Raphaël G. Git Repositories - youtubedl/blob - 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                             tube\.majestyc\.net/|
 142                             youtube\.googleapis\.com/)                        # the various hostnames, with wildcard subdomains
 143                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
 144                          (?:                                                  # the various things that can precede the ID:
 145                              (?:(?:v|embed|e)/)                               # v/ or embed/ or e/
 146                              |(?:                                             # or the v= param in all its forms
 147                                  (?:(?:watch|movie)(?:_popup)?(?:\.php)?)?    # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
 148                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
 149                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
 150                                  v=
 151                              )
 152                          ))
 153                          |youtu\.be/                                          # just youtu.be/xxxx
 154                          )
 155                      )?                                                       # all until now is optional -> you can pass the naked ID
 156                      ([0-9A-Za-z_-]{11})                                      # here is it! the YouTube video ID
 157                      (?(1).+)?                                                # if we found the ID, everything can follow
 158                      $"""
 159     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
 160     _formats = {
 161         '5': {'ext': 'flv', 'width': 400, 'height': 240},
 162         '6': {'ext': 'flv', 'width': 450, 'height': 270},
 163         '13': {'ext': '3gp'},
 164         '17': {'ext': '3gp', 'width': 176, 'height': 144},
 165         '18': {'ext': 'mp4', 'width': 640, 'height': 360},
 166         '22': {'ext': 'mp4', 'width': 1280, 'height': 720},
 167         '34': {'ext': 'flv', 'width': 640, 'height': 360},
 168         '35': {'ext': 'flv', 'width': 854, 'height': 480},
 169         '36': {'ext': '3gp', 'width': 320, 'height': 240},
 170         '37': {'ext': 'mp4', 'width': 1920, 'height': 1080},
 171         '38': {'ext': 'mp4', 'width': 4096, 'height': 3072},
 172         '43': {'ext': 'webm', 'width': 640, 'height': 360},
 173         '44': {'ext': 'webm', 'width': 854, 'height': 480},
 174         '45': {'ext': 'webm', 'width': 1280, 'height': 720},
 175         '46': {'ext': 'webm', 'width': 1920, 'height': 1080},
 176
 177
 178         # 3d videos
 179         '82': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': '3D', 'preference': -20},
 180         '83': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': '3D', 'preference': -20},
 181         '84': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': '3D', 'preference': -20},
 182         '85': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': '3D', 'preference': -20},
 183         '100': {'ext': 'webm', 'height': 360, 'resolution': '360p', 'format_note': '3D', 'preference': -20},
 184         '101': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': '3D', 'preference': -20},
 185         '102': {'ext': 'webm', 'height': 720, 'resolution': '720p', 'format_note': '3D', 'preference': -20},
 186
 187         # Apple HTTP Live Streaming
 188         '92': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'HLS', 'preference': -10},
 189         '93': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': 'HLS', 'preference': -10},
 190         '94': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': 'HLS', 'preference': -10},
 191         '95': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': 'HLS', 'preference': -10},
 192         '96': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'HLS', 'preference': -10},
 193         '132': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'HLS', 'preference': -10},
 194         '151': {'ext': 'mp4', 'height': 72, 'resolution': '72p', 'format_note': 'HLS', 'preference': -10},
 195
 196         # DASH mp4 video
 197         '133': {'ext': 'mp4', 'height': 240, 'resolution': '240p', 'format_note': 'DASH video', 'preference': -40},
 198         '134': {'ext': 'mp4', 'height': 360, 'resolution': '360p', 'format_note': 'DASH video', 'preference': -40},
 199         '135': {'ext': 'mp4', 'height': 480, 'resolution': '480p', 'format_note': 'DASH video', 'preference': -40},
 200         '136': {'ext': 'mp4', 'height': 720, 'resolution': '720p', 'format_note': 'DASH video', 'preference': -40},
 201         '137': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH video', 'preference': -40},
 202         '138': {'ext': 'mp4', 'height': 1081, 'resolution': '>1080p', 'format_note': 'DASH video', 'preference': -40},
 203         '160': {'ext': 'mp4', 'height': 192, 'resolution': '192p', 'format_note': 'DASH video', 'preference': -40},
 204         '264': {'ext': 'mp4', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH video', 'preference': -40},
 205
 206         # Dash mp4 audio
 207         '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50},
 208         '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50},
 209         '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50},
 210
 211         # Dash webm
 212         '167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 213         '168': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 214         '169': {'ext': 'webm', 'height': 720, 'width': 1280, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 215         '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 216         '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 217         '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'container': 'webm', 'vcodec': 'VP8', 'acodec': 'none', 'preference': -40},
 218         '242': {'ext': 'webm', 'height': 240, 'resolution': '240p', 'format_note': 'DASH webm', 'preference': -40},
 219         '243': {'ext': 'webm', 'height': 360, 'resolution': '360p', 'format_note': 'DASH webm', 'preference': -40},
 220         '244': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40},
 221         '245': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40},
 222         '246': {'ext': 'webm', 'height': 480, 'resolution': '480p', 'format_note': 'DASH webm', 'preference': -40},
 223         '247': {'ext': 'webm', 'height': 720, 'resolution': '720p', 'format_note': 'DASH webm', 'preference': -40},
 224         '248': {'ext': 'webm', 'height': 1080, 'resolution': '1080p', 'format_note': 'DASH webm', 'preference': -40},
 225
 226         # Dash webm audio
 227         '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm audio', 'abr': 48, 'preference': -50},
 228         '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH webm audio', 'abr': 256, 'preference': -50},
 229
 230         # RTMP (unnamed)
 231         '_rtmp': {'protocol': 'rtmp'},
 232     }
 233
 234     IE_NAME = u'youtube'
 235     _TESTS = [
 236         {
 237             u"url":  u"http://www.youtube.com/watch?v=BaW_jenozKc",
 238             u"file":  u"BaW_jenozKc.mp4",
 239             u"info_dict": {
 240                 u"title": u"youtube-dl test video \"'/\\ä↭𝕐",
 241                 u"uploader": u"Philipp Hagemeister",
 242                 u"uploader_id": u"phihag",
 243                 u"upload_date": u"20121002",
 244                 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 ."
 245             }
 246         },
 247         {
 248             u"url":  u"http://www.youtube.com/watch?v=UxxajLWwzqY",
 249             u"file":  u"UxxajLWwzqY.mp4",
 250             u"note": u"Test generic use_cipher_signature video (#897)",
 251             u"info_dict": {
 252                 u"upload_date": u"20120506",
 253                 u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
 254                 u"description": u"md5:5b292926389560516e384ac437c0ec07",
 255                 u"uploader": u"Icona Pop",
 256                 u"uploader_id": u"IconaPop"
 257             }
 258         },
 259         {
 260             u"url":  u"https://www.youtube.com/watch?v=07FYdnEawAQ",
 261             u"file":  u"07FYdnEawAQ.mp4",
 262             u"note": u"Test VEVO video with age protection (#956)",
 263             u"info_dict": {
 264                 u"upload_date": u"20130703",
 265                 u"title": u"Justin Timberlake - Tunnel Vision (Explicit)",
 266                 u"description": u"md5:64249768eec3bc4276236606ea996373",
 267                 u"uploader": u"justintimberlakeVEVO",
 268                 u"uploader_id": u"justintimberlakeVEVO"
 269             }
 270         },
 271         {
 272             u"url":  u"//www.YouTube.com/watch?v=yZIXLfi8CZQ",
 273             u"file":  u"yZIXLfi8CZQ.mp4",
 274             u"note": u"Embed-only video (#1746)",
 275             u"info_dict": {
 276                 u"upload_date": u"20120608",
 277                 u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012",
 278                 u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7",
 279                 u"uploader": u"SET India",
 280                 u"uploader_id": u"setindia"
 281             }
 282         },
 283         {
 284             u"url": u"http://www.youtube.com/watch?v=a9LDPn-MO4I",
 285             u"file": u"a9LDPn-MO4I.m4a",
 286             u"note": u"256k DASH audio (format 141) via DASH manifest",
 287             u"info_dict": {
 288                 u"upload_date": "20121002",
 289                 u"uploader_id": "8KVIDEO",
 290                 u"description": "No description available.",
 291                 u"uploader": "8KVIDEO",
 292                 u"title": "UHDTV TEST 8K VIDEO.mp4"
 293             },
 294             u"params": {
 295                 u"youtube_include_dash_manifest": True,
 296                 u"format": "141",
 297             },
 298         },
 299     ]
 300
 301
 302     @classmethod
 303     def suitable(cls, url):
 304         """Receives a URL and returns True if suitable for this IE."""
