_ Git - youtube-dl/blob - youtube_dl/extractor/youtube.py

   1 # coding: utf-8
   2
   3 import json
   4 import netrc
   5 import re
   6 import socket
   7
   8 from .common import InfoExtractor, SearchInfoExtractor
   9 from ..utils import (
  10     compat_http_client,
  11     compat_parse_qs,
  12     compat_urllib_error,
  13     compat_urllib_parse,
  14     compat_urllib_request,
  15     compat_str,
  16
  17     clean_html,
  18     get_element_by_id,
  19     ExtractorError,
  20     unescapeHTML,
  21     unified_strdate,
  22 )
  23
  24
  25 class YoutubeIE(InfoExtractor):
  26     """Information extractor for youtube.com."""
  27
  28     _VALID_URL = r"""^
  29                      (
  30                          (?:https?://)?                                       # http(s):// (optional)
  31                          (?:youtu\.be/|(?:\w+\.)?youtube(?:-nocookie)?\.com/|
  32                             tube\.majestyc\.net/)                             # the various hostnames, with wildcard subdomains
  33                          (?:.*?\#/)?                                          # handle anchor (#/) redirect urls
  34                          (?:                                                  # the various things that can precede the ID:
  35                              (?:(?:v|embed|e)/)                               # v/ or embed/ or e/
  36                              |(?:                                             # or the v= param in all its forms
  37                                  (?:watch(?:_popup)?(?:\.php)?)?              # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
  38                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
  39                                  (?:.*?&)?                                    # any other preceding param (like /?s=tuff&v=xxxx)
  40                                  v=
  41                              )
  42                          )?                                                   # optional -> youtube.com/xxxx is OK
  43                      )?                                                       # all until now is optional -> you can pass the naked ID
  44                      ([0-9A-Za-z_-]+)                                         # here is it! the YouTube video ID
  45                      (?(1).+)?                                                # if we found the ID, everything can follow
  46                      $"""
  47     _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1'
  48     _LOGIN_URL = 'https://accounts.google.com/ServiceLogin'
  49     _AGE_URL = 'http://www.youtube.com/verify_age?next_url=/&gl=US&hl=en'
  50     _NEXT_URL_RE = r'[\?&]next_url=([^&]+)'
  51     _NETRC_MACHINE = 'youtube'
  52     # Listed in order of quality
  53     _available_formats = ['38', '37', '46', '22', '45', '35', '44', '34', '18', '43', '6', '5', '17', '13']
  54     _available_formats_prefer_free = ['38', '46', '37', '45', '22', '44', '35', '43', '34', '18', '6', '5', '17', '13']
  55     _video_extensions = {
  56         '13': '3gp',
  57         '17': 'mp4',
  58         '18': 'mp4',
  59         '22': 'mp4',
  60         '37': 'mp4',
  61         '38': 'mp4',
  62         '43': 'webm',
  63         '44': 'webm',
  64         '45': 'webm',
  65         '46': 'webm',
  66     }
  67     _video_dimensions = {
  68         '5': '240x400',
  69         '6': '???',
  70         '13': '???',
  71         '17': '144x176',
  72         '18': '360x640',
  73         '22': '720x1280',
  74         '34': '360x640',
  75         '35': '480x854',
  76         '37': '1080x1920',
  77         '38': '3072x4096',
  78         '43': '360x640',
  79         '44': '480x854',
  80         '45': '720x1280',
  81         '46': '1080x1920',
  82     }
  83     IE_NAME = u'youtube'
  84
  85     @classmethod
  86     def suitable(cls, url):
  87         """Receives a URL and returns True if suitable for this IE."""
  88         if YoutubePlaylistIE.suitable(url): return False
  89         return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
  90
  91     def report_lang(self):
  92         """Report attempt to set language."""
  93         self.to_screen(u'Setting language')
  94
  95     def report_login(self):
  96         """Report attempt to log in."""
  97         self.to_screen(u'Logging in')
  98
  99     def report_video_webpage_download(self, video_id):
 100         """Report attempt to download video webpage."""
 101         self.to_screen(u'%s: Downloading video webpage' % video_id)
 102
 103     def report_video_info_webpage_download(self, video_id):
 104         """Report attempt to download video info webpage."""
 105         self.to_screen(u'%s: Downloading video info webpage' % video_id)
 106
 107     def report_video_subtitles_download(self, video_id):
 108         """Report attempt to download video info webpage."""
