youtube_dl/extractor/youtube.py

   1 # coding: utf-8
   2
   3 import collections
   4 import errno
   5 import io
   6 import itertools
   7 import json
   8 import os.path
   9 import re
  10 import string
  11 import struct
  12 import traceback
  13 import zlib
  14
  15 from .common import InfoExtractor, SearchInfoExtractor
  16 from .subtitles import SubtitlesInfoExtractor
  17 from ..utils import (
  18     compat_chr,
  19     compat_parse_qs,
  20     compat_urllib_parse,
  21     compat_urllib_request,
  22     compat_urlparse,
  23     compat_str,
  24
  25     clean_html,
  26     get_cachedir,
  27     get_element_by_id,
  28     get_element_by_attribute,
  29     ExtractorError,
  30     int_or_none,
  31     PagedList,
  32     RegexNotFoundError,
  33     unescapeHTML,
  34     unified_strdate,
  35     orderedSet,
  36     write_json_file,
  37     uppercase_escape,
  38 )
  39
  40 class YoutubeBaseInfoExtractor(InfoExtractor):
  41     """Provide base functions for Youtube extractors"""
  42     _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
  43     _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1'
  44     _AGE_URL = 'https://www.youtube.com/verify_age?next_url=/&gl=US&hl=en'
  45     _NETRC_MACHINE = 'youtube'
  46     # If True it will raise an error if no login info is provided
  47     _LOGIN_REQUIRED = False
  48
  49     def _set_language(self):
  50         return bool(self._download_webpage(
  51             self._LANG_URL, None,
  52             note=u'Setting language', errnote='unable to set language',
  53             fatal=False))
  54
  55     def _login(self):
  56         (username, password) = self._get_login_info()
  57         # No authentication to be performed
  58         if username is None:
  59             if self._LOGIN_REQUIRED:
  60                 raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True)
  61             return False
  62
  63         login_page = self._download_webpage(
  64             self._LOGIN_URL, None,
  65             note=u'Downloading login page',
  66             errnote=u'unable to fetch login page', fatal=False)
  67         if login_page is False:
  68             return
  69
  70         galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"',
  71                                   login_page, u'Login GALX parameter')
  72
  73         # Log in
  74         login_form_strs = {
  75                 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
  76                 u'Email': username,
  77                 u'GALX': galx,
  78                 u'Passwd': password,
  79                 u'PersistentCookie': u'yes',
  80                 u'_utf8': u'霱',
  81                 u'bgresponse': u'js_disabled',
  82                 u'checkConnection': u'',
  83                 u'checkedDomains': u'youtube',
  84                 u'dnConn': u'',
  85                 u'pstMsg': u'0',
  86                 u'rmShown': u'1',
  87                 u'secTok': u'',
  88                 u'signIn': u'Sign in',
  89                 u'timeStmp': u'',
  90                 u'service': u'youtube',
  91                 u'uilel': u'3',
  92                 u'hl': u'en_US',
  93         }
  94         # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
  95         # chokes on unicode
  96         login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items())
  97         login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
  98
  99         req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
 100         login_results = self._download_webpage(
 101             req, None,
 102             note=u'Logging in', errnote=u'unable to log in', fatal=False)
 103         if login_results is False:
 104             return False
 105         if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
 106             self._downloader.report_warning(u'unable to log in: bad username or password')
 107             return False
 108         return True
 109
 110     def _confirm_age(self):
 111         age_form = {
 112             'next_url': '/',
 113             'action_confirm': 'Confirm',
 114         }
 115         req = compat_urllib_request.Request(self._AGE_URL,
 116             compat_urllib_parse.urlencode(age_form).encode('ascii'))
 117
 118         self._download_webpage(
 119             req, None,
 120             note=u'Confirming age', errnote=u'Unable to confirm age')
 121         return True
 122
 123     def _real_initialize(self):
 124         if self._downloader is None:
 125             return
 126         if not self._set_language():
 127             return
 128         if not self._login():
 129             return
 130         self._confirm_age()
 131
 132
 133 class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
 134     IE_DESC = u'YouTube.com'
 135     _VALID_URL = r"""(?x)^
 136                      (
 137                          (?:https?://|//)?                                    # http(s):// or protocol-independent URL (optional)
 138                          (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/|
 139                             (?:www\.)?deturl\.com/www\.youtube\.com/|
 140                             (?:www\.)?pwnyoutube\.com/|
 141                             (?:www\.)?yourepeat\.com/|
 142                             tube\.majestyc\.net/|
 143                             youtube\.googleapis\.com/)                        # the various hostnames, with wildcard subdomains
 144                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
 145                          (?:                                                  # the various things that can precede the ID:
 146                              (?:(?:v|embed|e)/)                               # v/ or embed/ or e/
 147                              |(?:                                             # or the v= param in all its forms
 148                                  (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)?  # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
 149                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
 150                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
 151                                  v=
 152                              )
 153                          ))
 154                          |youtu\.be/                                          # just youtu.be/xxxx
 155                          )
 156                      )?                                                       # all until now is optional -> you can pass the naked ID
 157                      ([0-9A-Za-z_-]{11})                                      # here is it! the YouTube video ID
 158                      (?(1).+)?                                                # if we found the ID, everything can follow
 159                      $"""
 160     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
 161     _formats = {
 162         '5': {'ext': 'flv', 'width': 400, 'height': 240},
 163         '6': {'ext': 'flv', 'width': 450, 'height': 270},
 164         '13': {'ext': '3gp'},
 165         '17': {'ext': '3gp', 'width': 176, 'height': 144},
 166         '18': {'ext': 'mp4', 'width': 640, 'height': 360},
 167         '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
 168         '34': {'ext': 'flv', 'width': 640, 'height': 360},
 169         '35': {'ext': 'flv', 'width': 854, 'height': 480},
 170         '36': {'ext': '3gp', 'width': 320, 'height': 240},
 171         '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
 172         '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
 173         '43': {'ext': 'webm', 'width': 640, 'height': 360},
 174         '44': {'ext': 'webm', 'width': 854, 'height': 480},
 175         '45': {'ext': 'webm', 'width': 1280, 'height': 720},
 176         '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
 177
 178
 179         # 3d videos
 180         '82': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': '3D', 'preference': -20},
 181         '83': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': '3D', 'preference': -20},
 182         '84': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': '3D', 'preference': -20},
 183         '85': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': '3D', 'preference': -20},
 184         '100': {'ext': 'webm', 'height': 360, 'resolution': '360p', 'format_note': '3D', 'preference': -20},
 185         '101': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': '3D', 'preference': -20},
 186         '102': {'ext': 'webm', 'height': 720, 'resolution': '720p', 'format_note': '3D', 'preference': -20},
 187
 188         # Apple HTTP Live Streaming
 189         '92': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'HLS', 'preference': -10},
 190         '93': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': 'HLS', 'preference': -10},
 191         '94': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': 'HLS', 'preference': -10},
 192         '95': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': 'HLS', 'preference': -10},
 193         '96': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'HLS', 'preference': -10},
 194         '132': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'HLS', 'preference': -10},
 195         '151': {'ext': 'mp4', 'height': 72, 'resolution': '72p', 'format_note': 'HLS', 'preference': -10},
 196
 197         # DASH mp4 video
 198         '133': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'DASH video', 'preference': -40},
 199         '134': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': 'DASH video', 'preference': -40},
 200         '135': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': 'DASH video', 'preference': -40},
 201         '136': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': 'DASH video', 'preference': -40},
 202         '137': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH video', 'preference': -40},
 203         '138': {'ext': 'mp4', 'height': 1081, 'resolution': '>1080p', 'format_note': 'DASH video', 'preference': -40},
 204         '160': {'ext': 'mp4', 'height': 192, 'resolution': '192p', 'format_note': 'DASH video', 'preference': -40},
 205         '264': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH video', 'preference': -40},
 206
 207         # Dash mp4 audio
 208         '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50},
 209         '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50},
 210         '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50},
 211
 212         # Dash webm
 213         '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 214         '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 215         '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 216         '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 217         '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 218         '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 219         '242': {'ext': 'webm', 'height': 240, 'resolution': '240p', 'format_note': 'DASH webm', 'preference': -40},
 220         '243': {'ext': 'webm', 'height': 360, 'resolution': '360p', 'format_note': 'DASH webm', 'preference': -40},
 221         '244': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40},
 222         '245': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40},
 223         '246': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40},
 224         '247': {'ext': 'webm', 'height': 720, 'resolution': '720p', 'format_note': 'DASH webm', 'preference': -40},
 225         '248': {'ext': 'webm', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH webm', 'preference': -40},
 226
 227         # Dash webm audio
 228         '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm audio', 'abr': 48, 'preference': -50},
 229         '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm 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             }
 247         },
 248         {
 249             u"url":  u"http://www.youtube.com/watch?v=UxxajLWwzqY",
 250             u"file":  u"UxxajLWwzqY.mp4",
 251             u"note": u"Test generic use_cipher_signature video (#897)",
 252             u"info_dict": {
 253                 u"upload_date": u"20120506",
 254                 u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
 255                 u"description": u"md5:5b292926389560516e384ac437c0ec07",
 256                 u"uploader": u"Icona Pop",
 257                 u"uploader_id": u"IconaPop"
 258             }
 259         },
 260         {
 261             u"url":  u"https://www.youtube.com/watch?v=07FYdnEawAQ",
 262             u"file":  u"07FYdnEawAQ.mp4",
 263             u"note": u"Test VEVO video with age protection (#956)",
 264             u"info_dict": {
 265                 u"upload_date": u"20130703",
 266                 u"title": u"Justin Timberlake - Tunnel Vision (Explicit)",
 267                 u"description": u"md5:64249768eec3bc4276236606ea996373",
 268                 u"uploader": u"justintimberlakeVEVO",
 269                 u"uploader_id": u"justintimberlakeVEVO"
 270             }
 271         },
 272         {
 273             u"url":  u"//www.YouTube.com/watch?v=yZIXLfi8CZQ",
 274             u"file":  u"yZIXLfi8CZQ.mp4",
 275             u"note": u"Embed-only video (#1746)",
 276             u"info_dict": {
 277                 u"upload_date": u"20120608",
 278                 u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012",
 279                 u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7",
 280                 u"uploader": u"SET India",
 281                 u"uploader_id": u"setindia"
 282             }
 283         },
 284         {
 285             u"url": u"http://www.youtube.com/watch?v=a9LDPn-MO4I",
 286             u"file": u"a9LDPn-MO4I.m4a",
 287             u"note": u"256k DASH audio (format 141) via DASH manifest",
 288             u"info_dict": {
 289                 u"upload_date": "20121002",
 290                 u"uploader_id": "8KVIDEO",
 291                 u"description": "No description available.",
 292                 u"uploader": "8KVIDEO",
 293                 u"title": "UHDTV TEST 8K VIDEO.mp4"
 294             },
 295             u"params": {
 296                 u"youtube_include_dash_manifest": True,
 297                 u"format": "141",
 298             },
 299         },
 300     ]
 301
 302
 303     @classmethod
 304     def suitable(cls, url):
 305         """Receives a URL and returns True if suitable for this IE."""