 305         if YoutubePlaylistIE.suitable(url): return False
 306         return re.match(cls._VALID_URL, url) is not None
 307
 308     def __init__(self, *args, **kwargs):
 309         super(YoutubeIE, self).__init__(*args, **kwargs)
 310         self._player_cache = {}
 311
 312     def report_video_info_webpage_download(self, video_id):
 313         """Report attempt to download video info webpage."""
 314         self.to_screen(u'%s: Downloading video info webpage' % video_id)
 315
 316     def report_information_extraction(self, video_id):
 317         """Report attempt to extract video information."""
 318         self.to_screen(u'%s: Extracting video information' % video_id)
 319
 320     def report_unavailable_format(self, video_id, format):
 321         """Report extracted video URL."""
 322         self.to_screen(u'%s: Format %s not available' % (video_id, format))
 323
 324     def report_rtmp_download(self):
 325         """Indicate the download will use the RTMP protocol."""
 326         self.to_screen(u'RTMP download detected')
 327
 328     def _extract_signature_function(self, video_id, player_url, slen):
 329         id_m = re.match(r'.*-(?P<id>[a-zA-Z0-9_-]+)\.(?P<ext>[a-z]+)$',
 330                         player_url)
 331         player_type = id_m.group('ext')
 332         player_id = id_m.group('id')
 333
 334         # Read from filesystem cache
 335         func_id = '%s_%s_%d' % (player_type, player_id, slen)
 336         assert os.path.basename(func_id) == func_id
 337         cache_dir = get_cachedir(self._downloader.params)
 338
 339         cache_enabled = cache_dir is not None
 340         if cache_enabled:
 341             cache_fn = os.path.join(os.path.expanduser(cache_dir),
 342                                     u'youtube-sigfuncs',
 343                                     func_id + '.json')
 344             try:
 345                 with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
 346                     cache_spec = json.load(cachef)
 347                 return lambda s: u''.join(s[i] for i in cache_spec)
 348             except IOError:
 349                 pass  # No cache available
 350
 351         if player_type == 'js':
 352             code = self._download_webpage(
 353                 player_url, video_id,
 354                 note=u'Downloading %s player %s' % (player_type, player_id),
 355                 errnote=u'Download of %s failed' % player_url)
 356             res = self._parse_sig_js(code)
 357         elif player_type == 'swf':
 358             urlh = self._request_webpage(
 359                 player_url, video_id,
 360                 note=u'Downloading %s player %s' % (player_type, player_id),
 361                 errnote=u'Download of %s failed' % player_url)
 362             code = urlh.read()
 363             res = self._parse_sig_swf(code)
 364         else:
 365             assert False, 'Invalid player type %r' % player_type
 366
 367         if cache_enabled:
 368             try:
 369                 test_string = u''.join(map(compat_chr, range(slen)))
 370                 cache_res = res(test_string)
 371                 cache_spec = [ord(c) for c in cache_res]
 372                 try:
 373                     os.makedirs(os.path.dirname(cache_fn))
 374                 except OSError as ose:
 375                     if ose.errno != errno.EEXIST:
 376                         raise
 377                 write_json_file(cache_spec, cache_fn)
 378             except Exception:
 379                 tb = traceback.format_exc()
 380                 self._downloader.report_warning(
 381                     u'Writing cache to %r failed: %s' % (cache_fn, tb))
 382
 383         return res
 384
 385     def _print_sig_code(self, func, slen):
 386         def gen_sig_code(idxs):
 387             def _genslice(start, end, step):
 388                 starts = u'' if start == 0 else str(start)
 389                 ends = (u':%d' % (end+step)) if end + step >= 0 else u':'
 390                 steps = u'' if step == 1 else (u':%d' % step)
 391                 return u's[%s%s%s]' % (starts, ends, steps)
 392
 393             step = None
 394             start = '(Never used)'  # Quelch pyflakes warnings - start will be
 395                                     # set as soon as step is set
 396             for i, prev in zip(idxs[1:], idxs[:-1]):
 397                 if step is not None:
 398                     if i - prev == step:
 399                         continue
 400                     yield _genslice(start, prev, step)
 401                     step = None
 402                     continue
 403                 if i - prev in [-1, 1]:
 404                     step = i - prev
 405                     start = prev
 406                     continue
 407                 else:
 408                     yield u's[%d]' % prev
 409             if step is None:
 410                 yield u's[%d]' % i
 411             else:
 412                 yield _genslice(start, i, step)
 413
 414         test_string = u''.join(map(compat_chr, range(slen)))
 415         cache_res = func(test_string)
 416         cache_spec = [ord(c) for c in cache_res]
 417         expr_code = u' + '.join(gen_sig_code(cache_spec))
 418         code = u'if len(s) == %d:\n    return %s\n' % (slen, expr_code)
 419         self.to_screen(u'Extracted signature function:\n' + code)
 420
 421     def _parse_sig_js(self, jscode):
 422         funcname = self._search_regex(
 423             r'signature=([a-zA-Z]+)', jscode,
 424             u'Initial JS player signature function name')
 425
 426         functions = {}
 427
 428         def argidx(varname):
 429             return string.lowercase.index(varname)
 430
 431         def interpret_statement(stmt, local_vars, allow_recursion=20):
 432             if allow_recursion < 0:
 433                 raise ExtractorError(u'Recursion limit reached')
 434
 435             if stmt.startswith(u'var '):
 436                 stmt = stmt[len(u'var '):]
 437             ass_m = re.match(r'^(?P<out>[a-z]+)(?:\[(?P<index>[^\]]+)\])?' +
 438                              r'=(?P<expr>.*)$', stmt)
 439             if ass_m:
 440                 if ass_m.groupdict().get('index'):
 441                     def assign(val):
 442                         lvar = local_vars[ass_m.group('out')]
 443                         idx = interpret_expression(ass_m.group('index'),
 444                                                    local_vars, allow_recursion)
 445                         assert isinstance(idx, int)
 446                         lvar[idx] = val
 447                         return val
 448                     expr = ass_m.group('expr')
 449                 else:
 450                     def assign(val):
 451                         local_vars[ass_m.group('out')] = val
 452                         return val
 453                     expr = ass_m.group('expr')
 454             elif stmt.startswith(u'return '):
 455                 assign = lambda v: v
 456                 expr = stmt[len(u'return '):]
 457             else:
 458                 raise ExtractorError(
 459                     u'Cannot determine left side of statement in %r' % stmt)
 460
 461             v = interpret_expression(expr, local_vars, allow_recursion)
 462             return assign(v)
 463
 464         def interpret_expression(expr, local_vars, allow_recursion):
 465             if expr.isdigit():
 466                 return int(expr)
 467
 468             if expr.isalpha():
 469                 return local_vars[expr]
 470
 471             m = re.match(r'^(?P<in>[a-z]+)\.(?P<member>.*)$', expr)
 472             if m:
 473                 member = m.group('member')
 474                 val = local_vars[m.group('in')]
 475                 if member == 'split("")':
 476                     return list(val)
 477                 if member == 'join("")':
 478                     return u''.join(val)
 479                 if member == 'length':
 480                     return len(val)
 481                 if member == 'reverse()':
 482                     return val[::-1]
 483                 slice_m = re.match(r'slice\((?P<idx>.*)\)', member)
 484                 if slice_m:
 485                     idx = interpret_expression(
 486                         slice_m.group('idx'), local_vars, allow_recursion-1)
 487                     return val[idx:]
 488
 489             m = re.match(
 490                 r'^(?P<in>[a-z]+)\[(?P<idx>.+)\]$', expr)
 491             if m:
 492                 val = local_vars[m.group('in')]
 493                 idx = interpret_expression(m.group('idx'), local_vars,
 494                                            allow_recursion-1)
 495                 return val[idx]
 496
 497             m = re.match(r'^(?P<a>.+?)(?P<op>[%])(?P<b>.+?)$', expr)
 498             if m:
 499                 a = interpret_expression(m.group('a'),
 500                                          local_vars, allow_recursion)
 501                 b = interpret_expression(m.group('b'),
 502                                          local_vars, allow_recursion)
 503                 return a % b
 504
 505             m = re.match(
 506                 r'^(?P<func>[a-zA-Z$]+)\((?P<args>[a-z0-9,]+)\)$', expr)