 109         self.to_screen(u'%s: Checking available subtitles' % video_id)
 110
 111     def report_video_subtitles_request(self, video_id, sub_lang, format):
 112         """Report attempt to download video info webpage."""
 113         self.to_screen(u'%s: Downloading video subtitles for %s.%s' % (video_id, sub_lang, format))
 114
 115     def report_video_subtitles_available(self, video_id, sub_lang_list):
 116         """Report available subtitles."""
 117         sub_lang = ",".join(list(sub_lang_list.keys()))
 118         self.to_screen(u'%s: Available subtitles for video: %s' % (video_id, sub_lang))
 119
 120     def report_information_extraction(self, video_id):
 121         """Report attempt to extract video information."""
 122         self.to_screen(u'%s: Extracting video information' % video_id)
 123
 124     def report_unavailable_format(self, video_id, format):
 125         """Report extracted video URL."""
 126         self.to_screen(u'%s: Format %s not available' % (video_id, format))
 127
 128     def report_rtmp_download(self):
 129         """Indicate the download will use the RTMP protocol."""
 130         self.to_screen(u'RTMP download detected')
 131
 132     def _decrypt_signature(self, s):
 133         """Decrypt the key the two subkeys must have a length of 43"""
 134         if self._downloader.params.get('verbose'):
 135             self.to_screen('encrypted signature length %d' % (len(s)))
 136
 137         if len(s) == 88:
 138             return s[48] + s[81:67:-1] + s[82] + s[66:62:-1] + s[85] + s[61:48:-1] + s[67] + s[47:12:-1] + s[3] + s[11:3:-1] + s[2] + s[12]
 139         elif len(s) == 87:
 140             return s[62] + s[82:62:-1] + s[83] + s[61:52:-1] + s[0] + s[51:2:-1]
 141         elif len(s) == 86:
 142             return s[2:17] + s[0] + s[18:41] + s[79] + s[42:79] + s[82] + s[80:82] + s[41]
 143         elif len(s) == 85:
 144             return s[76] + s[82:76:-1] + s[83] + s[75:60:-1] + s[0] + s[59:50:-1] + s[1] + s[49:2:-1]
 145         elif len(s) == 84:
 146             return s[83:36:-1] + s[2] + s[35:26:-1] + s[3] + s[25:3:-1] + s[26]
 147         elif len(s) == 83:
 148             return s[52] + s[81:55:-1] + s[2] + s[54:52:-1] + s[82] + s[51:36:-1] + s[55] + s[35:2:-1] + s[36]
 149         elif len(s) == 82:
 150             return s[36] + s[79:67:-1] + s[81] + s[66:40:-1] + s[33] + s[39:36:-1] + s[40] + s[35] + s[0] + s[67] + s[32:0:-1] + s[34]
 151
 152         else:
 153             raise ExtractorError(u'Unable to decrypt signature, subkeys length %d not supported; retrying might work' % (len(s)))
 154
 155     def _get_available_subtitles(self, video_id):
 156         self.report_video_subtitles_download(video_id)
 157         request = compat_urllib_request.Request('http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id)
 158         try:
 159             sub_list = compat_urllib_request.urlopen(request).read().decode('utf-8')
 160         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 161             return (u'unable to download video subtitles: %s' % compat_str(err), None)
 162         sub_lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
 163         sub_lang_list = dict((l[1], l[0]) for l in sub_lang_list)
 164         if not sub_lang_list:
 165             return (u'video doesn\'t have subtitles', None)
 166         return sub_lang_list
 167
 168     def _list_available_subtitles(self, video_id):
 169         sub_lang_list = self._get_available_subtitles(video_id)
 170         self.report_video_subtitles_available(video_id, sub_lang_list)
 171
 172     def _request_subtitle(self, sub_lang, sub_name, video_id, format):
 173         """
 174         Return tuple:
 175         (error_message, sub_lang, sub)
 176         """
 177         self.report_video_subtitles_request(video_id, sub_lang, format)
 178         params = compat_urllib_parse.urlencode({
 179             'lang': sub_lang,
 180             'name': sub_name,
 181             'v': video_id,
 182             'fmt': format,
 183         })
 184         url = 'http://www.youtube.com/api/timedtext?' + params
 185         try:
 186             sub = compat_urllib_request.urlopen(url).read().decode('utf-8')
 187         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 188             return (u'unable to download video subtitles: %s' % compat_str(err), None, None)
 189         if not sub:
 190             return (u'Did not fetch video subtitles', None, None)
 191         return (None, sub_lang, sub)
 192
 193     def _request_automatic_caption(self, video_id, webpage):
 194         """We need the webpage for getting the captions url, pass it as an
 195            argument to speed up the process."""