 306         if YoutubePlaylistIE.suitable(url): return False
 307         return re.match(cls._VALID_URL, url) is not None
 308
 309     def __init__(self, *args, **kwargs):
 310         super(YoutubeIE, self).__init__(*args, **kwargs)
 311         self._player_cache = {}
 312
 313     def report_video_info_webpage_download(self, video_id):
 314         """Report attempt to download video info webpage."""
 315         self.to_screen(u'%s: Downloading video info webpage' % video_id)
 316
 317     def report_information_extraction(self, video_id):
 318         """Report attempt to extract video information."""
 319         self.to_screen(u'%s: Extracting video information' % video_id)
 320
 321     def report_unavailable_format(self, video_id, format):
 322         """Report extracted video URL."""
 323         self.to_screen(u'%s: Format %s not available' % (video_id, format))
 324
 325     def report_rtmp_download(self):
 326         """Indicate the download will use the RTMP protocol."""
 327         self.to_screen(u'RTMP download detected')
 328
 329     def _extract_signature_function(self, video_id, player_url, slen):
 330         id_m = re.match(r'.*-(?P<id>[a-zA-Z0-9_-]+)\.(?P<ext>[a-z]+)$',
 331                         player_url)
 332         player_type = id_m.group('ext')
 333         player_id = id_m.group('id')
 334
 335         # Read from filesystem cache
 336         func_id = '%s_%s_%d' % (player_type, player_id, slen)
 337         assert os.path.basename(func_id) == func_id
 338         cache_dir = get_cachedir(self._downloader.params)
 339
 340         cache_enabled = cache_dir is not None
 341         if cache_enabled:
 342             cache_fn = os.path.join(os.path.expanduser(cache_dir),
 343                                     u'youtube-sigfuncs',
 344                                     func_id + '.json')
 345             try:
 346                 with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
 347                     cache_spec = json.load(cachef)
 348                 return lambda s: u''.join(s[i] for i in cache_spec)
 349             except IOError:
 350                 pass  # No cache available
 351
 352         if player_type == 'js':
 353             code = self._download_webpage(
 354                 player_url, video_id,
 355                 note=u'Downloading %s player %s' % (player_type, player_id),
 356                 errnote=u'Download of %s failed' % player_url)
 357             res = self._parse_sig_js(code)
 358         elif player_type == 'swf':
 359             urlh = self._request_webpage(
 360                 player_url, video_id,
 361                 note=u'Downloading %s player %s' % (player_type, player_id),
 362                 errnote=u'Download of %s failed' % player_url)
 363             code = urlh.read()
 364             res = self._parse_sig_swf(code)
 365         else:
 366             assert False, 'Invalid player type %r' % player_type
 367
 368         if cache_enabled:
 369             try:
 370                 test_string = u''.join(map(compat_chr, range(slen)))
 371                 cache_res = res(test_string)
 372                 cache_spec = [ord(c) for c in cache_res]
 373                 try:
 374                     os.makedirs(os.path.dirname(cache_fn))
 375                 except OSError as ose:
 376                     if ose.errno != errno.EEXIST:
 377                         raise
 378                 write_json_file(cache_spec, cache_fn)
 379             except Exception:
 380                 tb = traceback.format_exc()
 381                 self._downloader.report_warning(
 382                     u'Writing cache to %r failed: %s' % (cache_fn, tb))
 383
 384         return res
 385
 386     def _print_sig_code(self, func, slen):
 387         def gen_sig_code(idxs):
 388             def _genslice(start, end, step):
 389                 starts = u'' if start == 0 else str(start)
 390                 ends = (u':%d' % (end+step)) if end + step >= 0 else u':'
 391                 steps = u'' if step == 1 else (u':%d' % step)
 392                 return u's[%s%s%s]' % (starts, ends, steps)
 393
 394             step = None
 395             start = '(Never used)'  # Quelch pyflakes warnings - start will be
 396                                     # set as soon as step is set
 397             for i, prev in zip(idxs[1:], idxs[:-1]):
 398                 if step is not None:
 399                     if i - prev == step:
 400                         continue
 401                     yield _genslice(start, prev, step)
 402                     step = None
 403                     continue
 404                 if i - prev in [-1, 1]:
 405                     step = i - prev
 406                     start = prev
 407                     continue
 408                 else:
 409                     yield u's[%d]' % prev
 410             if step is None:
 411                 yield u's[%d]' % i
 412             else:
 413                 yield _genslice(start, i, step)
 414
 415         test_string = u''.join(map(compat_chr, range(slen)))
 416         cache_res = func(test_string)
 417         cache_spec = [ord(c) for c in cache_res]
 418         expr_code = u' + '.join(gen_sig_code(cache_spec))
 419         code = u'if len(s) == %d:\n    return %s\n' % (slen, expr_code)
 420         self.to_screen(u'Extracted signature function:\n' + code)
 421
 422     def _parse_sig_js(self, jscode):
 423         funcname = self._search_regex(
 424             r'signature=([a-zA-Z]+)', jscode,
 425             u'Initial JS player signature function name')
 426
 427         functions = {}
 428
 429         def argidx(varname):
 430             return string.lowercase.index(varname)
 431
 432         def interpret_statement(stmt, local_vars, allow_recursion=20):
 433             if allow_recursion < 0:
 434                 raise ExtractorError(u'Recursion limit reached')
 435
 436             if stmt.startswith(u'var '):
 437                 stmt = stmt[len(u'var '):]
 438             ass_m = re.match(r'^(?P<out>[a-z]+)(?:\[(?P<index>[^\]]+)\])?' +
 439                              r'=(?P<expr>.*)$', stmt)
 440             if ass_m:
 441                 if ass_m.groupdict().get('index'):
 442                     def assign(val):
 443                         lvar = local_vars[ass_m.group('out')]
 444                         idx = interpret_expression(ass_m.group('index'),
 445                                                    local_vars, allow_recursion)
 446                         assert isinstance(idx, int)
 447                         lvar[idx] = val
 448                         return val
 449                     expr = ass_m.group('expr')
 450                 else:
 451                     def assign(val):
 452                         local_vars[ass_m.group('out')] = val
 453                         return val
 454                     expr = ass_m.group('expr')
 455             elif stmt.startswith(u'return '):
 456                 assign = lambda v: v
 457                 expr = stmt[len(u'return '):]
 458             else:
 459                 raise ExtractorError(
 460                     u'Cannot determine left side of statement in %r' % stmt)
 461
 462             v = interpret_expression(expr, local_vars, allow_recursion)
 463             return assign(v)
 464
 465         def interpret_expression(expr, local_vars, allow_recursion):
 466             if expr.isdigit():
 467                 return int(expr)
 468
 469             if expr.isalpha():
 470                 return local_vars[expr]
 471
 472             m = re.match(r'^(?P<in>[a-z]+)\.(?P<member>.*)$', expr)
 473             if m:
 474                 member = m.group('member')
 475                 val = local_vars[m.group('in')]
 476                 if member == 'split("")':
 477                     return list(val)
 478                 if member == 'join("")':
 479                     return u''.join(val)
 480                 if member == 'length':
 481                     return len(val)
 482                 if member == 'reverse()':
 483                     return val[::-1]
 484                 slice_m = re.match(r'slice\((?P<idx>.*)\)', member)
 485                 if slice_m:
 486                     idx = interpret_expression(
 487                         slice_m.group('idx'), local_vars, allow_recursion-1)
 488                     return val[idx:]
 489
 490             m = re.match(
 491                 r'^(?P<in>[a-z]+)\[(?P<idx>.+)\]$', expr)
 492             if m:
 493                 val = local_vars[m.group('in')]
 494                 idx = interpret_expression(m.group('idx'), local_vars,
 495                                            allow_recursion-1)
 496                 return val[idx]
 497
 498             m = re.match(r'^(?P<a>.+?)(?P<op>[%])(?P<b>.+?)$', expr)
 499             if m:
 500                 a = interpret_expression(m.group('a'),
 501                                          local_vars, allow_recursion)
 502                 b = interpret_expression(m.group('b'),
 503                                          local_vars, allow_recursion)
 504                 return a % b
 505
 506             m = re.match(
 507                 r'^(?P<func>[a-zA-Z$]+)\((?P<args>[a-z0-9,]+)\)$', expr)