 507             if m:
 508                 fname = m.group('func')
 509                 if fname not in functions:
 510                     functions[fname] = extract_function(fname)
 511                 argvals = [int(v) if v.isdigit() else local_vars[v]
 512                            for v in m.group('args').split(',')]
 513                 return functions[fname](argvals)
 514             raise ExtractorError(u'Unsupported JS expression %r' % expr)
 515
 516         def extract_function(funcname):
 517             func_m = re.search(
 518                 r'function ' + re.escape(funcname) +
 519                 r'\((?P<args>[a-z,]+)\){(?P<code>[^}]+)}',
 520                 jscode)
 521             argnames = func_m.group('args').split(',')
 522
 523             def resf(args):
 524                 local_vars = dict(zip(argnames, args))
 525                 for stmt in func_m.group('code').split(';'):
 526                     res = interpret_statement(stmt, local_vars)
 527                 return res
 528             return resf
 529
 530         initial_function = extract_function(funcname)
 531         return lambda s: initial_function([s])
 532
 533     def _parse_sig_swf(self, file_contents):
 534         if file_contents[1:3] != b'WS':
 535             raise ExtractorError(
 536                 u'Not an SWF file; header is %r' % file_contents[:3])
 537         if file_contents[:1] == b'C':
 538             content = zlib.decompress(file_contents[8:])
 539         else:
 540             raise NotImplementedError(u'Unsupported compression format %r' %
 541                                       file_contents[:1])
 542
 543         def extract_tags(content):
 544             pos = 0
 545             while pos < len(content):
 546                 header16 = struct.unpack('<H', content[pos:pos+2])[0]
 547                 pos += 2
 548                 tag_code = header16 >> 6
 549                 tag_len = header16 & 0x3f
 550                 if tag_len == 0x3f:
 551                     tag_len = struct.unpack('<I', content[pos:pos+4])[0]
 552                     pos += 4
 553                 assert pos+tag_len <= len(content)
 554                 yield (tag_code, content[pos:pos+tag_len])
 555                 pos += tag_len
 556
 557         code_tag = next(tag
 558                         for tag_code, tag in extract_tags(content)
 559                         if tag_code == 82)
 560         p = code_tag.index(b'\0', 4) + 1
 561         code_reader = io.BytesIO(code_tag[p:])
 562
 563         # Parse ABC (AVM2 ByteCode)
 564         def read_int(reader=None):
 565             if reader is None:
 566                 reader = code_reader
 567             res = 0
 568             shift = 0
 569             for _ in range(5):
 570                 buf = reader.read(1)
 571                 assert len(buf) == 1
 572                 b = struct.unpack('<B', buf)[0]
 573                 res = res | ((b & 0x7f) << shift)
 574                 if b & 0x80 == 0:
 575                     break
 576                 shift += 7
 577             return res
 578
 579         def u30(reader=None):
 580             res = read_int(reader)
 581             assert res & 0xf0000000 == 0
 582             return res
 583         u32 = read_int
 584
 585         def s32(reader=None):
 586             v = read_int(reader)
 587             if v & 0x80000000 != 0:
 588                 v = - ((v ^ 0xffffffff) + 1)
 589             return v
 590
 591         def read_string(reader=None):
 592             if reader is None:
 593                 reader = code_reader
 594             slen = u30(reader)
 595             resb = reader.read(slen)
 596             assert len(resb) == slen
 597             return resb.decode('utf-8')
 598
 599         def read_bytes(count, reader=None):
 600             if reader is None:
 601                 reader = code_reader
 602             resb = reader.read(count)
 603             assert len(resb) == count
 604             return resb
 605
 606         def read_byte(reader=None):
 607             resb = read_bytes(1, reader=reader)
 608             res = struct.unpack('<B', resb)[0]
 609             return res
 610
 611         # minor_version + major_version
 612         read_bytes(2 + 2)
 613
 614         # Constant pool
 615         int_count = u30()
 616         for _c in range(1, int_count):
 617             s32()
 618         uint_count = u30()
 619         for _c in range(1, uint_count):
 620             u32()
 621         double_count = u30()
 622         read_bytes((double_count-1) * 8)
 623         string_count = u30()
 624         constant_strings = [u'']
 625         for _c in range(1, string_count):
 626             s = read_string()
 627             constant_strings.append(s)
 628         namespace_count = u30()
 629         for _c in range(1, namespace_count):
 630             read_bytes(1)  # kind
 631             u30()  # name
 632         ns_set_count = u30()
 633         for _c in range(1, ns_set_count):
 634             count = u30()
 635             for _c2 in range(count):
 636                 u30()
 637         multiname_count = u30()
 638         MULTINAME_SIZES = {
 639             0x07: 2,  # QName
 640             0x0d: 2,  # QNameA
 641             0x0f: 1,  # RTQName
 642             0x10: 1,  # RTQNameA
 643             0x11: 0,  # RTQNameL
 644             0x12: 0,  # RTQNameLA
 645             0x09: 2,  # Multiname
 646             0x0e: 2,  # MultinameA
 647             0x1b: 1,  # MultinameL
 648             0x1c: 1,  # MultinameLA
 649         }
 650         multinames = [u'']
 651         for _c in range(1, multiname_count):
 652             kind = u30()
 653             assert kind in MULTINAME_SIZES, u'Invalid multiname kind %r' % kind
 654             if kind == 0x07:
 655                 u30()  # namespace_idx
 656                 name_idx = u30()
 657                 multinames.append(constant_strings[name_idx])
 658             else:
 659                 multinames.append('[MULTINAME kind: %d]' % kind)
 660                 for _c2 in range(MULTINAME_SIZES[kind]):
 661                     u30()
 662
 663         # Methods
 664         method_count = u30()
 665         MethodInfo = collections.namedtuple(
 666             'MethodInfo',
 667             ['NEED_ARGUMENTS', 'NEED_REST'])
 668         method_infos = []
 669         for method_id in range(method_count):
 670             param_count = u30()
 671             u30()  # return type
 672             for _ in range(param_count):
 673                 u30()  # param type
 674             u30()  # name index (always 0 for youtube)
 675             flags = read_byte()
 676             if flags & 0x08 != 0:
 677                 # Options present
 678                 option_count = u30()
 679                 for c in range(option_count):
 680                     u30()  # val
 681                     read_bytes(1)  # kind
 682             if flags & 0x80 != 0:
 683                 # Param names present
 684                 for _ in range(param_count):
 685                     u30()  # param name
 686             mi = MethodInfo(flags & 0x01 != 0, flags & 0x04 != 0)
 687             method_infos.append(mi)
 688
 689         # Metadata
 690         metadata_count = u30()
 691         for _c in range(metadata_count):
 692             u30()  # name
 693             item_count = u30()
 694             for _c2 in range(item_count):
 695                 u30()  # key
 696                 u30()  # value
 697
 698         def parse_traits_info():
 699             trait_name_idx = u30()
 700             kind_full = read_byte()
 701             kind = kind_full & 0x0f
 702             attrs = kind_full >> 4
 703             methods = {}
 704             if kind in [0x00, 0x06]:  # Slot or Const
 705                 u30()  # Slot id
 706                 u30()  # type_name_idx
 707                 vindex = u30()
 708                 if vindex != 0:
 709                     read_byte()  # vkind
 710             elif kind in [0x01, 0x02, 0x03]:  # Method / Getter / Setter
 711                 u30()  # disp_id
 712                 method_idx = u30()
 713                 methods[multinames[trait_name_idx]] = method_idx
 714             elif kind == 0x04:  # Class
 715                 u30()  # slot_id
 716                 u30()  # classi
 717             elif kind == 0x05:  # Function
 718                 u30()  # slot_id
 719                 function_idx = u30()
 720                 methods[function_idx] = multinames[trait_name_idx]
 721             else:
 722                 raise ExtractorError(u'Unsupported trait kind %d' % kind)
 723
 724             if attrs & 0x4 != 0:  # Metadata present
 725                 metadata_count = u30()
 726                 for _c3 in range(metadata_count):
 727                     u30()  # metadata index
 728
 729             return methods
 730
 731         # Classes
 732         TARGET_CLASSNAME = u'SignatureDecipher'
 733         searched_idx = multinames.index(TARGET_CLASSNAME)
 734         searched_class_id = None
 735         class_count = u30()
 736         for class_id in range(class_count):
 737             name_idx = u30()
 738             if name_idx == searched_idx:
 739                 # We found the class we're looking for!