 196         sub_lang = self._downloader.params.get('subtitleslang') or 'en'
 197         sub_format = self._downloader.params.get('subtitlesformat')
 198         self.to_screen(u'%s: Looking for automatic captions' % video_id)
 199         mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
 200         err_msg = u'Couldn\'t find automatic captions for "%s"' % sub_lang
 201         if mobj is None:
 202             return [(err_msg, None, None)]
 203         player_config = json.loads(mobj.group(1))
 204         try:
 205             args = player_config[u'args']
 206             caption_url = args[u'ttsurl']
 207             timestamp = args[u'timestamp']
 208             params = compat_urllib_parse.urlencode({
 209                 'lang': 'en',
 210                 'tlang': sub_lang,
 211                 'fmt': sub_format,
 212                 'ts': timestamp,
 213                 'kind': 'asr',
 214             })
 215             subtitles_url = caption_url + '&' + params
 216             sub = self._download_webpage(subtitles_url, video_id, u'Downloading automatic captions')
 217             return [(None, sub_lang, sub)]
 218         except KeyError:
 219             return [(err_msg, None, None)]
 220
 221     def _extract_subtitle(self, video_id):
 222         """
 223         Return a list with a tuple:
 224         [(error_message, sub_lang, sub)]
 225         """
 226         sub_lang_list = self._get_available_subtitles(video_id)
 227         sub_format = self._downloader.params.get('subtitlesformat')
 228         if  isinstance(sub_lang_list,tuple): #There was some error, it didn't get the available subtitles
 229             return [(sub_lang_list[0], None, None)]
 230         if self._downloader.params.get('subtitleslang', False):
 231             sub_lang = self._downloader.params.get('subtitleslang')
 232         elif 'en' in sub_lang_list:
 233             sub_lang = 'en'
 234         else:
 235             sub_lang = list(sub_lang_list.keys())[0]
 236         if not sub_lang in sub_lang_list:
 237             return [(u'no closed captions found in the specified language "%s"' % sub_lang, None, None)]
 238
 239         subtitle = self._request_subtitle(sub_lang, sub_lang_list[sub_lang].encode('utf-8'), video_id, sub_format)
 240         return [subtitle]
 241
 242     def _extract_all_subtitles(self, video_id):
 243         sub_lang_list = self._get_available_subtitles(video_id)
 244         sub_format = self._downloader.params.get('subtitlesformat')
 245         if  isinstance(sub_lang_list,tuple): #There was some error, it didn't get the available subtitles
 246             return [(sub_lang_list[0], None, None)]
 247         subtitles = []
 248         for sub_lang in sub_lang_list:
 249             subtitle = self._request_subtitle(sub_lang, sub_lang_list[sub_lang].encode('utf-8'), video_id, sub_format)
 250             subtitles.append(subtitle)
 251         return subtitles
 252
 253     def _print_formats(self, formats):
 254         print('Available formats:')
 255         for x in formats:
 256             print('%s\t:\t%s\t[%s]' %(x, self._video_extensions.get(x, 'flv'), self._video_dimensions.get(x, '???')))