 508             if m:
 509                 fname = m.group('func')
 510                 if fname not in functions:
 511                     functions[fname] = extract_function(fname)
 512                 argvals = [int(v) if v.isdigit() else local_vars[v]
 513                            for v in m.group('args').split(',')]
 514                 return functions[fname](argvals)
 515             raise ExtractorError(u'Unsupported JS expression %r' % expr)
 516
 517         def extract_function(funcname):
 518             func_m = re.search(
 519                 r'function ' + re.escape(funcname) +
 520                 r'\((?P<args>[a-z,]+)\){(?P<code>[^}]+)}',
 521                 jscode)
 522             argnames = func_m.group('args').split(',')
 523
 524             def resf(args):
 525                 local_vars = dict(zip(argnames, args))
 526                 for stmt in func_m.group('code').split(';'):
 527                     res = interpret_statement(stmt, local_vars)
 528                 return res
 529             return resf
 530
 531         initial_function = extract_function(funcname)
 532         return lambda s: initial_function([s])
 533
 534     def _parse_sig_swf(self, file_contents):
 535         if file_contents[1:3] != b'WS':
 536             raise ExtractorError(
 537                 u'Not an SWF file; header is %r' % file_contents[:3])
 538         if file_contents[:1] == b'C':
 539             content = zlib.decompress(file_contents[8:])
 540         else:
 541             raise NotImplementedError(u'Unsupported compression format %r' %
 542                                       file_contents[:1])
 543
 544         def extract_tags(content):
 545             pos = 0
 546             while pos < len(content):
 547                 header16 = struct.unpack('<H', content[pos:pos+2])[0]
 548                 pos += 2
 549                 tag_code = header16 >> 6
 550                 tag_len = header16 & 0x3f
 551                 if tag_len == 0x3f:
 552                     tag_len = struct.unpack('<I', content[pos:pos+4])[0]
 553                     pos += 4
 554                 assert pos+tag_len <= len(content)
 555                 yield (tag_code, content[pos:pos+tag_len])
 556                 pos += tag_len
 557
 558         code_tag = next(tag
 559                         for tag_code, tag in extract_tags(content)
 560                         if tag_code == 82)
 561         p = code_tag.index(b'\0', 4) + 1
 562         code_reader = io.BytesIO(code_tag[p:])
 563
 564         # Parse ABC (AVM2 ByteCode)
 565         def read_int(reader=None):
 566             if reader is None:
 567                 reader = code_reader
 568             res = 0
 569             shift = 0
 570             for _ in range(5):
 571                 buf = reader.read(1)
 572                 assert len(buf) == 1
 573                 b = struct.unpack('<B', buf)[0]
 574                 res = res | ((b & 0x7f) << shift)
 575                 if b & 0x80 == 0:
 576                     break
 577                 shift += 7
 578             return res
 579
 580         def u30(reader=None):
 581             res = read_int(reader)
 582             assert res & 0xf0000000 == 0
 583             return res
 584         u32 = read_int
 585
 586         def s32(reader=None):
 587             v = read_int(reader)
 588             if v & 0x80000000 != 0:
 589                 v = - ((v ^ 0xffffffff) + 1)
 590             return v
 591
 592         def read_string(reader=None):
 593             if reader is None:
 594                 reader = code_reader
 595             slen = u30(reader)
 596             resb = reader.read(slen)
 597             assert len(resb) == slen
 598             return resb.decode('utf-8')
 599
 600         def read_bytes(count, reader=None):
 601             if reader is None:
 602                 reader = code_reader
 603             resb = reader.read(count)
 604             assert len(resb) == count
 605             return resb
 606
 607         def read_byte(reader=None):
 608             resb = read_bytes(1, reader=reader)
 609             res = struct.unpack('<B', resb)[0]
 610             return res
 611
 612         # minor_version + major_version
 613         read_bytes(2 + 2)
 614
 615         # Constant pool
 616         int_count = u30()
 617         for _c in range(1, int_count):
 618             s32()
 619         uint_count = u30()
 620         for _c in range(1, uint_count):
 621             u32()
 622         double_count = u30()
 623         read_bytes((double_count-1) * 8)
 624         string_count = u30()
 625         constant_strings = [u'']
 626         for _c in range(1, string_count):
 627             s = read_string()
 628             constant_strings.append(s)
 629         namespace_count = u30()
 630         for _c in range(1, namespace_count):
 631             read_bytes(1)  # kind
 632             u30()  # name
 633         ns_set_count = u30()
 634         for _c in range(1, ns_set_count):
 635             count = u30()
 636             for _c2 in range(count):
 637                 u30()
 638         multiname_count = u30()
 639         MULTINAME_SIZES = {
 640             0x07: 2,  # QName
 641             0x0d: 2,  # QNameA
 642             0x0f: 1,  # RTQName
 643             0x10: 1,  # RTQNameA
 644             0x11: 0,  # RTQNameL
 645             0x12: 0,  # RTQNameLA
 646             0x09: 2,  # Multiname
 647             0x0e: 2,  # MultinameA
 648             0x1b: 1,  # MultinameL
 649             0x1c: 1,  # MultinameLA
 650         }
 651         multinames = [u'']
 652         for _c in range(1, multiname_count):
 653             kind = u30()
 654             assert kind in MULTINAME_SIZES, u'Invalid multiname kind %r' % kind
 655             if kind == 0x07:
 656                 u30()  # namespace_idx
 657                 name_idx = u30()
 658                 multinames.append(constant_strings[name_idx])
 659             else:
 660                 multinames.append('[MULTINAME kind: %d]' % kind)
 661                 for _c2 in range(MULTINAME_SIZES[kind]):
 662                     u30()
 663
 664         # Methods
 665         method_count = u30()
 666         MethodInfo = collections.namedtuple(
 667             'MethodInfo',
 668             ['NEED_ARGUMENTS', 'NEED_REST'])
 669         method_infos = []
 670         for method_id in range(method_count):
 671             param_count = u30()
 672             u30()  # return type
 673             for _ in range(param_count):
 674                 u30()  # param type
 675             u30()  # name index (always 0 for youtube)
 676             flags = read_byte()
 677             if flags & 0x08 != 0:
 678                 # Options present
 679                 option_count = u30()
 680                 for c in range(option_count):
 681                     u30()  # val
 682                     read_bytes(1)  # kind
 683             if flags & 0x80 != 0:
 684                 # Param names present
 685                 for _ in range(param_count):
 686                     u30()  # param name
 687             mi = MethodInfo(flags & 0x01 != 0, flags & 0x04 != 0)
 688             method_infos.append(mi)
 689
 690         # Metadata
 691         metadata_count = u30()
 692         for _c in range(metadata_count):
 693             u30()  # name
 694             item_count = u30()
 695             for _c2 in range(item_count):
 696                 u30()  # key
 697                 u30()  # value
 698
 699         def parse_traits_info():
 700             trait_name_idx = u30()
 701             kind_full = read_byte()
 702             kind = kind_full & 0x0f
 703             attrs = kind_full >> 4
 704             methods = {}
 705             if kind in [0x00, 0x06]:  # Slot or Const
 706                 u30()  # Slot id
 707                 u30()  # type_name_idx
 708                 vindex = u30()
 709                 if vindex != 0:
 710                     read_byte()  # vkind
 711             elif kind in [0x01, 0x02, 0x03]:  # Method / Getter / Setter
 712                 u30()  # disp_id
 713                 method_idx = u30()
 714                 methods[multinames[trait_name_idx]] = method_idx
 715             elif kind == 0x04:  # Class
 716                 u30()  # slot_id
 717                 u30()  # classi
 718             elif kind == 0x05:  # Function
 719                 u30()  # slot_id
 720                 function_idx = u30()
 721                 methods[function_idx] = multinames[trait_name_idx]
 722             else:
 723                 raise ExtractorError(u'Unsupported trait kind %d' % kind)
 724
 725             if attrs & 0x4 != 0:  # Metadata present
 726                 metadata_count = u30()
 727                 for _c3 in range(metadata_count):
 728                     u30()  # metadata index
 729
 730             return methods
 731
 732         # Classes
 733         TARGET_CLASSNAME = u'SignatureDecipher'
 734         searched_idx = multinames.index(TARGET_CLASSNAME)
 735         searched_class_id = None
 736         class_count = u30()
 737         for class_id in range(class_count):
 738             name_idx = u30()
 739             if name_idx == searched_idx:
 740                 # We found the class we're looking for!