 740                 searched_class_id = class_id
 741             u30()  # super_name idx
 742             flags = read_byte()
 743             if flags & 0x08 != 0:  # Protected namespace is present
 744                 u30()  # protected_ns_idx
 745             intrf_count = u30()
 746             for _c2 in range(intrf_count):
 747                 u30()
 748             u30()  # iinit
 749             trait_count = u30()
 750             for _c2 in range(trait_count):
 751                 parse_traits_info()
 752
 753         if searched_class_id is None:
 754             raise ExtractorError(u'Target class %r not found' %
 755                                  TARGET_CLASSNAME)
 756
 757         method_names = {}
 758         method_idxs = {}
 759         for class_id in range(class_count):
 760             u30()  # cinit
 761             trait_count = u30()
 762             for _c2 in range(trait_count):
 763                 trait_methods = parse_traits_info()
 764                 if class_id == searched_class_id:
 765                     method_names.update(trait_methods.items())
 766                     method_idxs.update(dict(
 767                         (idx, name)
 768                         for name, idx in trait_methods.items()))
 769
 770         # Scripts
 771         script_count = u30()
 772         for _c in range(script_count):
 773             u30()  # init
 774             trait_count = u30()
 775             for _c2 in range(trait_count):
 776                 parse_traits_info()
 777
 778         # Method bodies
 779         method_body_count = u30()
 780         Method = collections.namedtuple('Method', ['code', 'local_count'])
 781         methods = {}
 782         for _c in range(method_body_count):
 783             method_idx = u30()
 784             u30()  # max_stack
 785             local_count = u30()
 786             u30()  # init_scope_depth
 787             u30()  # max_scope_depth
 788             code_length = u30()
 789             code = read_bytes(code_length)
 790             if method_idx in method_idxs:
 791                 m = Method(code, local_count)
 792                 methods[method_idxs[method_idx]] = m
 793             exception_count = u30()
 794             for _c2 in range(exception_count):
 795                 u30()  # from
 796                 u30()  # to
 797                 u30()  # target
 798                 u30()  # exc_type
 799                 u30()  # var_name
 800             trait_count = u30()
 801             for _c2 in range(trait_count):
 802                 parse_traits_info()
 803
 804         assert p + code_reader.tell() == len(code_tag)
 805         assert len(methods) == len(method_idxs)
 806
 807         method_pyfunctions = {}
 808
 809         def extract_function(func_name):
 810             if func_name in method_pyfunctions:
 811                 return method_pyfunctions[func_name]
 812             if func_name not in methods:
 813                 raise ExtractorError(u'Cannot find function %r' % func_name)
 814             m = methods[func_name]
 815
 816             def resfunc(args):
 817                 registers = ['(this)'] + list(args) + [None] * m.local_count
 818                 stack = []
 819                 coder = io.BytesIO(m.code)
 820                 while True:
 821                     opcode = struct.unpack('!B', coder.read(1))[0]
 822                     if opcode == 36:  # pushbyte
 823                         v = struct.unpack('!B', coder.read(1))[0]
 824                         stack.append(v)
 825                     elif opcode == 44:  # pushstring
 826                         idx = u30(coder)
 827                         stack.append(constant_strings[idx])
 828                     elif opcode == 48:  # pushscope
 829                         # We don't implement the scope register, so we'll just
 830                         # ignore the popped value
 831                         stack.pop()
 832                     elif opcode == 70:  # callproperty
 833                         index = u30(coder)
 834                         mname = multinames[index]
 835                         arg_count = u30(coder)
 836                         args = list(reversed(
 837                             [stack.pop() for _ in range(arg_count)]))
 838                         obj = stack.pop()
 839                         if mname == u'split':
 840                             assert len(args) == 1
 841                             assert isinstance(args[0], compat_str)
 842                             assert isinstance(obj, compat_str)
 843                             if args[0] == u'':
 844                                 res = list(obj)
 845                             else:
 846                                 res = obj.split(args[0])
 847                             stack.append(res)
 848                         elif mname == u'slice':
 849                             assert len(args) == 1
 850                             assert isinstance(args[0], int)
 851                             assert isinstance(obj, list)
 852                             res = obj[args[0]:]
 853                             stack.append(res)
 854                         elif mname == u'join':
 855                             assert len(args) == 1
 856                             assert isinstance(args[0], compat_str)
 857                             assert isinstance(obj, list)
 858                             res = args[0].join(obj)
 859                             stack.append(res)
 860                         elif mname in method_pyfunctions:
 861                             stack.append(method_pyfunctions[mname](args))
 862                         else:
 863                             raise NotImplementedError(
 864                                 u'Unsupported property %r on %r'
 865                                 % (mname, obj))
 866                     elif opcode == 72:  # returnvalue
 867                         res = stack.pop()
 868                         return res
 869                     elif opcode == 79:  # callpropvoid
 870                         index = u30(coder)
 871                         mname = multinames[index]
 872                         arg_count = u30(coder)
 873                         args = list(reversed(
 874                             [stack.pop() for _ in range(arg_count)]))
 875                         obj = stack.pop()
 876                         if mname == u'reverse':
 877                             assert isinstance(obj, list)
 878                             obj.reverse()
 879                         else:
 880                             raise NotImplementedError(
 881                                 u'Unsupported (void) property %r on %r'
 882                                 % (mname, obj))
 883                     elif opcode == 93:  # findpropstrict
 884                         index = u30(coder)
 885                         mname = multinames[index]
 886                         res = extract_function(mname)
 887                         stack.append(res)
 888                     elif opcode == 97:  # setproperty
 889                         index = u30(coder)
 890                         value = stack.pop()
 891                         idx = stack.pop()
 892                         obj = stack.pop()
 893                         assert isinstance(obj, list)
 894                         assert isinstance(idx, int)
 895                         obj[idx] = value
 896                     elif opcode == 98:  # getlocal
 897                         index = u30(coder)
 898                         stack.append(registers[index])
 899                     elif opcode == 99:  # setlocal
 900                         index = u30(coder)
 901                         value = stack.pop()
 902                         registers[index] = value
 903                     elif opcode == 102:  # getproperty
 904                         index = u30(coder)
 905                         pname = multinames[index]
 906                         if pname == u'length':
 907                             obj = stack.pop()
 908                             assert isinstance(obj, list)
 909                             stack.append(len(obj))
 910                         else:  # Assume attribute access
 911                             idx = stack.pop()
 912                             assert isinstance(idx, int)
 913                             obj = stack.pop()
 914                             assert isinstance(obj, list)
 915                             stack.append(obj[idx])
 916                     elif opcode == 128:  # coerce
 917                         u30(coder)
 918                     elif opcode == 133:  # coerce_s
 919                         assert isinstance(stack[-1], (type(None), compat_str))
 920                     elif opcode == 164:  # modulo
 921                         value2 = stack.pop()
 922                         value1 = stack.pop()
 923                         res = value1 % value2
 924                         stack.append(res)
 925                     elif opcode == 208:  # getlocal_0
 926                         stack.append(registers[0])
 927                     elif opcode == 209:  # getlocal_1
 928                         stack.append(registers[1])
 929                     elif opcode == 210:  # getlocal_2
 930                         stack.append(registers[2])
 931                     elif opcode == 211:  # getlocal_3
 932                         stack.append(registers[3])
 933                     elif opcode == 214:  # setlocal_2
 934                         registers[2] = stack.pop()
 935                     elif opcode == 215:  # setlocal_3
 936                         registers[3] = stack.pop()
 937                     else:
 938                         raise NotImplementedError(
 939                             u'Unsupported opcode %d' % opcode)
 940
 941             method_pyfunctions[func_name] = resfunc
 942             return resfunc
 943
 944         initial_function = extract_function(u'decipher')
 945         return lambda s: initial_function([s])
 946
 947     def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
 948         """Turn the encrypted s field into a working signature"""
 949
 950         if player_url is not None:
 951             if player_url.startswith(u'//'):
 952                 player_url = u'https:' + player_url
 953             try:
 954                 player_id = (player_url, len(s))
 955                 if player_id not in self._player_cache:
 956                     func = self._extract_signature_function(
 957                         video_id, player_url, len(s)
 958                     )
 959                     self._player_cache[player_id] = func
 960                 func = self._player_cache[player_id]
 961                 if self._downloader.params.get('youtube_print_sig_code'):
 962                     self._print_sig_code(func, len(s))
 963                 return func(s)
 964             except Exception:
 965                 tb = traceback.format_exc()
 966                 self._downloader.report_warning(
 967                     u'Automatic signature extraction failed: ' + tb)
 968
 969             self._downloader.report_warning(
 970                 u'Warning: Falling back to static signature algorithm')
 971
 972         return self._static_decrypt_signature(
 973             s, video_id, player_url, age_gate)
 974
 975     def _static_decrypt_signature(self, s, video_id, player_url, age_gate):
 976         if age_gate:
 977             # The videos with age protection use another player, so the
 978             # algorithms can be different.