 257
 258     def _real_initialize(self):
 259         if self._downloader is None:
 260             return
 261
 262         username = None
 263         password = None
 264         downloader_params = self._downloader.params
 265
 266         # Attempt to use provided username and password or .netrc data
 267         if downloader_params.get('username', None) is not None:
 268             username = downloader_params['username']
 269             password = downloader_params['password']
 270         elif downloader_params.get('usenetrc', False):
 271             try:
 272                 info = netrc.netrc().authenticators(self._NETRC_MACHINE)
 273                 if info is not None:
 274                     username = info[0]
 275                     password = info[2]
 276                 else:
 277                     raise netrc.NetrcParseError('No authenticators for %s' % self._NETRC_MACHINE)
 278             except (IOError, netrc.NetrcParseError) as err:
 279                 self._downloader.report_warning(u'parsing .netrc: %s' % compat_str(err))
 280                 return
 281
 282         # Set language
 283         request = compat_urllib_request.Request(self._LANG_URL)
 284         try:
 285             self.report_lang()
 286             compat_urllib_request.urlopen(request).read()
 287         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 288             self._downloader.report_warning(u'unable to set language: %s' % compat_str(err))
 289             return
 290
 291         # No authentication to be performed
 292         if username is None:
 293             return
 294
 295         request = compat_urllib_request.Request(self._LOGIN_URL)
 296         try:
 297             login_page = compat_urllib_request.urlopen(request).read().decode('utf-8')
 298         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 299             self._downloader.report_warning(u'unable to fetch login page: %s' % compat_str(err))
 300             return
 301
 302         galx = None
 303         dsh = None
 304         match = re.search(re.compile(r'<input.+?name="GALX".+?value="(.+?)"', re.DOTALL), login_page)
 305         if match:
 306           galx = match.group(1)
 307
 308         match = re.search(re.compile(r'<input.+?name="dsh".+?value="(.+?)"', re.DOTALL), login_page)
 309         if match:
 310           dsh = match.group(1)
 311
 312         # Log in
 313         login_form_strs = {
 314                 u'continue': u'https://www.youtube.com/signin?action_handle_signin=true&feature=sign_in_button&hl=en_US&nomobiletemp=1',
 315                 u'Email': username,
 316                 u'GALX': galx,
 317                 u'Passwd': password,
 318                 u'PersistentCookie': u'yes',
 319                 u'_utf8': u'霱',
 320                 u'bgresponse': u'js_disabled',
 321                 u'checkConnection': u'',
 322                 u'checkedDomains': u'youtube',
 323                 u'dnConn': u'',
 324                 u'dsh': dsh,
 325                 u'pstMsg': u'0',
 326                 u'rmShown': u'1',
 327                 u'secTok': u'',
 328                 u'signIn': u'Sign in',
 329                 u'timeStmp': u'',
 330                 u'service': u'youtube',
 331                 u'uilel': u'3',
 332                 u'hl': u'en_US',
 333         }
 334         # Convert to UTF-8 *before* urlencode because Python 2.x's urlencode
 335         # chokes on unicode
 336         login_form = dict((k.encode('utf-8'), v.encode('utf-8')) for k,v in login_form_strs.items())
 337         login_data = compat_urllib_parse.urlencode(login_form).encode('ascii')
 338         request = compat_urllib_request.Request(self._LOGIN_URL, login_data)
 339         try:
 340             self.report_login()
 341             login_results = compat_urllib_request.urlopen(request).read().decode('utf-8')
 342             if re.search(r'(?i)<form[^>]* id="gaia_loginform"', login_results) is not None:
 343                 self._downloader.report_warning(u'unable to log in: bad username or password')
 344                 return
 345         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 346             self._downloader.report_warning(u'unable to log in: %s' % compat_str(err))
 347             return
 348
 349         # Confirm age
 350         age_form = {
 351                 'next_url':     '/',
 352                 'action_confirm':   'Confirm',
 353                 }
 354         request = compat_urllib_request.Request(self._AGE_URL, compat_urllib_parse.urlencode(age_form))
 355         try:
 356             self.report_age_confirmation()
 357             compat_urllib_request.urlopen(request).read().decode('utf-8')
 358         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 359             raise ExtractorError(u'Unable to confirm age: %s' % compat_str(err))
 360
 361     def _extract_id(self, url):
 362         mobj = re.match(self._VALID_URL, url, re.VERBOSE)
 363         if mobj is None:
 364             raise ExtractorError(u'Invalid URL: %s' % url)
 365         video_id = mobj.group(2)
 366         return video_id
 367
 368     def _real_extract(self, url):
 369         # Extract original video URL from URL with redirection, like age verification, using next_url parameter
 370         mobj = re.search(self._NEXT_URL_RE, url)
 371         if mobj:
 372             url = 'https://www.youtube.com/' + compat_urllib_parse.unquote(mobj.group(1)).lstrip('/')
 373         video_id = self._extract_id(url)
 374
 375         # Get video webpage
 376         self.report_video_webpage_download(video_id)