 741                 searched_class_id = class_id
 742             u30()  # super_name idx
 743             flags = read_byte()
 744             if flags & 0x08 != 0:  # Protected namespace is present
 745                 u30()  # protected_ns_idx
 746             intrf_count = u30()
 747             for _c2 in range(intrf_count):
 748                 u30()
 749             u30()  # iinit
 750             trait_count = u30()
 751             for _c2 in range(trait_count):
 752                 parse_traits_info()
 753
 754         if searched_class_id is None:
 755             raise ExtractorError(u'Target class %r not found' %
 756                                  TARGET_CLASSNAME)
 757
 758         method_names = {}
 759         method_idxs = {}
 760         for class_id in range(class_count):
 761             u30()  # cinit
 762             trait_count = u30()
 763             for _c2 in range(trait_count):
 764                 trait_methods = parse_traits_info()
 765                 if class_id == searched_class_id:
 766                     method_names.update(trait_methods.items())
 767                     method_idxs.update(dict(
 768                         (idx, name)
 769                         for name, idx in trait_methods.items()))
 770
 771         # Scripts
 772         script_count = u30()
 773         for _c in range(script_count):
 774             u30()  # init
 775             trait_count = u30()
 776             for _c2 in range(trait_count):
 777                 parse_traits_info()
 778
 779         # Method bodies
 780         method_body_count = u30()
 781         Method = collections.namedtuple('Method', ['code', 'local_count'])
 782         methods = {}
 783         for _c in range(method_body_count):
 784             method_idx = u30()
 785             u30()  # max_stack
 786             local_count = u30()
 787             u30()  # init_scope_depth
 788             u30()  # max_scope_depth
 789             code_length = u30()
 790             code = read_bytes(code_length)
 791             if method_idx in method_idxs:
 792                 m = Method(code, local_count)
 793                 methods[method_idxs[method_idx]] = m
 794             exception_count = u30()
 795             for _c2 in range(exception_count):
 796                 u30()  # from
 797                 u30()  # to
 798                 u30()  # target
 799                 u30()  # exc_type
 800                 u30()  # var_name
 801             trait_count = u30()
 802             for _c2 in range(trait_count):
 803                 parse_traits_info()
 804
 805         assert p + code_reader.tell() == len(code_tag)
 806         assert len(methods) == len(method_idxs)
 807
 808         method_pyfunctions = {}
 809
 810         def extract_function(func_name):
 811             if func_name in method_pyfunctions:
 812                 return method_pyfunctions[func_name]
 813             if func_name not in methods:
 814                 raise ExtractorError(u'Cannot find function %r' % func_name)
 815             m = methods[func_name]
 816
 817             def resfunc(args):
 818                 registers = ['(this)'] + list(args) + [None] * m.local_count
 819                 stack = []
 820                 coder = io.BytesIO(m.code)
 821                 while True:
 822                     opcode = struct.unpack('!B', coder.read(1))[0]
 823                     if opcode == 36:  # pushbyte
 824                         v = struct.unpack('!B', coder.read(1))[0]
 825                         stack.append(v)
 826                     elif opcode == 44:  # pushstring
 827                         idx = u30(coder)
 828                         stack.append(constant_strings[idx])
 829                     elif opcode == 48:  # pushscope
 830                         # We don't implement the scope register, so we'll just
 831                         # ignore the popped value
 832                         stack.pop()
 833                     elif opcode == 70:  # callproperty
 834                         index = u30(coder)
 835                         mname = multinames[index]
 836                         arg_count = u30(coder)
 837                         args = list(reversed(
 838                             [stack.pop() for _ in range(arg_count)]))
 839                         obj = stack.pop()
 840                         if mname == u'split':
 841                             assert len(args) == 1
 842                             assert isinstance(args[0], compat_str)
 843                             assert isinstance(obj, compat_str)
 844                             if args[0] == u'':
 845                                 res = list(obj)
 846                             else:
 847                                 res = obj.split(args[0])
 848                             stack.append(res)
 849                         elif mname == u'slice':
 850                             assert len(args) == 1
 851                             assert isinstance(args[0], int)
 852                             assert isinstance(obj, list)
 853                             res = obj[args[0]:]
 854                             stack.append(res)
 855                         elif mname == u'join':
 856                             assert len(args) == 1
 857                             assert isinstance(args[0], compat_str)
 858                             assert isinstance(obj, list)
 859                             res = args[0].join(obj)
 860                             stack.append(res)
 861                         elif mname in method_pyfunctions:
 862                             stack.append(method_pyfunctions[mname](args))
 863                         else:
 864                             raise NotImplementedError(
 865                                 u'Unsupported property %r on %r'
 866                                 % (mname, obj))
 867                     elif opcode == 72:  # returnvalue
 868                         res = stack.pop()
 869                         return res
 870                     elif opcode == 79:  # callpropvoid
 871                         index = u30(coder)
 872                         mname = multinames[index]
 873                         arg_count = u30(coder)
 874                         args = list(reversed(
 875                             [stack.pop() for _ in range(arg_count)]))
 876                         obj = stack.pop()
 877                         if mname == u'reverse':
 878                             assert isinstance(obj, list)
 879                             obj.reverse()
 880                         else:
 881                             raise NotImplementedError(
 882                                 u'Unsupported (void) property %r on %r'
 883                                 % (mname, obj))
 884                     elif opcode == 93:  # findpropstrict
 885                         index = u30(coder)
 886                         mname = multinames[index]
 887                         res = extract_function(mname)
 888                         stack.append(res)
 889                     elif opcode == 97:  # setproperty
 890                         index = u30(coder)
 891                         value = stack.pop()
 892                         idx = stack.pop()
 893                         obj = stack.pop()
 894                         assert isinstance(obj, list)
 895                         assert isinstance(idx, int)
 896                         obj[idx] = value
 897                     elif opcode == 98:  # getlocal
 898                         index = u30(coder)
 899                         stack.append(registers[index])
 900                     elif opcode == 99:  # setlocal
 901                         index = u30(coder)
 902                         value = stack.pop()
 903                         registers[index] = value
 904                     elif opcode == 102:  # getproperty
 905                         index = u30(coder)
 906                         pname = multinames[index]
 907                         if pname == u'length':
 908                             obj = stack.pop()
 909                             assert isinstance(obj, list)
 910                             stack.append(len(obj))
 911                         else:  # Assume attribute access
 912                             idx = stack.pop()
 913                             assert isinstance(idx, int)
 914                             obj = stack.pop()
 915                             assert isinstance(obj, list)
 916                             stack.append(obj[idx])
 917                     elif opcode == 128:  # coerce
 918                         u30(coder)
 919                     elif opcode == 133:  # coerce_s
 920                         assert isinstance(stack[-1], (type(None), compat_str))
 921                     elif opcode == 164:  # modulo
 922                         value2 = stack.pop()
 923                         value1 = stack.pop()
 924                         res = value1 % value2
 925                         stack.append(res)
 926                     elif opcode == 208:  # getlocal_0
 927                         stack.append(registers[0])
 928                     elif opcode == 209:  # getlocal_1
 929                         stack.append(registers[1])
 930                     elif opcode == 210:  # getlocal_2
 931                         stack.append(registers[2])
 932                     elif opcode == 211:  # getlocal_3
 933                         stack.append(registers[3])
 934                     elif opcode == 214:  # setlocal_2
 935                         registers[2] = stack.pop()
 936                     elif opcode == 215:  # setlocal_3
 937                         registers[3] = stack.pop()
 938                     else:
 939                         raise NotImplementedError(
 940                             u'Unsupported opcode %d' % opcode)
 941
 942             method_pyfunctions[func_name] = resfunc
 943             return resfunc
 944
 945         initial_function = extract_function(u'decipher')
 946         return lambda s: initial_function([s])
 947
 948     def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
 949         """Turn the encrypted s field into a working signature"""
 950
 951         if player_url is not None:
 952             if player_url.startswith(u'//'):
 953                 player_url = u'https:' + player_url
 954             try:
 955                 player_id = (player_url, len(s))
 956                 if player_id not in self._player_cache:
 957                     func = self._extract_signature_function(
 958                         video_id, player_url, len(s)
 959                     )
 960                     self._player_cache[player_id] = func
 961                 func = self._player_cache[player_id]
 962                 if self._downloader.params.get('youtube_print_sig_code'):
 963                     self._print_sig_code(func, len(s))
 964                 return func(s)
 965             except Exception:
 966                 tb = traceback.format_exc()
 967                 self._downloader.report_warning(
 968                     u'Automatic signature extraction failed: ' + tb)
 969
 970             self._downloader.report_warning(
 971                 u'Warning: Falling back to static signature algorithm')
 972
 973         return self._static_decrypt_signature(
 974             s, video_id, player_url, age_gate)
 975
 976     def _static_decrypt_signature(self, s, video_id, player_url, age_gate):
 977         if age_gate:
 978             # The videos with age protection use another player, so the
 979             # algorithms can be different.
