X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fyoutube.py;h=f6ffb86c3f3dcdfc40cbebffa67ae65fff239919;hb=18b4e04f1c663e0ea695f6501b860f85af9d7ca1;hp=1599dd484cf79815a6cf9da0069352802b0cad1b;hpb=37b6d5f684d409365bbac6d3f2b8074b57e643a8;p=youtube-dl diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 1599dd484..f6ffb86c3 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -7,6 +7,7 @@ import socket import itertools from .common import InfoExtractor, SearchInfoExtractor +from .subtitles import SubtitlesIE from ..utils import ( compat_http_client, compat_parse_qs, @@ -23,114 +24,67 @@ from ..utils import ( orderedSet, ) -class YoutubeBaseInfoExtractor(InfoExtractor): - """Provide base functions for Youtube extractors""" - _LOGIN_URL = 'https://accounts.google.com/ServiceLogin' - _LANG_URL = r'https://www.youtube.com/?hl=en&persist_hl=1&gl=US&persist_gl=1&opt_out_ackd=1' - _AGE_URL = 'http://www.youtube.com/verify_age?next_url=/&gl=US&hl=en' - _NETRC_MACHINE = 'youtube' - # If True it will raise an error if no login info is provided - _LOGIN_REQUIRED = False - - def report_lang(self): - """Report attempt to set language.""" - self.to_screen(u'Setting language') - - def _set_language(self): - request = compat_urllib_request.Request(self._LANG_URL) - try: - self.report_lang() - compat_urllib_request.urlopen(request).read() - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - self._downloader.report_warning(u'unable to set language: %s' % compat_str(err)) - return False - return True - - def _login(self): - (username, password) = self._get_login_info() - # No authentication to be performed - if username is None: - if self._LOGIN_REQUIRED: - raise ExtractorError(u'No login info available, needed for using %s.' % self.IE_NAME, expected=True) - return False - - request = compat_urllib_request.Request(self._LOGIN_URL) - try: - login_page = compat_urllib_request.urlopen(request).read().decode('utf-8') - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - self._downloader.report_warning(u'unable to fetch login page: %s' % compat_str(err)) - return False - - galx = None - dsh = None - match = re.search(re.compile(r']* id="gaia_loginform"', login_results) is not None: - self._downloader.report_warning(u'unable to log in: bad username or password') - return False + sub_list = compat_urllib_request.urlopen(request).read().decode('utf-8') except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - self._downloader.report_warning(u'unable to log in: %s' % compat_str(err)) - return False - return True - - def _confirm_age(self): - age_form = { - 'next_url': '/', - 'action_confirm': 'Confirm', - } - request = compat_urllib_request.Request(self._AGE_URL, compat_urllib_parse.urlencode(age_form)) + self._downloader.report_warning(u'unable to download video subtitles: %s' % compat_str(err)) + return {} + lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list) + + sub_lang_list = {} + for l in lang_list: + lang = l[1] + params = compat_urllib_parse.urlencode({ + 'lang': lang, + 'v': video_id, + 'fmt': self._downloader.params.get('subtitlesformat'), + }) + url = u'http://www.youtube.com/api/timedtext?' + params + sub_lang_list[lang] = url + if not sub_lang_list: + self._downloader.report_warning(u'video doesn\'t have subtitles') + return {} + return sub_lang_list + + def _request_automatic_caption(self, video_id, webpage): + """We need the webpage for getting the captions url, pass it as an + argument to speed up the process.""" + sub_lang = self._downloader.params.get('subtitleslang') or 'en' + sub_format = self._downloader.params.get('subtitlesformat') + self.to_screen(u'%s: Looking for automatic captions' % video_id) + mobj = re.search(r';ytplayer.config = ({.*?});', webpage) + err_msg = u'Couldn\'t find automatic captions for "%s"' % sub_lang + if mobj is None: + self._downloader.report_warning(err_msg) + return {} + player_config = json.loads(mobj.group(1)) try: - self.report_age_confirmation() - compat_urllib_request.urlopen(request).read().decode('utf-8') - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: - raise ExtractorError(u'Unable to confirm age: %s' % compat_str(err)) - return True + args = player_config[u'args'] + caption_url = args[u'ttsurl'] + timestamp = args[u'timestamp'] + params = compat_urllib_parse.urlencode({ + 'lang': 'en', + 'tlang': sub_lang, + 'fmt': sub_format, + 'ts': timestamp, + 'kind': 'asr', + }) + subtitles_url = caption_url + '&' + params + sub = self._download_webpage(subtitles_url, video_id, u'Downloading automatic captions') + return {sub_lang: sub} + # An extractor error can be raise by the download process if there are + # no automatic captions but there are subtitles + except (KeyError, ExtractorError): + self._downloader.report_warning(err_msg) + return {} - def _real_initialize(self): - if self._downloader is None: - return - if not self._set_language(): - return - if not self._login(): - return - self._confirm_age() -class YoutubeIE(YoutubeBaseInfoExtractor): +class YoutubeIE(YoutubeSubtitlesIE): IE_DESC = u'YouTube.com' _VALID_URL = r"""^ ( @@ -141,7 +95,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): (?: # the various things that can precede the ID: (?:(?:v|embed|e)/) # v/ or embed/ or e/ |(?: # or the v= param in all its forms - (?:(?:watch|movie)(?:_popup)?