 377         url = 'https://www.youtube.com/watch?v=%s&gl=US&hl=en&has_verified=1' % video_id
 378         request = compat_urllib_request.Request(url)
 379         try:
 380             video_webpage_bytes = compat_urllib_request.urlopen(request).read()
 381         except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 382             raise ExtractorError(u'Unable to download video webpage: %s' % compat_str(err))
 383
 384         video_webpage = video_webpage_bytes.decode('utf-8', 'ignore')
 385
 386         # Attempt to extract SWF player URL
 387         mobj = re.search(r'swfConfig.*?"(http:\\/\\/.*?watch.*?-.*?\.swf)"', video_webpage)
 388         if mobj is not None:
 389             player_url = re.sub(r'\\(.)', r'\1', mobj.group(1))
 390         else:
 391             player_url = None
 392
 393         # Get video info
 394         self.report_video_info_webpage_download(video_id)
 395         for el_type in ['&el=embedded', '&el=detailpage', '&el=vevo', '']:
 396             video_info_url = ('https://www.youtube.com/get_video_info?&video_id=%s%s&ps=default&eurl=&gl=US&hl=en'
 397                     % (video_id, el_type))
 398             video_info_webpage = self._download_webpage(video_info_url, video_id,
 399                                     note=False,
 400                                     errnote='unable to download video info webpage')
 401             video_info = compat_parse_qs(video_info_webpage)
 402             if 'token' in video_info:
 403                 break
 404         if 'token' not in video_info:
 405             if 'reason' in video_info:
 406                 raise ExtractorError(u'YouTube said: %s' % video_info['reason'][0])
 407             else:
 408                 raise ExtractorError(u'"token" parameter not in video info for unknown reason')
 409
 410         # Check for "rental" videos
 411         if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
 412             raise ExtractorError(u'"rental" videos not supported')
 413
 414         # Start extracting information
 415         self.report_information_extraction(video_id)
 416
 417         # uploader
 418         if 'author' not in video_info:
 419             raise ExtractorError(u'Unable to extract uploader name')
 420         video_uploader = compat_urllib_parse.unquote_plus(video_info['author'][0])
 421
 422         # uploader_id
 423         video_uploader_id = None
 424         mobj = re.search(r'<link itemprop="url" href="http://www.youtube.com/(?:user|channel)/([^"]+)">', video_webpage)
 425         if mobj is not None:
 426             video_uploader_id = mobj.group(1)
 427         else:
 428             self._downloader.report_warning(u'unable to extract uploader nickname')
 429
 430         # title
 431         if 'title' not in video_info:
 432             raise ExtractorError(u'Unable to extract video title')
 433         video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
 434
 435         # thumbnail image
 436         if 'thumbnail_url' not in video_info:
 437             self._downloader.report_warning(u'unable to extract video thumbnail')
 438             video_thumbnail = ''
 439         else:   # don't panic if we can't find it
 440             video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
 441
 442         # upload date
 443         upload_date = None
 444         mobj = re.search(r'id="eow-date.*?>(.*?)</span>', video_webpage, re.DOTALL)
 445         if mobj is not None:
 446             upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
 447             upload_date = unified_strdate(upload_date)
 448
 449         # description
 450         video_description = get_element_by_id("eow-description", video_webpage)
 451         if video_description:
 452             video_description = clean_html(video_description)
 453         else:
 454             fd_mobj = re.search(r'<meta name="description" content="([^"]+)"', video_webpage)
 455             if fd_mobj:
 456                 video_description = unescapeHTML(fd_mobj.group(1))
 457             else:
 458                 video_description = u''
 459
 460         # subtitles
 461         video_subtitles = None
 462
 463         if self._downloader.params.get('writesubtitles', False):
 464             video_subtitles = self._extract_subtitle(video_id)
 465             if video_subtitles:
 466                 (sub_error, sub_lang, sub) = video_subtitles[0]
 467                 if sub_error:
 468                     self._downloader.report_warning(sub_error)
 469
 470         if self._downloader.params.get('writeautomaticsub', False):
 471             video_subtitles = self._request_automatic_caption(video_id, video_webpage)
 472             (sub_error, sub_lang, sub) = video_subtitles[0]
 473             if sub_error:
 474                 self._downloader.report_warning(sub_error)
 475
 476         if self._downloader.params.get('allsubtitles', False):
 477             video_subtitles = self._extract_all_subtitles(video_id)
 478             for video_subtitle in video_subtitles:
 479                 (sub_error, sub_lang, sub) = video_subtitle
 480                 if sub_error:
 481                     self._downloader.report_warning(sub_error)
 482
 483         if self._downloader.params.get('listsubtitles', False):
 484             self._list_available_subtitles(video_id)
 485             return
 486
 487         if 'length_seconds' not in video_info:
 488             self._downloader.report_warning(u'unable to extract video duration')
 489             video_duration = ''