 980             if len(s) == 86:
 981                 return s[2:63] + s[82] + s[64:82] + s[63]
 982
 983         if len(s) == 93:
 984             return s[86:29:-1] + s[88] + s[28:5:-1]
 985         elif len(s) == 92:
 986             return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83]
 987         elif len(s) == 91:
 988             return s[84:27:-1] + s[86] + s[26:5:-1]
 989         elif len(s) == 90:
 990             return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81]
 991         elif len(s) == 89:
 992             return s[84:78:-1] + s[87] + s[77:60:-1] + s[0] + s[59:3:-1]
 993         elif len(s) == 88:
 994             return s[7:28] + s[87] + s[29:45] + s[55] + s[46:55] + s[2] + s[56:87] + s[28]
 995         elif len(s) == 87:
 996             return s[6:27] + s[4] + s[28:39] + s[27] + s[40:59] + s[2] + s[60:]
 997         elif len(s) == 86:
 998             return s[80:72:-1] + s[16] + s[71:39:-1] + s[72] + s[38:16:-1] + s[82] + s[15::-1]
 999         elif len(s) == 85:
1000             return s[3:11] + s[0] + s[12:55] + s[84] + s[56:84]
1001         elif len(s) == 84:
1002             return s[78:70:-1] + s[14] + s[69:37:-1] + s[70] + s[36:14:-1] + s[80] + s[:14][::-1]
1003         elif len(s) == 83:
1004             return s[80:63:-1] + s[0] + s[62:0:-1] + s[63]
1005         elif len(s) == 82:
1006             return s[80:37:-1] + s[7] + s[36:7:-1] + s[0] + s[6:0:-1] + s[37]
1007         elif len(s) == 81:
1008             return s[56] + s[79:56:-1] + s[41] + s[55:41:-1] + s[80] + s[40:34:-1] + s[0] + s[33:29:-1] + s[34] + s[28:9:-1] + s[29] + s[8:0:-1] + s[9]
1009         elif len(s) == 80:
1010             return s[1:19] + s[0] + s[20:68] + s[19] + s[69:80]
1011         elif len(s) == 79:
1012             return s[54] + s[77:54:-1] + s[39] + s[53:39:-1] + s[78] + s[38:34:-1] + s[0] + s[33:29:-1] + s[34] + s[28:9:-1] + s[29] + s[8:0:-1] + s[9]
1013
1014         else:
1015             raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s)))
1016
1017     def _get_available_subtitles(self, video_id, webpage):
1018         try:
1019             sub_list = self._download_webpage(
1020                 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
1021                 video_id, note=False)
1022         except ExtractorError as err:
1023             self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
1024             return {}
1025         lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
1026
1027         sub_lang_list = {}
1028         for l in lang_list:
1029             lang = l[1]
1030             params = compat_urllib_parse.urlencode({
1031                 'lang': lang,
1032                 'v': video_id,
1033                 'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
1034                 'name': unescapeHTML(l[0]).encode('utf-8'),
1035             })
1036             url = u'https://www.youtube.com/api/timedtext?' + params
1037             sub_lang_list[lang] = url
1038         if not sub_lang_list:
1039             self._downloader.report_warning(u'video doesn\'t have subtitles')
1040             return {}
1041         return sub_lang_list
1042
1043     def _get_available_automatic_caption(self, video_id, webpage):
1044         """We need the webpage for getting the captions url, pass it as an
1045            argument to speed up the process."""
1046         sub_format = self._downloader.params.get('subtitlesformat', 'srt')
1047         self.to_screen(u'%s: Looking for automatic captions' % video_id)
1048         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
1049         err_msg = u'Couldn\'t find automatic captions for %s' % video_id
1050         if mobj is None:
1051             self._downloader.report_warning(err_msg)
1052             return {}
1053         player_config = json.loads(mobj.group(1))
1054         try:
1055             args = player_config[u'args']
1056             caption_url = args[u'ttsurl']
1057             timestamp = args[u'timestamp']
1058             # We get the available subtitles
1059             list_params = compat_urllib_parse.urlencode({
1060                 'type': 'list',
1061                 'tlangs': 1,
1062                 'asrs': 1,
1063             })
1064             list_url = caption_url + '&' + list_params
1065             caption_list = self._download_xml(list_url, video_id)
1066             original_lang_node = caption_list.find('track')
1067             if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' :
1068                 self._downloader.report_warning(u'Video doesn\'t have automatic captions')
1069                 return {}
1070             original_lang = original_lang_node.attrib['lang_code']
1071
1072             sub_lang_list = {}
1073             for lang_node in caption_list.findall('target'):
1074                 sub_lang = lang_node.attrib['lang_code']
1075                 params = compat_urllib_parse.urlencode({
1076                     'lang': original_lang,
1077                     'tlang': sub_lang,
1078                     'fmt': sub_format,
1079                     'ts': timestamp,
1080                     'kind': 'asr',
1081                 })
1082                 sub_lang_list[sub_lang] = caption_url + '&' + params
1083             return sub_lang_list
1084         # An extractor error can be raise by the download process if there are
1085         # no automatic captions but there are subtitles
1086         except (KeyError, ExtractorError):
1087             self._downloader.report_warning(err_msg)
1088             return {}
1089
1090     @classmethod
1091     def extract_id(cls, url):
1092         mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
1093         if mobj is None:
1094             raise ExtractorError(u'Invalid URL: %s' % url)
1095         video_id = mobj.group(2)
1096         return video_id
1097
1098     def _extract_from_m3u8(self, manifest_url, video_id):
1099         url_map = {}
1100         def _get_urls(_manifest):
1101             lines = _manifest.split('\n')
1102             urls = filter(lambda l: l and not l.startswith('#'),
1103                             lines)
1104             return urls
1105         manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest')
1106         formats_urls = _get_urls(manifest)
1107         for format_url in formats_urls:
1108             itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
1109             url_map[itag] = format_url
1110         return url_map
1111
1112     def _extract_annotations(self, video_id):
1113         url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
1114         return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.')
