def _real_initialize(self):
1053         self._login()
1054
1055     def _ids_to_results(self, ids):
1056         return [self.url_result(vid_id, 'Youtube', video_id=vid_id)
1057                        for vid_id in ids]
1058
1059     def _extract_mix(self, playlist_id):
1060         # The mixes are generated from a a single video
1061         # the id of the playlist is just 'RD' + video_id
1062         url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id)
1063         webpage = self._download_webpage(url, playlist_id, u'Downloading Youtube mix')
1064         search_title = lambda class_name: get_element_by_attribute('class', class_name, webpage)
1065         title_span = (search_title('playlist-title') or
1066             search_title('title long-title') or search_title('title'))
1067         title = clean_html(title_span)
1068         video_re = r'''(?x)data-video-username=".*?".*?
1069                        href="/watch\?v=([0-9A-Za-z_-]{11})&amp;[^"]*?list=%s''' % re.escape(playlist_id)
1070         ids = orderedSet(re.findall(video_re, webpage, flags=re.DOTALL))
1071         url_results = self._ids_to_results(ids)
1072
1073         return self.playlist_result(url_results, playlist_id, title)
1074
1075     def _real_extract(self, url):
1076         # Extract playlist id
1077         mobj = re.match(self._VALID_URL, url)
1078         if mobj is None:
1079             raise ExtractorError(u'Invalid URL: %s' % url)
1080         playlist_id = mobj.group(1) or mobj.group(2)
1081
1082         # Check if it's a video-specific URL
1083         query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1084         if 'v' in query_dict:
1085             video_id = query_dict['v'][0]
1086             if self._downloader.params.get('noplaylist'):
1087                 self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
1088                 return self.url_result(video_id, 'Youtube', video_id=video_id)
1089             else:
1090                 self.to_screen(u'Downloading playlist %s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1091
1092         if playlist_id.startswith('RD'):
1093             # Mixes require a custom extraction process
1094             return self._extract_mix(playlist_id)
1095         if playlist_id.startswith('TL'):
1096             raise ExtractorError(u'For downloading YouTube.com top lists, use '
1097                 u'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True)
1098
1099         url = self._TEMPLATE_URL % playlist_id
1100         page = self._download_webpage(url, playlist_id)
1101         more_widget_html = content_html = page
1102
1103         # Check if the playlist exists or is private
1104         if re.search(r'<div class="yt-alert-message">[^<]*?(The|This) playlist (does not exist|is private)[^<]*?</div>', page) is not None:
1105             raise ExtractorError(
1106                 u'The playlist doesn\'t exist or is private, use --username or '
1107                 '--netrc to access it.',
1108                 expected=True)
1109
1110         # Extract the video ids from the playlist pages
1111         ids = []
1112
1113         for page_num in itertools.count(1):
1114             matches = re.finditer(self._VIDEO_RE, content_html)
1115             # We remove the duplicates and the link with index 0
1116             # (it's not the first video of the playlist)
1117             new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
1118             ids.extend(new_ids)
1119
1120             mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
1121             if not mobj:
1122                 break
1123
1124             more = self._download_json(
1125                 'https://youtube.com/%s' % mobj.group('more'), playlist_id,
1126                 'Downloading page #%s' % page_num,
1127                 transform_source=uppercase_escape)
1128             content_html = more['content_html']
1129             more_widget_html = more['load_more_widget_html']
1130
1131         playlist_title = self._html_search_regex(
1132             r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>',
1133             page, u'title')
1134
1135         url_results = self._ids_to_results(ids)
1136         return self.playlist_result(url_results, playlist_id, playlist_title)
1137
1138
1139 class YoutubeTopListIE(YoutubePlaylistIE):
1140     IE_NAME = u'youtube:toplist'
1141     IE_DESC = (u'YouTube.com top lists, "yttoplist:{channel}:{list title}"'
1142         u' (Example: "yttoplist:music:Top Tracks")')
1143     _VALID_URL = r'yttoplist:(?P<chann>.*?):(?P<title>.*?)$'
1144
1145     def _real_extract(self, url):
1146         mobj = re.match(self._VALID_URL, url)
1147         channel = mobj.group('chann')
1148         title = mobj.group('title')
1149         query = compat_urllib_parse.urlencode({'title': title})