 490         else:
 491             video_duration = compat_urllib_parse.unquote_plus(video_info['length_seconds'][0])
 492
 493         # Decide which formats to download
 494         req_format = self._downloader.params.get('format', None)
 495
 496         try:
 497             mobj = re.search(r';ytplayer.config = ({.*?});', video_webpage)
 498             if not mobj:
 499                 raise ValueError('Could not find vevo ID')
 500             info = json.loads(mobj.group(1))
 501             args = info['args']
 502             # Easy way to know if the 's' value is in url_encoded_fmt_stream_map
 503             # this signatures are encrypted
 504             m_s = re.search(r'[&,]s=', args['url_encoded_fmt_stream_map'])
 505             if m_s is not None:
 506                 self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
 507                 video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
 508         except ValueError:
 509             pass
 510
 511         if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
 512             self.report_rtmp_download()
 513             video_url_list = [(None, video_info['conn'][0])]
 514         elif 'url_encoded_fmt_stream_map' in video_info and len(video_info['url_encoded_fmt_stream_map']) >= 1:
 515             url_map = {}
 516             for url_data_str in video_info['url_encoded_fmt_stream_map'][0].split(','):
 517                 url_data = compat_parse_qs(url_data_str)
 518                 if 'itag' in url_data and 'url' in url_data:
 519                     url = url_data['url'][0]
 520                     if 'sig' in url_data:
 521                         url += '&signature=' + url_data['sig'][0]
 522                     elif 's' in url_data:
 523                         signature = self._decrypt_signature(url_data['s'][0])
 524                         url += '&signature=' + signature
 525                     if 'ratebypass' not in url:
 526                         url += '&ratebypass=yes'
 527                     url_map[url_data['itag'][0]] = url
 528
 529             format_limit = self._downloader.params.get('format_limit', None)
 530             available_formats = self._available_formats_prefer_free if self._downloader.params.get('prefer_free_formats', False) else self._available_formats
 531             if format_limit is not None and format_limit in available_formats:
 532                 format_list = available_formats[available_formats.index(format_limit):]
 533             else:
 534                 format_list = available_formats
 535             existing_formats = [x for x in format_list if x in url_map]
 536             if len(existing_formats) == 0:
 537                 raise ExtractorError(u'no known formats available for video')
 538             if self._downloader.params.get('listformats', None):
 539                 self._print_formats(existing_formats)
 540                 return
 541             if req_format is None or req_format == 'best':
 542                 video_url_list = [(existing_formats[0], url_map[existing_formats[0]])] # Best quality
 543             elif req_format == 'worst':
 544                 video_url_list = [(existing_formats[len(existing_formats)-1], url_map[existing_formats[len(existing_formats)-1]])] # worst quality
 545             elif req_format in ('-1', 'all'):
 546                 video_url_list = [(f, url_map[f]) for f in existing_formats] # All formats
 547             else:
 548                 # Specific formats. We pick the first in a slash-delimeted sequence.
 549                 # For example, if '1/2/3/4' is requested and '2' and '4' are available, we pick '2'.
 550                 req_formats = req_format.split('/')
 551                 video_url_list = None
 552                 for rf in req_formats:
 553                     if rf in url_map:
 554                         video_url_list = [(rf, url_map[rf])]
 555                         break
 556                 if video_url_list is None:
 557                     raise ExtractorError(u'requested format not available')
 558         else:
 559             raise ExtractorError(u'no conn or url_encoded_fmt_stream_map information found in video info')
 560
 561         results = []
 562         for format_param, video_real_url in video_url_list:
 563             # Extension
 564             video_extension = self._video_extensions.get(format_param, 'flv')
 565
 566             video_format = '{0} - {1}'.format(format_param if format_param else video_extension,
 567                                               self._video_dimensions.get(format_param, '???'))
 568
 569             results.append({
 570                 'id':       video_id,
 571                 'url':      video_real_url,
 572                 'uploader': video_uploader,
 573                 'uploader_id': video_uploader_id,
 574                 'upload_date':  upload_date,
 575                 'title':    video_title,
 576                 'ext':      video_extension,
 577                 'format':   video_format,
 578                 'thumbnail':    video_thumbnail,
 579                 'description':  video_description,
 580                 'player_url':   player_url,
 581                 'subtitles':    video_subtitles,
 582                 'duration':     video_duration
 583             })
 584         return results
 585
 586 class YoutubePlaylistIE(InfoExtractor):
 587     """Information Extractor for YouTube playlists."""
 588
 589     _VALID_URL = r"""(?:
 590                         (?:https?://)?
 591                         (?:\w+\.)?