1115
1116     def _real_extract(self, url):
1117         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
1118         mobj = re.search(self._NEXT_URL_RE, url)
1119         if mobj:
1120             url = 'https://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
1121         video_id = self.extract_id(url)
1122
1123         # Get video webpage
1124         url = 'https://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
1125         video_webpage = self._download_webpage(url, video_id)
1126
1127         # Attempt to extract SWF player URL
1128         mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
1129         if mobj is not None:
1130             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
1131         else:
1132             player_url = None
1133
1134         # Get video info
1135         self.report_video_info_webpage_download(video_id)
1136         if re.search(r'player-age-gate-content">', video_webpage) is not None:
1137             self.report_age_confirmation()
1138             age_gate = True
1139             # We simulate the access to the video from www.youtube.com/v/{video_id}
1140             # this can be viewed without login into Youtube
1141             data = compat_urllib_parse.urlencode({'video_id': video_id,
1142                                                   'el': 'player_embedded',
1143                                                   'gl': 'US',
1144                                                   'hl': 'en',
1145                                                   'eurl': 'https://youtube.googleapis.com/v/' + video_id,
1146                                                   'asv': 3,
1147                                                   'sts':'1588',
1148                                                   })
1149             video_info_url = 'https://www.youtube.com/get_video_info?' + data
1150             video_info_webpage = self._download_webpage(video_info_url, video_id,
1151                                     note=False,
1152                                     errnote='unable to download video info webpage')
1153             video_info = compat_parse_qs(video_info_webpage)
1154         else:
1155             age_gate = False
1156             for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
1157                 video_info_url = ('https://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
1158                         % (video_id, el_type))
1159                 video_info_webpage = self._download_webpage(video_info_url, video_id,
1160                                         note=False,
1161                                         errnote='unable to download video info webpage')
1162                 video_info = compat_parse_qs(video_info_webpage)
1163                 if 'token' in video_info:
1164                     break
1165         if 'token' not in video_info:
1166             if 'reason' in video_info:
1167                 raise ExtractorError(u'YouTube said: %s' % video_info['reason'][0], expected=True)
1168             else:
1169                 raise ExtractorError(u'"token" parameter not in video info for unknown reason')
1170
1171         if 'view_count' in video_info:
1172             view_count = int(video_info['view_count'][0])
1173         else:
1174             view_count = None
1175
1176         # Check for "rental" videos
1177         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
1178             raise ExtractorError(u'"rental" videos not supported')
1179
1180         # Start extracting information
1181         self.report_information_extraction(video_id)
1182
1183         # uploader
1184         if 'author' not in video_info:
1185             raise ExtractorError(u'Unable to extract uploader name')
1186         video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
1187
1188         # uploader_id
1189         video_uploader_id = None
1190         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
1191         if mobj is not None:
1192             video_uploader_id = mobj.group(1)
1193         else:
1194             self._downloader.report_warning(u'unable to extract uploader nickname')
1195
1196         # title
1197         if 'title' in video_info:
1198             video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
1199         else:
1200             self._downloader.report_warning(u'Unable to extract video title')
1201             video_title = u'_'
1202
1203         # thumbnail image
1204         # We try first to get a high quality image:
1205         m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
1206                             video_webpage, re.DOTALL)
1207         if m_thumb is not None:
1208             video_thumbnail = m_thumb.group(1)
1209         elif 'thumbnail_url' not in video_info:
1210             self._downloader.report_warning(u'unable to extract video thumbnail')
1211             video_thumbnail = None
1212         else:   # don't panic if we can't find it
1213             video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
1214
1215         # upload date
1216         upload_date = None
1217         mobj = re.search(r'id="eow-date.*?>(.*?)</span>', video_webpage, re.DOTALL)
1218         if mobj is not None:
1219             upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
1220             upload_date = unified_strdate(upload_date)
1221
1222         # description
1223         video_description = get_element_by_id("eow-description", video_webpage)
1224         if video_description:
1225             video_description = re.sub(r'''(?x)
1226                 <a\s+
1227                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1228                     title="([^"]+)"\s+
1229                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1230                     class="yt-uix-redirect-link"\s*>
1231                 [^<]+
1232                 </a>
1233             ''', r'\1', video_description)
1234             video_description = clean_html(video_description)
1235         else:
1236             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
1237             if fd_mobj:
1238                 video_description = unescapeHTML(fd_mobj.group(1))
1239             else:
1240                 video_description = u''
1241
1242         def _extract_count(klass):
1243             count = self._search_regex(
1244                 r'class="%s">([\d,]+)</span>' % re.escape(klass),
1245                 video_webpage, klass, default=None)
1246             if count is not None:
1247                 return int(count.replace(',', ''))
1248             return None
1249         like_count = _extract_count(u'likes-count')
1250         dislike_count = _extract_count(u'dislikes-count')
1251
1252         # subtitles
1253         video_subtitles = self.extract_subtitles(video_id, video_webpage)
1254
1255         if self._downloader.params.get('listsubtitles', False):
1256             self._list_available_subtitles(video_id, video_webpage)
1257             return
1258
1259         if 'length_seconds' not in video_info:
1260             self._downloader.report_warning(u'unable to extract video duration')
1261             video_duration = None
1262         else:
1263             video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]))
1264
1265         # annotations
1266         video_annotations = None
1267         if self._downloader.params.get('writeannotations', False):
1268                 video_annotations = self._extract_annotations(video_id)
1269
1270         # Decide which formats to download
1271         try:
1272             mobj = re.search(r';ytplayer.config = ({.*?});', video_webpage)
1273             if not mobj:
1274                 raise ValueError('Could not find vevo ID')
1275             info = json.loads(mobj.group(1))
1276             args = info['args']
1277             # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
1278             # this signatures are encrypted
1279             if 'url_encoded_fmt_stream_map' not in args:
1280                 raise ValueError(u'No stream_map present')  # caught below
1281             re_signature = re.compile(r'[&,]s=')
1282             m_s = re_signature.search(args['url_encoded_fmt_stream_map'])
1283             if m_s is not None:
1284                 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
1285                 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
1286             m_s = re_signature.search(args.get('adaptive_fmts', u''))
1287             if m_s is not None:
1288                 if 'adaptive_fmts' in video_info:
1289                     video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts']
1290                 else:
1291                     video_info['adaptive_fmts'] = [args['adaptive_fmts']]
1292         except ValueError:
1293             pass
1294
1295         def _map_to_format_list(urlmap):
1296             formats = []
1297             for itag, video_real_url in urlmap.items():
1298                 dct = {
1299                     'format_id': itag,
1300                     'url': video_real_url,
1301                     'player_url': player_url,
1302                 }
1303                 if itag in self._formats:
1304                     dct.update(self._formats[itag])
1305                 formats.append(dct)
1306             return formats
1307
1308         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
1309             self.report_rtmp_download()
1310             formats = [{
1311                 'format_id': '_rtmp',
1312                 'protocol': 'rtmp',
1313                 'url': video_info['conn'][0],
1314                 'player_url': player_url,
1315             }]
1316         elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1:
1317             encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0]
1318             if 'rtmpe%3Dyes' in encoded_url_map:
1319                 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
1320             url_map = {}
1321             for url_data_str in encoded_url_map.split(','):
1322                 url_data = compat_parse_qs(url_data_str)
1323                 if 'itag' in url_data and 'url' in url_data:
1324                     url = url_data['url'][0]
1325                     if 'sig' in url_data:
1326                         url += '&signature=' + url_data['sig'][0]
1327                     elif 's' in url_data:
1328                         encrypted_sig = url_data['s'][0]
1329                         if self._downloader.params.get('verbose'):
1330                             if age_gate:
1331                                 if player_url is None:
1332                                     player_version = 'unknown'
1333                                 else:
1334                                     player_version = self._search_regex(
1335                                         r'-(.+)\.swf$', player_url,
1336                                         u'flash player', fatal=False)
1337                                 player_desc = 'flash player %s' % player_version
1338                             else:
1339                                 player_version = self._search_regex(
1340                                     r'html5player-(.+?)\.js', video_webpage,
1341                                     'html5 player', fatal=False)
1342                                 player_desc = u'html5 player %s' % player_version
1343
1344                             parts_sizes = u'.'.join(compat_str(len(part)) for part in encrypted_sig.split('.'))
1345                             self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' %
1346                                 (len(encrypted_sig), parts_sizes, url_data['itag'][0], player_desc))
1347
1348                         if not age_gate:
1349                             jsplayer_url_json = self._search_regex(
1350                                 r'"assets":.+?"js":\s*("[^"]+")',
1351                                 video_webpage, u'JS player URL')
1352                             player_url = json.loads(jsplayer_url_json)
1353
1354                         signature = self._decrypt_signature(
1355                             encrypted_sig, video_id, player_url, age_gate)
1356                         url += '&signature=' + signature
1357                     if 'ratebypass' not in url:
1358                         url += '&ratebypass=yes'
1359                     url_map[url_data['itag'][0]] = url
1360             formats = _map_to_format_list(url_map)
1361         elif video_info.get('hlsvp'):
1362             manifest_url = video_info['hlsvp'][0]
1363             url_map = self._extract_from_m3u8(manifest_url, video_id)
1364             formats = _map_to_format_list(url_map)
1365         else:
1366             raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
1367
1368         # Look for the DASH manifest
1369         if (self._downloader.params.get('youtube_include_dash_manifest', False)):
1370             try:
1371                 # The DASH manifest used needs to be the one from the original video_webpage.
1372                 # The one found in get_video_info seems to be using different signatures.
1373                 # However, in the case of an age restriction there won't be any embedded dashmpd in the video_webpage.
1374                 # Luckily, it seems, this case uses some kind of default signature (len == 86), so the
1375                 # combination of get_video_info and the _static_decrypt_signature() decryption fallback will work here.