1150         playlist_re = 'href="([^"]+?%s.*?)"' % re.escape(query)
1151         channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title)
1152         link = self._html_search_regex(playlist_re, channel_page, u'list')
1153         url = compat_urlparse.urljoin('https://www.youtube.com/', link)
1154
1155         video_re = r'data-index="\d+".*?data-video-id="([0-9A-Za-z_-]{11})"'
1156         ids = []
1157         # sometimes the webpage doesn't contain the videos
1158         # retry until we get them
1159         for i in itertools.count(0):
1160             msg = u'Downloading Youtube mix'
1161             if i > 0:
1162                 msg += ', retry #%d' % i
1163             webpage = self._download_webpage(url, title, msg)
1164             ids = orderedSet(re.findall(video_re, webpage))
1165             if ids:
1166                 break
1167         url_results = self._ids_to_results(ids)
1168         return self.playlist_result(url_results, playlist_title=title)
1169
1170
1171 class YoutubeChannelIE(InfoExtractor):
1172     IE_DESC = u'YouTube.com channels'
1173     _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
1174     _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
1175     _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'
1176     IE_NAME = u'youtube:channel'
1177
1178     def extract_videos_from_page(self, page):
1179         ids_in_page = []
1180         for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
1181             if mobj.group(1) not in ids_in_page:
1182                 ids_in_page.append(mobj.group(1))
1183         return ids_in_page
1184
1185     def _real_extract(self, url):
1186         # Extract channel id
1187         mobj = re.match(self._VALID_URL, url)
1188         if mobj is None:
1189             raise ExtractorError(u'Invalid URL: %s' % url)
1190
1191         # Download channel page
1192         channel_id = mobj.group(1)
1193         video_ids = []
1194         url = 'https://www.youtube.com/channel/%s/videos' % channel_id
1195         channel_page = self._download_webpage(url, channel_id)
1196         autogenerated = re.search(r'''(?x)
1197                 class="[^"]*?(?:
1198                     channel-header-autogenerated-label|
1199                     yt-channel-title-autogenerated
1200                 )[^"]*"''', channel_page) is not None
1201
1202         if autogenerated:
1203             # The videos are contained in a single page
1204             # the ajax pages can't be used, they are empty
1205             video_ids = self.extract_videos_from_page(channel_page)
1206         else:
1207             # Download all channel pages using the json-based channel_ajax query
1208             for pagenum in itertools.count(1):
1209                 url = self._MORE_PAGES_URL % (pagenum, channel_id)
1210                 page = self._download_json(
1211                     url, channel_id, note=u'Downloading page #%s' % pagenum,
1212                     transform_source=uppercase_escape)
1213
1214                 ids_in_page = self.extract_videos_from_page(page['content_html'])
1215                 video_ids.extend(ids_in_page)
1216
1217                 if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
1218                     break
1219
1220         self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
1221
1222         url_entries = [self.url_result(video_id, 'Youtube', video_id=video_id)
1223                        for video_id in video_ids]
1224         return self.playlist_result(url_entries, channel_id)
1225
1226
1227 class YoutubeUserIE(InfoExtractor):
1228     IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
1229     _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
1230     _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s'
1231     _GDATA_PAGE_SIZE = 50
1232     _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
1233     IE_NAME = u'youtube:user'
1234
1235     @classmethod
1236     def suitable(cls, url):
1237         # Don't return True if the url can be extracted with other youtube
1238         # extractor, the regex would is too permissive and it would match.
1239         other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1240         if any(ie.suitable(url) for ie in other_ies): return False
1241         else: return super(YoutubeUserIE, cls).suitable(url)
1242
1243     def _real_extract(self, url):
1244         # Extract username
1245         mobj = re.match(self._VALID_URL, url)
1246         if mobj is None:
1247             raise ExtractorError(u'Invalid URL: %s' % url)
1248
1249         username = mobj.group(1)
1250
1251         # Download video ids using YouTube Data API. Result size per
1252         # query is limited (currently to 50 videos) so we need to query
1253         # page by page until there are no video ids - it means we got
1254         # all of them.