 592                         youtube\.com/
 593                         (?:
 594                            (?:course|view_play_list|my_playlists|artist|playlist|watch)
 595                            \? (?:.*?&)*? (?:p|a|list)=
 596                         |  p/
 597                         )
 598                         ((?:PL|EC|UU)?[0-9A-Za-z-_]{10,})
 599                         .*
 600                      |
 601                         ((?:PL|EC|UU)[0-9A-Za-z-_]{10,})
 602                      )"""
 603     _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/playlists/%s?max-results=%i&start-index=%i&v=2&alt=json&safeSearch=none'
 604     _MAX_RESULTS = 50
 605     IE_NAME = u'youtube:playlist'
 606
 607     @classmethod
 608     def suitable(cls, url):
 609         """Receives a URL and returns True if suitable for this IE."""
 610         return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
 611
 612     def _real_extract(self, url):
 613         # Extract playlist id
 614         mobj = re.match(self._VALID_URL, url, re.VERBOSE)
 615         if mobj is None:
 616             raise ExtractorError(u'Invalid URL: %s' % url)
 617
 618         # Download playlist videos from API
 619         playlist_id = mobj.group(1) or mobj.group(2)
 620         page_num = 1
 621         videos = []
 622
 623         while True:
 624             url = self._TEMPLATE_URL % (playlist_id, self._MAX_RESULTS, self._MAX_RESULTS * (page_num - 1) + 1)
 625             page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
 626
 627             try:
 628                 response = json.loads(page)
 629             except ValueError as err:
 630                 raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
 631
 632             if 'feed' not in response:
 633                 raise ExtractorError(u'Got a malformed response from YouTube API')
 634             playlist_title = response['feed']['title']['$t']
 635             if 'entry' not in response['feed']:
 636                 # Number of videos is a multiple of self._MAX_RESULTS
 637                 break
 638
 639             for entry in response['feed']['entry']:
 640                 index = entry['yt$position']['$t']
 641                 if 'media$group' in entry and 'media$player' in entry['media$group']:
 642                     videos.append((index, entry['media$group']['media$player']['url']))
 643
 644             if len(response['feed']['entry']) < self._MAX_RESULTS:
 645                 break
 646             page_num += 1
 647
 648         videos = [v[1] for v in sorted(videos)]
 649
 650         url_results = [self.url_result(url, 'Youtube') for url in videos]
 651         return [self.playlist_result(url_results, playlist_id, playlist_title)]
 652
 653
 654 class YoutubeChannelIE(InfoExtractor):
 655     """Information Extractor for YouTube channels."""
 656
 657     _VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
 658     _TEMPLATE_URL = 'http://www.youtube.com/channel/%s/videos?sort=da&flow=list&view=0&page=%s&gl=US&hl=en'
 659     _MORE_PAGES_INDICATOR = 'yt-uix-load-more'
 660     _MORE_PAGES_URL = 'http://www.youtube.com/channel_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s'
 661     IE_NAME = u'youtube:channel'
 662
 663     def extract_videos_from_page(self, page):
 664         ids_in_page = []
 665         for mobj in re.finditer(r'href="/watch\?v=([0-9A-Za-z_-]+)&?', page):
 666             if mobj.group(1) not in ids_in_page:
 667                 ids_in_page.append(mobj.group(1))
 668         return ids_in_page
 669
 670     def _real_extract(self, url):
 671         # Extract channel id
 672         mobj = re.match(self._VALID_URL, url)
 673         if mobj is None:
 674             raise ExtractorError(u'Invalid URL: %s' % url)
 675
 676         # Download channel page
 677         channel_id = mobj.group(1)
 678         video_ids = []
 679         pagenum = 1
 680
 681         url = self._TEMPLATE_URL % (channel_id, pagenum)
 682         page = self._download_webpage(url, channel_id,
 683                                       u'Downloading page #%s' % pagenum)
 684
 685         # Extract video identifiers
 686         ids_in_page = self.extract_videos_from_page(page)
 687         video_ids.extend(ids_in_page)
 688
 689         # Download any subsequent channel pages using the json-based channel_ajax query
 690         if self._MORE_PAGES_INDICATOR in page:
 691             while True:
 692                 pagenum = pagenum + 1
 693
 694                 url = self._MORE_PAGES_URL % (pagenum, channel_id)
 695                 page = self._download_webpage(url, channel_id,
 696                                               u'Downloading page #%s' % pagenum)
 697
 698                 page = json.loads(page)
 699
 700                 ids_in_page = self.extract_videos_from_page(page['content_html'])
 701                 video_ids.extend(ids_in_page)
 702
 703                 if self._MORE_PAGES_INDICATOR  not in page['load_more_widget_html']:
 704                     break
 705
 706         self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
 707
 708         urls = ['http://www.youtube.com/watch?v=%s' % id for id in video_ids]
 709         url_entries = [self.url_result(url, 'Youtube') for url in urls]
 710         return [self.playlist_result(url_entries, channel_id)]
 711
 712
 713 class YoutubeUserIE(InfoExtractor):
 714     """Information Extractor for YouTube users."""