(?:\.php)?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx) + (?:watch|movie(?:_popup)?(?:\.php)?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx) (?:\?|\#!?) # the params delimiter ? or # or #! (?:.*?&)? # any other preceding param (like /?s=tuff&v=xxxx) v= @@ -155,19 +109,22 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # Listed in order of quality _available_formats = ['38', '37', '46', '22', '45', '35', '44', '34', '18', '43', '6', '5', '17', '13', '95', '94', '93', '92', '132', '151', - '85', '84', '102', '83', '101', '82', '100', # 3D - '138', '137', '136', '135', '134', '133', '160', # Dash video mp4 - '141', '140', '139', # Dash auido mp4 - '248', '247', '246', '245', '244', '243', '242', # Dash video webm - '172', '171', # Dash audio webm + # 3D + '85', '84', '102', '83', '101', '82', '100', + # Dash video + '138', '137', '248', '136', '247', '135', '246', + '245', '244', '134', '243', '133', '242', '160', + # Dash audio + '141', '172', '140', '171', '139', ] _available_formats_prefer_free = ['38', '46', '37', '45', '22', '44', '35', '43', '34', '18', '6', '5', '17', '13', '95', '94', '93', '92', '132', '151', '85', '102', '84', '101', '83', '100', '82', - '248', '247', '246', '245', '244', '243', '242', # Dash video webm - '172', '171', # Dash audio webm - '138', '137', '136', '135', '134', '133', '160', # Dash video mp4 - '141', '140', '139', # Dash auido mp4 + # Dash video + '138', '248', '137', '247', '136', '246', '245', + '244', '135', '243', '134', '242', '133', '160', + # Dash audio + '172', '141', '171', '140', '139', ] _video_extensions = { '13': '3gp', @@ -333,7 +290,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): u"upload_date": u"20120506", u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]", u"description": u"md5:b085c9804f5ab69f4adea963a2dceb3c", - u"uploader": u"IconaPop", + u"uploader": u"Icona Pop", u"uploader_id": u"IconaPop" } }, @@ -381,19 +338,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor): """Report attempt to download video info webpage.""" self.to_screen(u'%s: Downloading video info webpage' % video_id) - def report_video_subtitles_download(self, video_id): - """Report attempt to download video info webpage.""" - self.to_screen(u'%s: Checking available subtitles' % video_id) - - def report_video_subtitles_request(self, video_id, sub_lang, format): - """Report attempt to download video info webpage.""" - self.to_screen(u'%s: Downloading video subtitles for %s.%s' % (video_id, sub_lang, format)) - - def report_video_subtitles_available(self, video_id, sub_lang_list): - """Report available subtitles.""" - sub_lang = ",".join(list(sub_lang_list.keys())) - self.to_screen(u'%s: Available subtitles for video: %s' % (video_id, sub_lang)) - def report_information_extraction(self, video_id): """Report attempt to extract video information.""" self.to_screen(u'%s: Extracting video information' % video_id) @@ -733,25 +677,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # subtitles video_subtitles = None - if self._downloader.params.get('writesubtitles', False): - video_subtitles = self._extract_subtitle(video_id) - if video_subtitles: - (sub_error, sub_lang, sub) = video_subtitles[0] - if sub_error: - self._downloader.report_warning(sub_error) - - if self._downloader.params.get('writeautomaticsub', False): + if self._downloader.params.get('writesubtitles', False) or self._downloader.params.get('allsubtitles', False): + video_subtitles = self._extract_subtitles(video_id) + elif self._downloader.params.get('writeautomaticsub', False): video_subtitles = self._request_automatic_caption(video_id, video_webpage) - (sub_error, sub_lang, sub) = video_subtitles[0] - if sub_error: - self._downloader.report_warning(sub_error) - - if self._downloader.params.get('allsubtitles', False): - video_subtitles = self._extract_all_subtitles(video_id) - for video_subtitle in video_subtitles: - (sub_error, sub_lang, sub) = video_subtitle - if sub_error: - self._downloader.report_warning(sub_error) if self._downloader.params.get('listsubtitles', False): self._list_available_subtitles(video_id)