1255
1256         def download_page(pagenum):
1257             start_index = pagenum * self._GDATA_PAGE_SIZE + 1
1258
1259             gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
1260             page = self._download_webpage(
1261                 gdata_url, username,
1262                 u'Downloading video ids from %d to %d' % (
1263                     start_index, start_index + self._GDATA_PAGE_SIZE))
1264
1265             try:
1266                 response = json.loads(page)
1267             except ValueError as err:
1268                 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1269             if 'entry' not in response['feed']:
1270                 return
1271
1272             # Extract video identifiers
1273             entries = response['feed']['entry']
1274             for entry in entries:
1275                 title = entry['title']['$t']
1276                 video_id = entry['id']['$t'].split('/')[-1]
1277                 yield {
1278                     '_type': 'url',
1279                     'url': video_id,
1280                     'ie_key': 'Youtube',
1281                     'id': video_id,
1282                     'title': title,
1283                 }
1284         url_results = PagedList(download_page, self._GDATA_PAGE_SIZE)
1285
1286         return self.playlist_result(url_results, playlist_title=username)
1287
1288
1289 class YoutubeSearchIE(SearchInfoExtractor):
1290     IE_DESC = u'YouTube.com searches'
1291     _API_URL = u'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
1292     _MAX_RESULTS = 1000
1293     IE_NAME = u'youtube:search'
1294     _SEARCH_KEY = 'ytsearch'
1295
1296     def _get_n_results(self, query, n):
1297         """Get a specified number of results for a query"""
1298
1299         video_ids = []
1300         pagenum = 0
1301         limit = n
1302         PAGE_SIZE = 50
1303
1304         while (PAGE_SIZE * pagenum) < limit:
1305             result_url = self._API_URL % (
1306                 compat_urllib_parse.quote_plus(query.encode('utf-8')),
1307                 (PAGE_SIZE * pagenum) + 1)
1308             data_json = self._download_webpage(
1309                 result_url, video_id=u'query "%s"' % query,
1310                 note=u'Downloading page %s' % (pagenum + 1),
1311                 errnote=u'Unable to download API page')
1312             data = json.loads(data_json)
1313             api_response = data['data']
1314
1315             if 'items' not in api_response:
1316                 raise ExtractorError(
1317                     u'[youtube] No video results', expected=True)
1318
1319             new_ids = list(video['id'] for video in api_response['items'])
1320             video_ids += new_ids
1321
1322             limit = min(n, api_response['totalItems'])
1323             pagenum += 1
1324
1325         if len(video_ids) > n:
1326             video_ids = video_ids[:n]
1327         videos = [self.url_result(video_id, 'Youtube', video_id=video_id)
1328                   for video_id in video_ids]
1329         return self.playlist_result(videos, query)
1330
1331
1332 class YoutubeSearchDateIE(YoutubeSearchIE):
1333     IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
1334     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published'
1335     _SEARCH_KEY = 'ytsearchdate'
1336     IE_DESC = u'YouTube.com searches, newest videos first'
1337
1338
1339 class YoutubeSearchURLIE(InfoExtractor):
1340     IE_DESC = u'YouTube.com search URLs'
1341     IE_NAME = u'youtube:search_url'
1342     _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?search_query=(?P<query>[^&]+)(?:[&]|$)'
1343
1344     def _real_extract(self, url):
1345         mobj = re.match(self._VALID_URL, url)
1346         query = compat_urllib_parse.unquote_plus(mobj.group('query'))
1347
1348         webpage = self._download_webpage(url, query)
1349         result_code = self._search_regex(
1350             r'(?s)<ol class="item-section"(.*?)</ol>', webpage, u'result HTML')
1351
1352         part_codes = re.findall(
1353             r'(?s)<h3 class="yt-lockup-title">(.*?)</h3>', result_code)
1354         entries = []
1355         for part_code in part_codes:
1356             part_title = self._html_search_regex(
1357                 [r'(?s)title="([^"]+)"', r'>([^<]+)</a>'], part_code, 'item title', fatal=False)
1358             part_url_snippet = self._html_search_regex(
1359                 r'(?s)href="([^"]+)"', part_code, 'item URL')
1360             part_url = compat_urlparse.urljoin(
1361                 'https://www.youtube.com/', part_url_snippet)
1362             entries.append({
1363                 '_type': 'url',
1364                 'url': part_url,
1365                 'title': part_title,
1366             })
1367
1368         return {
1369             '_type': 'playlist',
1370             'entries': entries,
1371             'title': query,
1372         }
1373
1374
1375 class YoutubeShowIE(InfoExtractor):
1376     IE_DESC = u'YouTube.com (multi-season) shows'
1377     _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
1378     IE_NAME = u'youtube:show'
1379
1380     def _real_extract(self, url):