 979             if len(s) == 86:
 980                 return s[2:63] + s[82] + s[64:82] + s[63]
 981
 982         if len(s) == 93:
 983             return s[86:29:-1] + s[88] + s[28:5:-1]
 984         elif len(s) == 92:
 985             return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83]
 986         elif len(s) == 91:
 987             return s[84:27:-1] + s[86] + s[26:5:-1]
 988         elif len(s) == 90:
 989             return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81]
 990         elif len(s) == 89:
 991             return s[84:78:-1] + s[87] + s[77:60:-1] + s[0] + s[59:3:-1]
 992         elif len(s) == 88:
 993             return s[7:28] + s[87] + s[29:45] + s[55] + s[46:55] + s[2] + s[56:87] + s[28]
 994         elif len(s) == 87:
 995             return s[6:27] + s[4] + s[28:39] + s[27] + s[40:59] + s[2] + s[60:]
 996         elif len(s) == 86:
 997             return s[80:72:-1] + s[16] + s[71:39:-1] + s[72] + s[38:16:-1] + s[82] + s[15::-1]
 998         elif len(s) == 85:
 999             return s[3:11] + s[0] + s[12:55] + s[84] + s[56:84]
1000         elif len(s) == 84:
1001             return s[78:70:-1] + s[14] + s[69:37:-1] + s[70] + s[36:14:-1] + s[80] + s[:14][::-1]
1002         elif len(s) == 83:
1003             return s[80:63:-1] + s[0] + s[62:0:-1] + s[63]
1004         elif len(s) == 82:
1005             return s[80:37:-1] + s[7] + s[36:7:-1] + s[0] + s[6:0:-1] + s[37]
1006         elif len(s) == 81:
1007             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]
1008         elif len(s) == 80:
1009             return s[1:19] + s[0] + s[20:68] + s[19] + s[69:80]
1010         elif len(s) == 79:
1011             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]
1012
1013         else:
1014             raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s)))
1015
1016     def _get_available_subtitles(self, video_id, webpage):
1017         try:
1018             sub_list = self._download_webpage(
1019                 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
1020                 video_id, note=False)
1021         except ExtractorError as err:
1022             self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err))
1023             return {}
1024         lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
1025
1026         sub_lang_list = {}
1027         for l in lang_list:
1028             lang = l[1]
1029             params = compat_urllib_parse.urlencode({
1030                 'lang': lang,
1031                 'v': video_id,
1032                 'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
1033                 'name': unescapeHTML(l[0]).encode('utf-8'),
1034             })
1035             url = u'https://www.youtube.com/api/timedtext?' + params
1036             sub_lang_list[lang] = url
1037         if not sub_lang_list:
1038             self._downloader.report_warning(u'video doesn\'t have subtitles')
1039             return {}
1040         return sub_lang_list
1041
1042     def _get_available_automatic_caption(self, video_id, webpage):
1043         """We need the webpage for getting the captions url, pass it as an
1044            argument to speed up the process."""
1045         sub_format = self._downloader.params.get('subtitlesformat', 'srt')
1046         self.to_screen(u'%s: Looking for automatic captions' % video_id)
1047         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
1048         err_msg = u'Couldn\'t find automatic captions for %s' % video_id
1049         if mobj is None:
1050             self._downloader.report_warning(err_msg)
1051             return {}
1052         player_config = json.loads(mobj.group(1))
1053         try:
1054             args = player_config[u'args']
1055             caption_url = args[u'ttsurl']
1056             timestamp = args[u'timestamp']
1057             # We get the available subtitles
1058             list_params = compat_urllib_parse.urlencode({
1059                 'type': 'list',
1060                 'tlangs': 1,
1061                 'asrs': 1,
1062             })
1063             list_url = caption_url + '&' + list_params
1064             caption_list = self._download_xml(list_url, video_id)
1065             original_lang_node = caption_list.find('track')
1066             if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' :
1067                 self._downloader.report_warning(u'Video doesn\'t have automatic captions')
1068                 return {}
1069             original_lang = original_lang_node.attrib['lang_code']
1070
1071             sub_lang_list = {}
1072             for lang_node in caption_list.findall('target'):
1073                 sub_lang = lang_node.attrib['lang_code']
1074                 params = compat_urllib_parse.urlencode({
1075                     'lang': original_lang,
1076                     'tlang': sub_lang,
1077                     'fmt': sub_format,
1078                     'ts': timestamp,
1079                     'kind': 'asr',
1080                 })
1081                 sub_lang_list[sub_lang] = caption_url + '&' + params
1082             return sub_lang_list
1083         # An extractor error can be raise by the download process if there are
1084         # no automatic captions but there are subtitles
1085         except (KeyError, ExtractorError):
1086             self._downloader.report_warning(err_msg)
1087             return {}
1088
1089     @classmethod
1090     def extract_id(cls, url):
1091         mobj = re.match(cls._VALID_URL, url, re.VERBOSE)
1092         if mobj is None:
1093             raise ExtractorError(u'Invalid URL: %s' % url)
1094         video_id = mobj.group(2)
1095         return video_id
1096
1097     def _extract_from_m3u8(self, manifest_url, video_id):
1098         url_map = {}
1099         def _get_urls(_manifest):
1100             lines = _manifest.split('\n')
1101             urls = filter(lambda l: l and not l.startswith('#'),
1102                             lines)
1103             return urls
1104         manifest = self._download_webpage(manifest_url, video_id, u'Downloading formats manifest')
1105         formats_urls = _get_urls(manifest)
1106         for format_url in formats_urls:
1107             itag = self._search_regex(r'itag/(\d+?)/', format_url, 'itag')
1108             url_map[itag] = format_url
1109         return url_map
1110
1111     def _extract_annotations(self, video_id):
1112         url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
1113         return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.')
1114
1115     def _real_extract(self, url):
1116         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
1117         mobj = re.search(self._NEXT_URL_RE, url)
1118         if mobj:
1119             url = 'https://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
1120         video_id = self.extract_id(url)
1121
1122         # Get video webpage
1123         url = 'https://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
1124         video_webpage = self._download_webpage(url, video_id)
1125
1126         # Attempt to extract SWF player URL
1127         mobj = re.search(r'swfConfig.*?"(https?:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
1128         if mobj is not None:
1129             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
1130         else:
1131             player_url = None
1132
1133         # Get video info
1134         self.report_video_info_webpage_download(video_id)
1135         if re.search(r'player-age-gate-content">', video_webpage) is not None:
1136             self.report_age_confirmation()
1137             age_gate = True
1138             # We simulate the access to the video from www.youtube.com/v/{video_id}
1139             # this can be viewed without login into Youtube
1140             data = compat_urllib_parse.urlencode({'video_id': video_id,
1141                                                   'el': 'player_embedded',
1142                                                   'gl': 'US',
1143                                                   'hl': 'en',
1144                                                   'eurl': 'https://youtube.googleapis.com/v/' + video_id,
1145                                                   'asv': 3,
1146                                                   'sts':'1588',
1147                                                   })
1148             video_info_url = 'https://www.youtube.com/get_video_info?' + data
1149             video_info_webpage = self._download_webpage(video_info_url, video_id,
1150                                     note=False,
1151                                     errnote='unable to download video info webpage')
1152             video_info = compat_parse_qs(video_info_webpage)
1153         else:
1154             age_gate = False
1155             for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
1156                 video_info_url = ('https://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
1157                         % (video_id, el_type))
1158                 video_info_webpage = self._download_webpage(video_info_url, video_id,
1159                                         note=False,
1160                                         errnote='unable to download video info webpage')
1161                 video_info = compat_parse_qs(video_info_webpage)
1162                 if 'token' in video_info:
1163                     break
1164         if 'token' not in video_info:
1165             if 'reason' in video_info:
1166                 raise ExtractorError(u'YouTube said: %s' % video_info['reason'][0], expected=True)
1167             else:
1168                 raise ExtractorError(u'"token" parameter not in video info for unknown reason')
1169
1170         if 'view_count' in video_info:
1171             view_count = int(video_info['view_count'][0])
1172         else:
1173             view_count = None
1174
1175         # Check for "rental" videos
1176         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
1177             raise ExtractorError(u'"rental" videos not supported')
1178
1179         # Start extracting information
1180         self.report_information_extraction(video_id)
1181
1182         # uploader
1183         if 'author' not in video_info:
1184             raise ExtractorError(u'Unable to extract uploader name')
1185         video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
1186
1187         # uploader_id
1188         video_uploader_id = None
1189         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
1190         if mobj is not None:
1191             video_uploader_id = mobj.group(1)
1192         else:
1193             self._downloader.report_warning(u'unable to extract uploader nickname')
1194
1195         # title
1196         if 'title' in video_info:
1197             video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
1198         else:
1199             self._downloader.report_warning(u'Unable to extract video title')
1200             video_title = u'_'
1201
1202         # thumbnail image
1203         # We try first to get a high quality image:
1204         m_thumb = re.search(r'<span itemprop="thumbnail".*?href="(.*?)">',
1205                             video_webpage, re.DOTALL)
1206         if m_thumb is not None:
1207             video_thumbnail = m_thumb.group(1)
1208         elif 'thumbnail_url' not in video_info:
1209             self._downloader.report_warning(u'unable to extract video thumbnail')
1210             video_thumbnail = None
1211         else:   # don't panic if we can't find it
1212             video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
1213
1214         # upload date
1215         upload_date = None
1216         mobj = re.search(r'id="eow-date.*?>(.*?)</span>', video_webpage, re.DOTALL)
1217         if mobj is not None:
1218             upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
1219             upload_date = unified_strdate(upload_date)
1220
1221         # description
1222         video_description = get_element_by_id("eow-description", video_webpage)
1223         if video_description:
1224             video_description = re.sub(r'''(?x)
1225                 <a\s+
1226                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1227                     title="([^"]+)"\s+
1228                     (?:[a-zA-Z-]+="[^"]+"\s+)*?