1376                 if age_gate:
1377                     dash_manifest_url = video_info.get('dashmpd')[0];
1378                 else:
1379                     x = re.search(r'ytplayer\.config = ({.*});', video_webpage)
1380                     x = json.loads(x.group(1));
1381                     dash_manifest_url = x['args']['dashmpd']
1382                 def decrypt_sig(mobj):
1383                     s = mobj.group(1)
1384                     dec_s = self._decrypt_signature(s, video_id, player_url, age_gate)
1385                     return '/signature/%s' % dec_s
1386                 dash_manifest_url = re.sub(r'/s/([\w\.]+)', decrypt_sig, dash_manifest_url)
1387                 dash_doc = self._download_xml(
1388                     dash_manifest_url, video_id,
1389                     note=u'Downloading DASH manifest',
1390                     errnote=u'Could not download DASH manifest')
1391                 for r in dash_doc.findall(u'.//{urn:mpeg:DASH:schema:MPD:2011}Representation'):
1392                     url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL')
1393                     if url_el is None:
1394                         continue
1395                     format_id = r.attrib['id']
1396                     video_url = url_el.text
1397                     filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
1398                     f = {
1399                         'format_id': format_id,
1400                         'url': video_url,
1401                         'width': int_or_none(r.attrib.get('width')),
1402                         'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
1403                         'asr': int_or_none(r.attrib.get('audioSamplingRate')),
1404                         'filesize': filesize,
1405                     }
1406                     try:
1407                         existing_format = next(
1408                             fo for fo in formats
1409                             if fo['format_id'] == format_id)
1410                     except StopIteration:
1411                         f.update(self._formats.get(format_id, {}))
1412                         formats.append(f)
1413                     else:
1414                         existing_format.update(f)
1415
1416             except (ExtractorError, KeyError) as e:
1417                 self.report_warning(u'Skipping DASH manifest: %s' % e, video_id)
1418
1419         self._sort_formats(formats)
1420
1421         return {
1422             'id':           video_id,
1423             'uploader':     video_uploader,
1424             'uploader_id':  video_uploader_id,
1425             'upload_date':  upload_date,
1426             'title':        video_title,
1427             'thumbnail':    video_thumbnail,
1428             'description':  video_description,
1429             'subtitles':    video_subtitles,
1430             'duration':     video_duration,
1431             'age_limit':    18 if age_gate else 0,
1432             'annotations':  video_annotations,
1433             'webpage_url': 'https://www.youtube.com/watch?v=%s' % video_id,
1434             'view_count':   view_count,
1435             'like_count': like_count,
1436             'dislike_count': dislike_count,
1437             'formats':      formats,
1438         }
1439
1440 class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
1441     IE_DESC = u'YouTube.com playlists'
1442     _VALID_URL = r"""(?x)(?:
1443                         (?:https?://)?
1444                         (?:\w+\.)?
1445                         youtube\.com/
1446                         (?:
1447                            (?:course|view_play_list|my_playlists|artist|playlist|watch)
1448                            \? (?:.*?&)*? (?:p|a|list)=
1449                         |  p/
1450                         )
1451                         (
1452                             (?:PL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,}
1453                             # Top tracks, they can also include dots
1454                             |(?:MC)[\w\.]*
1455                         )
1456                         .*
1457                      |
1458                         ((?:PL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,})
1459                      )"""
1460     _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s&page=%s'
1461     _MORE_PAGES_INDICATOR = r'data-link-type="next"'
1462     _VIDEO_RE = r'href="/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&amp;[^"]*?index=(?P<index>\d+)'
1463     IE_NAME = u'youtube:playlist'
1464
1465     def _real_initialize(self):
1466         self._login()
1467
1468     def _ids_to_results(self, ids):
1469         return [self.url_result(vid_id, 'Youtube', video_id=vid_id)
1470                        for vid_id in ids]
1471
1472     def _extract_mix(self, playlist_id):
1473         # The mixes are generated from a a single video
1474         # the id of the playlist is just 'RD' + video_id
1475         url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id)
1476         webpage = self._download_webpage(url, playlist_id, u'Downloading Youtube mix')
1477         title_span = (get_element_by_attribute('class', 'title long-title', webpage) or
1478             get_element_by_attribute('class', 'title ', webpage))
1479         title = clean_html(title_span)
1480         video_re = r'data-index="\d+".*?href="/watch\?v=([0-9A-Za-z_-]{11})&amp;[^"]*?list=%s' % re.escape(playlist_id)
1481         ids = orderedSet(re.findall(video_re, webpage))
1482         url_results = self._ids_to_results(ids)
1483
1484         return self.playlist_result(url_results, playlist_id, title)
1485
1486     def _real_extract(self, url):
1487         # Extract playlist id
1488         mobj = re.match(self._VALID_URL, url)
1489         if mobj is None:
1490             raise ExtractorError(u'Invalid URL: %s' % url)
1491         playlist_id = mobj.group(1) or mobj.group(2)
1492
1493         # Check if it's a video-specific URL
1494         query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1495         if 'v' in query_dict:
1496             video_id = query_dict['v'][0]
1497             if self._downloader.params.get('noplaylist'):
1498                 self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
1499                 return self.url_result(video_id, 'Youtube', video_id=video_id)
1500             else:
1501                 self.to_screen(u'Downloading playlist PL%s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1502
1503         if playlist_id.startswith('RD'):
1504             # Mixes require a custom extraction process
1505             return self._extract_mix(playlist_id)
1506         if playlist_id.startswith('TL'):
1507             raise ExtractorError(u'For downloading YouTube.com top lists, use '
1508                 u'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True)
1509
1510         # Extract the video ids from the playlist pages
1511         ids = []
1512
1513         for page_num in itertools.count(1):
1514             url = self._TEMPLATE_URL % (playlist_id, page_num)
1515             page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
1516             matches = re.finditer(self._VIDEO_RE, page)
1517             # We remove the duplicates and the link with index 0
1518             # (it's not the first video of the playlist)
1519             new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
1520             ids.extend(new_ids)
1521
1522             if re.search(self._MORE_PAGES_INDICATOR, page) is None:
1523                 break
1524
1525         try:
1526             playlist_title = self._og_search_title(page)
1527         except RegexNotFoundError:
1528             self.report_warning(
1529                 u'Playlist page is missing OpenGraph title, falling back ...',
1530                 playlist_id)
1531             playlist_title = self._html_search_regex(
1532                 r'<h1 class="pl-header-title">(.*?)</h1>', page, u'title')
1533
1534         url_results = self._ids_to_results(ids)
1535         return self.playlist_result(url_results, playlist_id, playlist_title)
1536
1537
1538 class YoutubeTopListIE(YoutubePlaylistIE):
1539     IE_NAME = u'youtube:toplist'
1540     IE_DESC = (u'YouTube.com top lists, "yttoplist:{channel}:{list title}"'
1541         u' (Example: "yttoplist:music:Top Tracks")')
1542     _VALID_URL = r'yttoplist:(?P<chann>.*?):(?P<title>.*?)$'
1543
1544     def _real_extract(self, url):
1545         mobj = re.match(self._VALID_URL, url)
1546         channel = mobj.group('chann')
1547         title = mobj.group('title')
1548         query = compat_urllib_parse.urlencode({'title': title})
1549         playlist_re = 'href="([^"]+?%s.*?)"' % re.escape(query)
1550         channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title)
1551         link = self._html_search_regex(playlist_re, channel_page, u'list')
1552         url = compat_urlparse.urljoin('https://www.youtube.com/', link)
1553
1554         video_re = r'data-index="\d+".*?data-video-id="([0-9A-Za-z_-]{11})"'
1555         ids = []
1556         # sometimes the webpage doesn't contain the videos
1557         # retry until we get them
1558         for i in itertools.count(0):
1559             msg = u'Downloading Youtube mix'
1560             if i > 0:
1561                 msg += ', retry #%d' % i
1562             webpage = self._download_webpage(url, title, msg)
1563             ids = orderedSet(re.findall(video_re, webpage))
1564             if ids:
1565                 break
1566         url_results = self._ids_to_results(ids)
1567         return self.playlist_result(url_results, playlist_title=title)
1568
1569
1570 class YoutubeChannelIE(InfoExtractor):
1571     IE_DESC = u'YouTube.com channels'
1572     _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
1573     _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
1574     _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'
1575     IE_NAME = u'youtube:channel'
1576
1577     def extract_videos_from_page(self, page):
1578         ids_in_page = []
1579         for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
1580             if mobj.group(1) not in ids_in_page:
1581                 ids_in_page.append(mobj.group(1))
1582         return ids_in_page
1583
1584     def _real_extract(self, url):
1585         # Extract channel id
1586         mobj = re.match(self._VALID_URL, url)
1587         if mobj is None:
1588             raise ExtractorError(u'Invalid URL: %s' % url)
1589
1590         # Download channel page
1591         channel_id = mobj.group(1)
1592         video_ids = []
1593         url = 'https://www.youtube.com/channel/%s/videos' % channel_id
1594         channel_page = self._download_webpage(url, channel_id)
1595         autogenerated = re.search(r'''(?x)
1596                 class="[^"]*?(?:
1597                     channel-header-autogenerated-label|
1598                     yt-channel-title-autogenerated
1599                 )[^"]*"''', channel_page) is not None
1600
1601         if autogenerated:
1602             # The videos are contained in a single page
1603             # the ajax pages can't be used, they are empty
1604             video_ids = self.extract_videos_from_page(channel_page)
1605         else:
1606             # Download all channel pages using the json-based channel_ajax query
1607             for pagenum in itertools.count(1):
1608                 url = self._MORE_PAGES_URL % (pagenum, channel_id)
1609                 page = self._download_json(
1610                     url, channel_id, note=u'Downloading page #%s' % pagenum,
1611                     transform_source=uppercase_escape)
1612
1613                 ids_in_page = self.extract_videos_from_page(page['content_html'])
1614                 video_ids.extend(ids_in_page)
1615
1616                 if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
1617                     break
1618
1619         self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
1620
1621         url_entries = [self.url_result(video_id, 'Youtube', video_id=video_id)
1622                        for video_id in video_ids]
1623         return self.playlist_result(url_entries, channel_id)
1624
1625
1626 class YoutubeUserIE(InfoExtractor):
1627     IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
1628     _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
1629     _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s'
1630     _GDATA_PAGE_SIZE = 50
1631     _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
1632     IE_NAME = u'youtube:user'
1633
1634     @classmethod
1635     def suitable(cls, url):
1636         # Don't return True if the url can be extracted with other youtube
1637         # extractor, the regex would is too permissive and it would match.