1381         mobj = re.match(self._VALID_URL, url)
1382         show_name = mobj.group(1)
1383         webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
1384         # There's one playlist for each season of the show
1385         m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1386         self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
1387         return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
1388
1389
1390 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1391     """
1392     Base class for extractors that fetch info from
1393     http://www.youtube.com/feed_ajax
1394     Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1395     """
1396     _LOGIN_REQUIRED = True
1397     # use action_load_personal_feed instead of action_load_system_feed
1398     _PERSONAL_FEED = False
1399
1400     @property
1401     def _FEED_TEMPLATE(self):
1402         action = 'action_load_system_feed'
1403         if self._PERSONAL_FEED:
1404             action = 'action_load_personal_feed'
1405         return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
1406
1407     @property
1408     def IE_NAME(self):
1409         return u'youtube:%s' % self._FEED_NAME
1410
1411     def _real_initialize(self):
1412         self._login()
1413
1414     def _real_extract(self, url):
1415         feed_entries = []
1416         paging = 0
1417         for i in itertools.count(1):
1418             info = self._download_json(self._FEED_TEMPLATE % paging,
1419                                           u'%s feed' % self._FEED_NAME,
1420                                           u'Downloading page %s' % i)
1421             feed_html = info.get('feed_html') or info.get('content_html')
1422             m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
1423             ids = orderedSet(m.group(1) for m in m_ids)
1424             feed_entries.extend(
1425                 self.url_result(video_id, 'Youtube', video_id=video_id)
1426                 for video_id in ids)
1427             mobj = re.search(
1428                 r'data-uix-load-more-href="/?[^"]+paging=(?P<paging>\d+)',
1429                 feed_html)
1430             if mobj is None:
1431                 break
1432             paging = mobj.group('paging')
1433         return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
1434
1435 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1436     IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
1437     _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1438     _FEED_NAME = 'subscriptions'
1439     _PLAYLIST_TITLE = u'Youtube Subscriptions'
1440
1441 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1442     IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
1443     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1444     _FEED_NAME = 'recommended'
1445     _PLAYLIST_TITLE = u'Youtube Recommended videos'
1446
1447 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
1448     IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
1449     _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
1450     _FEED_NAME = 'watch_later'
1451     _PLAYLIST_TITLE = u'Youtube Watch Later'
1452     _PERSONAL_FEED = True
1453
1454 class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
1455     IE_DESC = u'Youtube watch history, "ythistory" keyword (requires authentication)'
1456     _VALID_URL = u'https?://www\.youtube\.com/feed/history|:ythistory'
1457     _FEED_NAME = 'history'
1458     _PERSONAL_FEED = True
1459     _PLAYLIST_TITLE = u'Youtube Watch History'
1460
1461 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1462     IE_NAME = u'youtube:favorites'
1463     IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
1464     _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1465     _LOGIN_REQUIRED = True
1466
1467     def _real_extract(self, url):
1468         webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1469         playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
1470         return self.url_result(playlist_id, 'YoutubePlaylist')
1471
1472
1473 class YoutubeTruncatedURLIE(InfoExtractor):
1474     IE_NAME = 'youtube:truncated_url'
1475     IE_DESC = False  # Do not list
1476     _VALID_URL = r'''(?x)
1477         (?:https?://)?[^/]+/watch\?(?:
1478             feature=[a-z_]+|
1479             annotation_id=annotation_[^&]+
1480         )?$|
1481         (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$
1482     '''
1483
1484     _TESTS = [{
1485         'url': 'http://www.youtube.com/watch?annotation_id=annotation_3951667041',
1486         'only_matching': True,
1487     }, {
1488         'url': 'http://www.youtube.com/watch?',
1489         'only_matching': True,
1490     }]
1491
1492     def _real_extract(self, url):
1493         raise ExtractorError(
1494             u'Did you forget to quote the URL? Remember that & is a meta '
1495             u'character in most shells, so you want to put the URL in quotes, '
1496             u'like  youtube-dl '
1497             u'"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
1498             u' or simply  youtube-dl BaW_jenozKc  .',
1499             expected=True)