1229                     class="yt-uix-redirect-link"\s*>
1230                 [^<]+
1231                 </a>
1232             ''', r'\1', video_description)
1233             video_description = clean_html(video_description)
1234         else:
1235             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
1236             if fd_mobj:
1237                 video_description = unescapeHTML(fd_mobj.group(1))
1238             else:
1239                 video_description = u''
1240
1241         def _extract_count(klass):
1242             count = self._search_regex(
1243                 r'class="%s">([\d,]+)</span>' % re.escape(klass),
1244                 video_webpage, klass, default=None)
1245             if count is not None:
1246                 return int(count.replace(',', ''))
1247             return None
1248         like_count = _extract_count(u'likes-count')
1249         dislike_count = _extract_count(u'dislikes-count')
1250
1251         # subtitles
1252         video_subtitles = self.extract_subtitles(video_id, video_webpage)
1253
1254         if self._downloader.params.get('listsubtitles', False):
1255             self._list_available_subtitles(video_id, video_webpage)
1256             return
1257
1258         if 'length_seconds' not in video_info:
1259             self._downloader.report_warning(u'unable to extract video duration')
1260             video_duration = None
1261         else:
1262             video_duration = int(compat_urllib_parse.unquote_plus(video_info['length_seconds'][0]))
1263
1264         # annotations
1265         video_annotations = None
1266         if self._downloader.params.get('writeannotations', False):
1267                 video_annotations = self._extract_annotations(video_id)
1268
1269         # Decide which formats to download
1270         try:
1271             mobj = re.search(r';ytplayer.config = ({.*?});', video_webpage)
1272             if not mobj:
1273                 raise ValueError('Could not find vevo ID')
1274             info = json.loads(mobj.group(1))
1275             args = info['args']
1276             # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
1277             # this signatures are encrypted
1278             if 'url_encoded_fmt_stream_map' not in args:
1279                 raise ValueError(u'No stream_map present')  # caught below
1280             re_signature = re.compile(r'[&,]s=')
1281             m_s = re_signature.search(args['url_encoded_fmt_stream_map'])
1282             if m_s is not None:
1283                 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
1284                 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
1285             m_s = re_signature.search(args.get('adaptive_fmts', u''))
1286             if m_s is not None:
1287                 if 'adaptive_fmts' in video_info:
1288                     video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts']
1289                 else:
1290                     video_info['adaptive_fmts'] = [args['adaptive_fmts']]
1291         except ValueError:
1292             pass
1293
1294         def _map_to_format_list(urlmap):
1295             formats = []
1296             for itag, video_real_url in urlmap.items():
1297                 dct = {
1298                     'format_id': itag,
1299                     'url': video_real_url,
1300                     'player_url': player_url,
1301                 }
1302                 if itag in self._formats:
1303                     dct.update(self._formats[itag])
1304                 formats.append(dct)
1305             return formats
1306
1307         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
1308             self.report_rtmp_download()
1309             formats = [{
1310                 'format_id': '_rtmp',
1311                 'protocol': 'rtmp',
1312                 'url': video_info['conn'][0],
1313                 'player_url': player_url,
1314             }]
1315         elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1:
1316             encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0]
1317             if 'rtmpe%3Dyes' in encoded_url_map:
1318                 raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
1319             url_map = {}
1320             for url_data_str in encoded_url_map.split(','):
1321                 url_data = compat_parse_qs(url_data_str)
1322                 if 'itag' in url_data and 'url' in url_data:
1323                     url = url_data['url'][0]
1324                     if 'sig' in url_data:
1325                         url += '&signature=' + url_data['sig'][0]
1326                     elif 's' in url_data:
1327                         encrypted_sig = url_data['s'][0]
1328                         if self._downloader.params.get('verbose'):
1329                             if age_gate:
1330                                 if player_url is None:
1331                                     player_version = 'unknown'
1332                                 else:
1333                                     player_version = self._search_regex(
1334                                         r'-(.+)\.swf$', player_url,
1335                                         u'flash player', fatal=False)
1336                                 player_desc = 'flash player %s' % player_version
1337                             else:
1338                                 player_version = self._search_regex(
1339                                     r'html5player-(.+?)\.js', video_webpage,
1340                                     'html5 player', fatal=False)
1341                                 player_desc = u'html5 player %s' % player_version
1342
1343                             parts_sizes = u'.'.join(compat_str(len(part)) for part in encrypted_sig.split('.'))
1344                             self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' %
1345                                 (len(encrypted_sig), parts_sizes, url_data['itag'][0], player_desc))
1346
1347                         if not age_gate:
1348                             jsplayer_url_json = self._search_regex(
1349                                 r'"assets":.+?"js":\s*("[^"]+")',
1350                                 video_webpage, u'JS player URL')
1351                             player_url = json.loads(jsplayer_url_json)
1352
1353                         signature = self._decrypt_signature(
1354                             encrypted_sig, video_id, player_url, age_gate)
1355                         url += '&signature=' + signature
1356                     if 'ratebypass' not in url:
1357                         url += '&ratebypass=yes'
1358                     url_map[url_data['itag'][0]] = url
1359             formats = _map_to_format_list(url_map)
1360         elif video_info.get('hlsvp'):
1361             manifest_url = video_info['hlsvp'][0]
1362             url_map = self._extract_from_m3u8(manifest_url, video_id)
1363             formats = _map_to_format_list(url_map)
1364         else:
1365             raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
1366
1367         # Look for the DASH manifest
1368         dash_manifest_url_lst = video_info.get('dashmpd')
1369         if (dash_manifest_url_lst and dash_manifest_url_lst[0] and
1370                 self._downloader.params.get('youtube_include_dash_manifest', False)):
1371             try:
1372                 dash_doc = self._download_xml(
1373                     dash_manifest_url_lst[0], video_id,
1374                     note=u'Downloading DASH manifest',
1375                     errnote=u'Could not download DASH manifest')
1376                 for r in dash_doc.findall(u'.//{urn:mpeg:DASH:schema:MPD:2011}Representation'):
1377                     url_el = r.find('{urn:mpeg:DASH:schema:MPD:2011}BaseURL')
1378                     if url_el is None:
1379                         continue
1380                     format_id = r.attrib['id']
1381                     video_url = url_el.text
1382                     filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength'))
1383                     f = {
1384                         'format_id': format_id,
1385                         'url': video_url,
1386                         'width': int_or_none(r.attrib.get('width')),
1387                         'tbr': int_or_none(r.attrib.get('bandwidth'), 1000),
1388                         'asr': int_or_none(r.attrib.get('audioSamplingRate')),
1389                         'filesize': filesize,
1390                     }
1391                     try:
1392                         existing_format = next(
1393                             fo for fo in formats
1394                             if fo['format_id'] == format_id)
1395                     except StopIteration:
1396                         f.update(self._formats.get(format_id, {}))
1397                         formats.append(f)
1398                     else:
1399                         existing_format.update(f)
1400
1401             except (ExtractorError, KeyError) as e:
1402                 self.report_warning(u'Skipping DASH manifest: %s' % e, video_id)
1403
1404         self._sort_formats(formats)
1405
1406         return {
1407             'id':           video_id,
1408             'uploader':     video_uploader,
1409             'uploader_id':  video_uploader_id,
1410             'upload_date':  upload_date,
1411             'title':        video_title,
1412             'thumbnail':    video_thumbnail,
1413             'description':  video_description,
1414             'subtitles':    video_subtitles,
1415             'duration':     video_duration,
1416             'age_limit':    18 if age_gate else 0,
1417             'annotations':  video_annotations,
1418             'webpage_url': 'https://www.youtube.com/watch?v=%s' % video_id,
1419             'view_count':   view_count,
1420             'like_count': like_count,
1421             'dislike_count': dislike_count,
1422             'formats':      formats,
1423         }
1424
1425 class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
1426     IE_DESC = u'YouTube.com playlists'
1427     _VALID_URL = r"""(?x)(?:
1428                         (?:https?://)?