1638         other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1639         if any(ie.suitable(url) for ie in other_ies): return False
1640         else: return super(YoutubeUserIE, cls).suitable(url)
1641
1642     def _real_extract(self, url):
1643         # Extract username
1644         mobj = re.match(self._VALID_URL, url)
1645         if mobj is None:
1646             raise ExtractorError(u'Invalid URL: %s' % url)
1647
1648         username = mobj.group(1)
1649
1650         # Download video ids using YouTube Data API. Result size per
1651         # query is limited (currently to 50 videos) so we need to query
1652         # page by page until there are no video ids - it means we got
1653         # all of them.
1654
1655         def download_page(pagenum):
1656             start_index = pagenum * self._GDATA_PAGE_SIZE + 1
1657
1658             gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
1659             page = self._download_webpage(
1660                 gdata_url, username,
1661                 u'Downloading video ids from %d to %d' % (
1662                     start_index, start_index + self._GDATA_PAGE_SIZE))
1663
1664             try:
1665                 response = json.loads(page)
1666             except ValueError as err:
1667                 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1668             if 'entry' not in response['feed']:
1669                 return
1670
1671             # Extract video identifiers
1672             entries = response['feed']['entry']
1673             for entry in entries:
1674                 title = entry['title']['$t']
1675                 video_id = entry['id']['$t'].split('/')[-1]
1676                 yield {
1677                     '_type': 'url',
1678                     'url': video_id,
1679                     'ie_key': 'Youtube',
1680                     'id': video_id,
1681                     'title': title,
1682                 }
1683         url_results = PagedList(download_page, self._GDATA_PAGE_SIZE)
1684
1685         return self.playlist_result(url_results, playlist_title=username)
1686
1687
1688 class YoutubeSearchIE(SearchInfoExtractor):
1689     IE_DESC = u'YouTube.com searches'
1690     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
1691     _MAX_RESULTS = 1000
1692     IE_NAME = u'youtube:search'
1693     _SEARCH_KEY = 'ytsearch'
1694
1695     def _get_n_results(self, query, n):
1696         """Get a specified number of results for a query"""
1697
1698         video_ids = []
1699         pagenum = 0
1700         limit = n
1701
1702         while (50 * pagenum) < limit:
1703             result_url = self._API_URL % (compat_urllib_parse.quote_plus(query), (50*pagenum)+1)
1704             data_json = self._download_webpage(
1705                 result_url, video_id=u'query "%s"' % query,
1706                 note=u'Downloading page %s' % (pagenum + 1),
1707                 errnote=u'Unable to download API page')
1708             data = json.loads(data_json)
1709             api_response = data['data']
1710
1711             if 'items' not in api_response:
1712                 raise ExtractorError(
1713                     u'[youtube] No video results', expected=True)
1714
1715             new_ids = list(video['id'] for video in api_response['items'])
1716             video_ids += new_ids
1717
1718             limit = min(n, api_response['totalItems'])
1719             pagenum += 1
1720
1721         if len(video_ids) > n:
1722             video_ids = video_ids[:n]
1723         videos = [self.url_result(video_id, 'Youtube', video_id=video_id)
1724                   for video_id in video_ids]
1725         return self.playlist_result(videos, query)
1726
1727 class YoutubeSearchDateIE(YoutubeSearchIE):
1728     IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
1729     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published'
1730     _SEARCH_KEY = 'ytsearchdate'
1731     IE_DESC = u'YouTube.com searches, newest videos first'
1732
1733 class YoutubeShowIE(InfoExtractor):
1734     IE_DESC = u'YouTube.com (multi-season) shows'
1735     _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
1736     IE_NAME = u'youtube:show'
1737
1738     def _real_extract(self, url):
1739         mobj = re.match(self._VALID_URL, url)
1740         show_name = mobj.group(1)
1741         webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
1742         # There's one playlist for each season of the show
1743         m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1744         self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
1745         return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
1746
1747
1748 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1749     """
1750     Base class for extractors that fetch info from
1751     http://www.youtube.com/feed_ajax
1752     Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1753     """
1754     _LOGIN_REQUIRED = True
1755     # use action_load_personal_feed instead of action_load_system_feed
1756     _PERSONAL_FEED = False
1757
1758     @property
1759     def _FEED_TEMPLATE(self):
1760         action = 'action_load_system_feed'
1761         if self._PERSONAL_FEED:
1762             action = 'action_load_personal_feed'
1763         return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
1764
1765     @property
1766     def IE_NAME(self):
1767         return u'youtube:%s' % self._FEED_NAME
1768
1769     def _real_initialize(self):
1770         self._login()
1771
1772     def _real_extract(self, url):
1773         feed_entries = []
1774         paging = 0
1775         for i in itertools.count(1):
1776             info = self._download_webpage(self._FEED_TEMPLATE % paging,
1777                                           u'%s feed' % self._FEED_NAME,
1778                                           u'Downloading page %s' % i)
1779             info = json.loads(info)
1780             feed_html = info['feed_html']
1781             m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
1782             ids = orderedSet(m.group(1) for m in m_ids)
1783             feed_entries.extend(
1784                 self.url_result(video_id, 'Youtube', video_id=video_id)
1785                 for video_id in ids)
1786             if info['paging'] is None:
1787                 break
1788             paging = info['paging']
1789         return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
1790
1791 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1792     IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword(requires authentication)'
1793     _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1794     _FEED_NAME = 'subscriptions'
1795     _PLAYLIST_TITLE = u'Youtube Subscriptions'
1796
1797 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1798     IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
1799     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1800     _FEED_NAME = 'recommended'
1801     _PLAYLIST_TITLE = u'Youtube Recommended videos'
1802
1803 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
1804     IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
1805     _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
1806     _FEED_NAME = 'watch_later'
1807     _PLAYLIST_TITLE = u'Youtube Watch Later'
1808     _PERSONAL_FEED = True
1809
1810 class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
1811     IE_DESC = u'Youtube watch history, "ythistory" keyword (requires authentication)'
1812     _VALID_URL = u'https?://www\.youtube\.com/feed/history|:ythistory'
1813     _FEED_NAME = 'history'
1814     _PERSONAL_FEED = True
1815     _PLAYLIST_TITLE = u'Youtube Watch History'
1816
1817 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1818     IE_NAME = u'youtube:favorites'
1819     IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
1820     _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1821     _LOGIN_REQUIRED = True
1822
1823     def _real_extract(self, url):
1824         webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1825         playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
1826         return self.url_result(playlist_id, 'YoutubePlaylist')
1827
1828
1829 class YoutubeTruncatedURLIE(InfoExtractor):
1830     IE_NAME = 'youtube:truncated_url'
1831     IE_DESC = False  # Do not list
1832     _VALID_URL = r'''(?x)
1833         (?:https?://)?[^/]+/watch\?(?:feature=[a-z_]+)?$|
1834         (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$
1835     '''
1836
1837     def _real_extract(self, url):
1838         raise ExtractorError(
1839             u'Did you forget to quote the URL? Remember that & is a meta '
1840             u'character in most shells, so you want to put the URL in quotes, '
1841             u'like  youtube-dl '
1842             u'"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
1843             u' or simply  youtube-dl BaW_jenozKc  .',
1844             expected=True)