 715
 716     _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/user/)|ytuser:)([A-Za-z0-9_-]+)'
 717     _TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s'
 718     _GDATA_PAGE_SIZE = 50
 719     _GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d'
 720     _VIDEO_INDICATOR = r'/watch\?v=(.+?)[\<&]'
 721     IE_NAME = u'youtube:user'
 722
 723     def _real_extract(self, url):
 724         # Extract username
 725         mobj = re.match(self._VALID_URL, url)
 726         if mobj is None:
 727             raise ExtractorError(u'Invalid URL: %s' % url)
 728
 729         username = mobj.group(1)
 730
 731         # Download video ids using YouTube Data API. Result size per
 732         # query is limited (currently to 50 videos) so we need to query
 733         # page by page until there are no video ids - it means we got
 734         # all of them.
 735
 736         video_ids = []
 737         pagenum = 0
 738
 739         while True:
 740             start_index = pagenum * self._GDATA_PAGE_SIZE + 1
 741
 742             gdata_url = self._GDATA_URL % (username, self._GDATA_PAGE_SIZE, start_index)
 743             page = self._download_webpage(gdata_url, username,
 744                                           u'Downloading video ids from %d to %d' % (start_index, start_index + self._GDATA_PAGE_SIZE))
 745
 746             # Extract video identifiers
 747             ids_in_page = []
 748
 749             for mobj in re.finditer(self._VIDEO_INDICATOR, page):
 750                 if mobj.group(1) not in ids_in_page:
 751                     ids_in_page.append(mobj.group(1))
 752
 753             video_ids.extend(ids_in_page)
 754
 755             # A little optimization - if current page is not
 756             # "full", ie. does not contain PAGE_SIZE video ids then
 757             # we can assume that this page is the last one - there
 758             # are no more ids on further pages - no need to query
 759             # again.
 760
 761             if len(ids_in_page) < self._GDATA_PAGE_SIZE:
 762                 break
 763
 764             pagenum += 1
 765
 766         urls = ['http://www.youtube.com/watch?v=%s' % video_id for video_id in video_ids]
 767         url_results = [self.url_result(url, 'Youtube') for url in urls]
 768         return [self.playlist_result(url_results, playlist_title = username)]
 769
 770 class YoutubeSearchIE(SearchInfoExtractor):
 771     """Information Extractor for YouTube search queries."""
 772     _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc'
 773     _MAX_RESULTS = 1000
 774     IE_NAME = u'youtube:search'
 775     _SEARCH_KEY = 'ytsearch'
 776
 777     def report_download_page(self, query, pagenum):
 778         """Report attempt to download search page with given number."""
 779         self._downloader.to_screen(u'[youtube] query "%s": Downloading page %s' % (query, pagenum))
 780
 781     def _get_n_results(self, query, n):
 782         """Get a specified number of results for a query"""
 783
 784         video_ids = []
 785         pagenum = 0
 786         limit = n
 787
 788         while (50 * pagenum) < limit:
 789             self.report_download_page(query, pagenum+1)
 790             result_url = self._API_URL % (compat_urllib_parse.quote_plus(query), (50*pagenum)+1)
 791             request = compat_urllib_request.Request(result_url)
 792             try:
 793                 data = compat_urllib_request.urlopen(request).read().decode('utf-8')
 794             except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
 795                 raise ExtractorError(u'Unable to download API page: %s' % compat_str(err))
 796             api_response = json.loads(data)['data']
 797
 798             if not 'items' in api_response:
 799                 raise ExtractorError(u'[youtube] No video results')
 800
 801             new_ids = list(video['id'] for video in api_response['items'])
 802             video_ids += new_ids
 803
 804             limit = min(n, api_response['totalItems'])
 805             pagenum += 1
 806
 807         if len(video_ids) > n:
 808             video_ids = video_ids[:n]
 809         videos = [self.url_result('http://www.youtube.com/watch?v=%s' % id, 'Youtube') for id in video_ids]
 810         return self.playlist_result(videos, query)