1429                         (?:\w+\.)?
1430                         youtube\.com/
1431                         (?:
1432                            (?:course|view_play_list|my_playlists|artist|playlist|watch)
1433                            \? (?:.*?&)*? (?:p|a|list)=
1434                         |  p/
1435                         )
1436                         (
1437                             (?:PL|EC|UU|FL|RD)?[0-9A-Za-z-_]{10,}
1438                             # Top tracks, they can also include dots
1439                             |(?:MC)[\w\.]*
1440                         )
1441                         .*
1442                      |
1443                         ((?:PL|EC|UU|FL|RD)[0-9A-Za-z-_]{10,})
1444                      )"""
1445     _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s&page=%s'
1446     _MORE_PAGES_INDICATOR = r'data-link-type="next"'
1447     _VIDEO_RE = r'href="/watch\?v=(?P<id>[0-9A-Za-z_-]{11})&amp;[^"]*?index=(?P<index>\d+)'
1448     IE_NAME = u'youtube:playlist'
1449
1450     def _real_initialize(self):
1451         self._login()
1452
1453     def _ids_to_results(self, ids):
1454         return [self.url_result(vid_id, 'Youtube', video_id=vid_id)
1455                        for vid_id in ids]
1456
1457     def _extract_mix(self, playlist_id):
1458         # The mixes are generated from a a single video
1459         # the id of the playlist is just 'RD' + video_id
1460         url = 'https://youtube.com/watch?v=%s&list=%s' % (playlist_id[-11:], playlist_id)
1461         webpage = self._download_webpage(url, playlist_id, u'Downloading Youtube mix')
1462         title_span = (get_element_by_attribute('class', 'title long-title', webpage) or
1463             get_element_by_attribute('class', 'title ', webpage))
1464         title = clean_html(title_span)
1465         video_re = r'data-index="\d+".*?href="/watch\?v=([0-9A-Za-z_-]{11})&amp;[^"]*?list=%s' % re.escape(playlist_id)
1466         ids = orderedSet(re.findall(video_re, webpage))
1467         url_results = self._ids_to_results(ids)
1468
1469         return self.playlist_result(url_results, playlist_id, title)
1470
1471     def _real_extract(self, url):
1472         # Extract playlist id
1473         mobj = re.match(self._VALID_URL, url)
1474         if mobj is None:
1475             raise ExtractorError(u'Invalid URL: %s' % url)
1476         playlist_id = mobj.group(1) or mobj.group(2)
1477
1478         # Check if it's a video-specific URL
1479         query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
1480         if 'v' in query_dict:
1481             video_id = query_dict['v'][0]
1482             if self._downloader.params.get('noplaylist'):
1483                 self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
1484                 return self.url_result(video_id, 'Youtube', video_id=video_id)
1485             else:
1486                 self.to_screen(u'Downloading playlist PL%s - add --no-playlist to just download video %s' % (playlist_id, video_id))
1487
1488         if playlist_id.startswith('RD'):
1489             # Mixes require a custom extraction process
1490             return self._extract_mix(playlist_id)
1491         if playlist_id.startswith('TL'):
1492             raise ExtractorError(u'For downloading YouTube.com top lists, use '
1493                 u'the "yttoplist" keyword, for example "youtube-dl \'yttoplist:music:Top Tracks\'"', expected=True)
1494
1495         # Extract the video ids from the playlist pages
1496         ids = []
1497
1498         for page_num in itertools.count(1):
1499             url = self._TEMPLATE_URL % (playlist_id, page_num)
1500             page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
1501             matches = re.finditer(self._VIDEO_RE, page)
1502             # We remove the duplicates and the link with index 0
1503             # (it's not the first video of the playlist)
1504             new_ids = orderedSet(m.group('id') for m in matches if m.group('index') != '0')
1505             ids.extend(new_ids)
1506
1507             if re.search(self._MORE_PAGES_INDICATOR, page) is None:
1508                 break
1509
1510         try:
1511             playlist_title = self._og_search_title(page)
1512         except RegexNotFoundError:
1513             self.report_warning(
1514                 u'Playlist page is missing OpenGraph title, falling back ...',
1515                 playlist_id)
1516             playlist_title = self._html_search_regex(
1517                 r'<h1 class="pl-header-title">(.*?)</h1>', page, u'title')
1518
1519         url_results = self._ids_to_results(ids)
1520         return self.playlist_result(url_results, playlist_id, playlist_title)
1521
1522
1523 class YoutubeTopListIE(YoutubePlaylistIE):
1524     IE_NAME = u'youtube:toplist'
1525     IE_DESC = (u'YouTube.com top lists, "yttoplist:{channel}:{list title}"'
1526         u' (Example: "yttoplist:music:Top Tracks")')
1527     _VALID_URL = r'yttoplist:(?P<chann>.*?):(?P<title>.*?)$'
1528
1529     def _real_extract(self, url):
1530         mobj = re.match(self._VALID_URL, url)
1531         channel = mobj.group('chann')
1532         title = mobj.group('title')
1533         query = compat_urllib_parse.urlencode({'title': title})
1534         playlist_re = 'href="([^"]+?%s.*?)"' % re.escape(query)
1535         channel_page = self._download_webpage('https://www.youtube.com/%s' % channel, title)
1536         link = self._html_search_regex(playlist_re, channel_page, u'list')
1537         url = compat_urlparse.urljoin('https://www.youtube.com/', link)
1538
1539         video_re = r'data-index="\d+".*?data-video-id="([0-9A-Za-z_-]{11})"'
1540         ids = []
1541         # sometimes the webpage doesn't contain the videos
1542         # retry until we get them
1543         for i in itertools.count(0):
1544             msg = u'Downloading Youtube mix'
1545             if i > 0:
1546                 msg += ', retry #%d' % i
1547             webpage = self._download_webpage(url, title, msg)
1548             ids = orderedSet(re.findall(video_re, webpage))
1549             if ids:
1550                 break
1551         url_results = self._ids_to_results(ids)
1552         return self.playlist_result(url_results, playlist_title=title)
1553
1554
1555 class YoutubeChannelIE(InfoExtractor):
1556     IE_DESC = u'YouTube.com channels'
1557     _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
1558     _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
1559     _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'
1560     IE_NAME = u'youtube:channel'
1561
1562     def extract_videos_from_page(self, page):
1563         ids_in_page = []
1564         for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
1565             if mobj.group(1) not in ids_in_page:
1566                 ids_in_page.append(mobj.group(1))
1567         return ids_in_page
1568
1569     def _real_extract(self, url):
1570         # Extract channel id
1571         mobj = re.match(self._VALID_URL, url)
1572         if mobj is None:
1573             raise ExtractorError(u'Invalid URL: %s' % url)
1574
1575         # Download channel page
1576         channel_id = mobj.group(1)
1577         video_ids = []
1578         url = 'https://www.youtube.com/channel/%s/videos' % channel_id
1579         channel_page = self._download_webpage(url, channel_id)
1580         autogenerated = re.search(r'''(?x)
1581                 class="[^"]*?(?:
1582                     channel-header-autogenerated-label|
1583                     yt-channel-title-autogenerated
1584                 )[^"]*"''', channel_page) is not None
1585
1586         if autogenerated:
1587             # The videos are contained in a single page
1588             # the ajax pages can't be used, they are empty
1589             video_ids = self.extract_videos_from_page(channel_page)
1590         else:
1591             # Download all channel pages using the json-based channel_ajax query
1592             for pagenum in itertools.count(1):
1593                 url = self._MORE_PAGES_URL % (pagenum, channel_id)
1594                 page = self._download_json(
1595                     url, channel_id, note=u'Downloading page #%s' % pagenum,
1596                     transform_source=uppercase_escape)
1597
1598                 ids_in_page = self.extract_videos_from_page(page['content_html'])
1599                 video_ids.extend(ids_in_page)
1600
1601                 if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
1602                     break
1603
1604         self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
1605
1606         url_entries = [self.url_result(video_id, 'Youtube', video_id=video_id)
1607                        for video_id in video_ids]
1608         return self.playlist_result(url_entries, channel_id)
1609
1610
1611 class YoutubeUserIE(InfoExtractor):
1612     IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
1613     _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
1614     _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/users/%s'
1615     _GDATA_PAGE_SIZE = 50
1616     _GDATA_URL = 'https://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
1617     IE_NAME = u'youtube:user'
1618
1619     @classmethod
1620     def suitable(cls, url):
1621         # Don't return True if the url can be extracted with other youtube
1622         # extractor, the regex would is too permissive and it would match.
1623         other_ies = iter(klass for (name, klass) in globals().items() if name.endswith('IE') and klass is not cls)
1624         if any(ie.suitable(url) for ie in other_ies): return False
1625         else: return super(YoutubeUserIE, cls).suitable(url)
1626
1627     def _real_extract(self, url):
1628         # Extract username
1629         mobj = re.match(self._VALID_URL, url)
1630         if mobj is None:
1631             raise ExtractorError(u'Invalid URL: %s' % url)
1632
1633         username = mobj.group(1)
1634
1635         # Download video ids using YouTube Data API. Result size per
1636         # query is limited (currently to 50 videos) so we need to query
1637         # page by page until there are no video ids - it means we got
1638         # all of them.
1639
1640         def download_page(pagenum):
1641             start_index = pagenum * self._GDATA_PAGE_SIZE + 1
1642
1643             gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
1644             page = self._download_webpage(
1645                 gdata_url, username,
1646                 u'Downloading video ids from %d to %d' % (
1647                     start_index, start_index + self._GDATA_PAGE_SIZE))
1648
1649             try:
1650                 response = json.loads(page)
1651             except ValueError as err:
1652                 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
1653             if 'entry' not in response['feed']:
1654                 return
1655
1656             # Extract video identifiers
1657             entries = response['feed']['entry']
1658             for entry in entries:
1659                 title = entry['title']['$t']
1660                 video_id = entry['id']['$t'].split('/')[-1]
1661                 yield {
1662                     '_type': 'url',
1663                     'url': video_id,
1664                     'ie_key': 'Youtube',
1665                     'id': video_id,
1666                     'title': title,
1667                 }
1668         url_results = PagedList(download_page, self._GDATA_PAGE_SIZE)
1669
1670         return self.playlist_result(url_results, playlist_title=username)
1671
1672
1673 class YoutubeSearchIE(SearchInfoExtractor):
1674     IE_DESC = u'YouTube.com searches'
1675     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
1676     _MAX_RESULTS = 1000
1677     IE_NAME = u'youtube:search'
1678     _SEARCH_KEY = 'ytsearch'
1679
1680     def _get_n_results(self, query, n):
1681         """Get a specified number of results for a query"""
1682
1683         video_ids = []
1684         pagenum = 0
1685         limit = n
1686
1687         while (50 * pagenum) < limit:
1688             result_url = self._API_URL % (compat_urllib_parse.quote_plus(query), (50*pagenum)+1)
1689             data_json = self._download_webpage(
1690                 result_url, video_id=u'query "%s"' % query,
1691                 note=u'Downloading page %s' % (pagenum + 1),
1692                 errnote=u'Unable to download API page')
1693             data = json.loads(data_json)
1694             api_response = data['data']
1695
1696             if 'items' not in api_response:
1697                 raise ExtractorError(
1698                     u'[youtube] No video results', expected=True)
1699
1700             new_ids = list(video['id'] for video in api_response['items'])
1701             video_ids += new_ids
1702
1703             limit = min(n, api_response['totalItems'])
1704             pagenum += 1
1705
1706         if len(video_ids) > n:
1707             video_ids = video_ids[:n]
1708         videos = [self.url_result(video_id, 'Youtube', video_id=video_id)
1709                   for video_id in video_ids]
1710         return self.playlist_result(videos, query)
1711
1712 class YoutubeSearchDateIE(YoutubeSearchIE):
1713     IE_NAME = YoutubeSearchIE.IE_NAME + ':date'
1714     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published'
1715     _SEARCH_KEY = 'ytsearchdate'
1716     IE_DESC = u'YouTube.com searches, newest videos first'
1717
1718 class YoutubeShowIE(InfoExtractor):
1719     IE_DESC = u'YouTube.com (multi-season) shows'
1720     _VALID_URL = r'https?://www\.youtube\.com/show/(.*)'
1721     IE_NAME = u'youtube:show'
1722
1723     def _real_extract(self, url):
1724         mobj = re.match(self._VALID_URL, url)
1725         show_name = mobj.group(1)
1726         webpage = self._download_webpage(url, show_name, u'Downloading show webpage')
1727         # There's one playlist for each season of the show
1728         m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage))
1729         self.to_screen(u'%s: Found %s seasons' % (show_name, len(m_seasons)))
1730         return [self.url_result('https://www.youtube.com' + season.group(1), 'YoutubePlaylist') for season in m_seasons]
1731
1732
1733 class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
1734     """
1735     Base class for extractors that fetch info from
1736     http://www.youtube.com/feed_ajax
1737     Subclasses must define the _FEED_NAME and _PLAYLIST_TITLE properties.
1738     """
1739     _LOGIN_REQUIRED = True
1740     # use action_load_personal_feed instead of action_load_system_feed
1741     _PERSONAL_FEED = False
1742
1743     @property
1744     def _FEED_TEMPLATE(self):
1745         action = 'action_load_system_feed'
1746         if self._PERSONAL_FEED:
1747             action = 'action_load_personal_feed'
1748         return 'https://www.youtube.com/feed_ajax?%s=1&feed_name=%s&paging=%%s' % (action, self._FEED_NAME)
1749
1750     @property
1751     def IE_NAME(self):
1752         return u'youtube:%s' % self._FEED_NAME
1753
1754     def _real_initialize(self):
1755         self._login()
1756
1757     def _real_extract(self, url):
1758         feed_entries = []
1759         paging = 0
1760         for i in itertools.count(1):
1761             info = self._download_webpage(self._FEED_TEMPLATE % paging,
1762                                           u'%s feed' % self._FEED_NAME,
1763                                           u'Downloading page %s' % i)
1764             info = json.loads(info)
1765             feed_html = info['feed_html']
1766             m_ids = re.finditer(r'"/watch\?v=(.*?)["&]', feed_html)
1767             ids = orderedSet(m.group(1) for m in m_ids)
1768             feed_entries.extend(
1769                 self.url_result(video_id, 'Youtube', video_id=video_id)
1770                 for video_id in ids)
1771             if info['paging'] is None:
1772                 break
1773             paging = info['paging']
1774         return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
1775
1776 class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
1777     IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword(requires authentication)'
1778     _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
1779     _FEED_NAME = 'subscriptions'
1780     _PLAYLIST_TITLE = u'Youtube Subscriptions'
1781
1782 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
1783     IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
1784     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
1785     _FEED_NAME = 'recommended'
1786     _PLAYLIST_TITLE = u'Youtube Recommended videos'
1787
1788 class YoutubeWatchLaterIE(YoutubeFeedsInfoExtractor):
1789     IE_DESC = u'Youtube watch later list, "ytwatchlater" keyword (requires authentication)'
1790     _VALID_URL = r'https?://www\.youtube\.com/feed/watch_later|:ytwatchlater'
1791     _FEED_NAME = 'watch_later'
1792     _PLAYLIST_TITLE = u'Youtube Watch Later'
1793     _PERSONAL_FEED = True
1794
1795 class YoutubeHistoryIE(YoutubeFeedsInfoExtractor):
1796     IE_DESC = u'Youtube watch history, "ythistory" keyword (requires authentication)'
1797     _VALID_URL = u'https?://www\.youtube\.com/feed/history|:ythistory'
1798     _FEED_NAME = 'history'
1799     _PERSONAL_FEED = True
1800     _PLAYLIST_TITLE = u'Youtube Watch History'
1801
1802 class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
1803     IE_NAME = u'youtube:favorites'
1804     IE_DESC = u'YouTube.com favourite videos, "ytfav" keyword (requires authentication)'
1805     _VALID_URL = r'https?://www\.youtube\.com/my_favorites|:ytfav(?:ou?rites)?'
1806     _LOGIN_REQUIRED = True
1807
1808     def _real_extract(self, url):
1809         webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
1810         playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
1811         return self.url_result(playlist_id, 'YoutubePlaylist')
1812
1813
1814 class YoutubeTruncatedURLIE(InfoExtractor):
1815     IE_NAME = 'youtube:truncated_url'
1816     IE_DESC = False  # Do not list
1817     _VALID_URL = r'''(?x)
1818         (?:https?://)?[^/]+/watch\?feature=[a-z_]+$|
1819         (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$
1820     '''
1821
1822     def _real_extract(self, url):
1823         raise ExtractorError(
1824             u'Did you forget to quote the URL? Remember that & is a meta '
1825             u'character in most shells, so you want to put the URL in quotes, '
1826             u'like  youtube-dl '
1827             u'"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
1828             u' or simply  youtube-dl BaW_jenozKc  .',
1829             expected=True)