X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;ds=sidebyside;f=youtube_dl%2Fextractor%2Fsubtitles.py;h=4b4c5235d09ea8a6f75b7d182abf87bfc573557f;hb=2b35c9ef742bf261078ea10c6c0bba848db1a0df;hp=caacea5fe5c380433628138bade91db746645978;hpb=69df680b973841b61594c246a9cf4a708f09cb17;p=youtube-dl diff --git a/youtube_dl/extractor/subtitles.py b/youtube_dl/extractor/subtitles.py index caacea5fe..4b4c5235d 100644 --- a/youtube_dl/extractor/subtitles.py +++ b/youtube_dl/extractor/subtitles.py @@ -1,44 +1,62 @@ -import socket - from .common import InfoExtractor from ..utils import ( - compat_http_client, - compat_urllib_error, - compat_urllib_request, compat_str, + ExtractorError, ) -class SubtitlesIE(InfoExtractor): +class SubtitlesInfoExtractor(InfoExtractor): + @property + def _have_to_download_any_subtitles(self): + return any([self._downloader.params.get('writesubtitles', False), + self._downloader.params.get('writeautomaticsub')]) - def _list_available_subtitles(self, video_id): + def _list_available_subtitles(self, video_id, webpage): """ outputs the available subtitles for the video """ - sub_lang_list = self._get_available_subtitles(video_id) + sub_lang_list = self._get_available_subtitles(video_id, webpage) + auto_captions_list = self._get_available_automatic_caption(video_id, webpage) sub_lang = ",".join(list(sub_lang_list.keys())) self.to_screen(u'%s: Available subtitles for video: %s' % (video_id, sub_lang)) + auto_lang = ",".join(auto_captions_list.keys()) + self.to_screen(u'%s: Available automatic captions for video: %s' % + (video_id, auto_lang)) - def _extract_subtitles(self, video_id): - """ returns {sub_lang: sub} or {} if subtitles not found """ - sub_lang_list = self._get_available_subtitles(video_id) - if not sub_lang_list: # error, it didn't get the available subtitles - return {} - + def extract_subtitles(self, video_id, webpage): + """ + returns {sub_lang: sub} ,{} if subtitles not found or None if the + subtitles aren't requested. + """ + if not self._have_to_download_any_subtitles: + return None + available_subs_list = {} + if self._downloader.params.get('writeautomaticsub', False): + available_subs_list.update(self._get_available_automatic_caption(video_id, webpage)) if self._downloader.params.get('writesubtitles', False): - if self._downloader.params.get('subtitleslang', False): - sub_lang = self._downloader.params.get('subtitleslang') - elif 'en' in sub_lang_list: - sub_lang = 'en' + available_subs_list.update(self._get_available_subtitles(video_id, webpage)) + + if not available_subs_list: # error, it didn't get the available subtitles + return {} + if self._downloader.params.get('allsubtitles', False): + sub_lang_list = available_subs_list + else: + if self._downloader.params.get('subtitleslangs', False): + requested_langs = self._downloader.params.get('subtitleslangs') + elif 'en' in available_subs_list: + requested_langs = ['en'] else: - sub_lang = list(sub_lang_list.keys())[0] - if not sub_lang in sub_lang_list: - self._downloader.report_warning(u'no closed captions found in the specified language "%s"' % sub_lang) - return {} - sub_lang_list = {sub_lang: sub_lang_list[sub_lang]} + requested_langs = [list(available_subs_list.keys())[0]] + + sub_lang_list = {} + for sub_lang in requested_langs: + if not sub_lang in available_subs_list: + self._downloader.report_warning(u'no closed captions found in the specified language "%s"' % sub_lang) + continue + sub_lang_list[sub_lang] = available_subs_list[sub_lang] subtitles = {} - for sub_lang, url in sub_lang_list.iteritems(): + for sub_lang, url in sub_lang_list.items(): subtitle = self._request_subtitle_url(sub_lang, url) if subtitle: subtitles[sub_lang] = subtitle @@ -47,8 +65,8 @@ class SubtitlesIE(InfoExtractor): def _request_subtitle_url(self, sub_lang, url): """ makes the http request for the subtitle """ try: - sub = compat_urllib_request.urlopen(url).read().decode('utf-8') - except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err: + sub = self._download_webpage(url, None, note=False) + except ExtractorError as err: self._downloader.report_warning(u'unable to download video subtitles for %s: %s' % (sub_lang, compat_str(err))) return if not sub: @@ -56,20 +74,18 @@ class SubtitlesIE(InfoExtractor): return return sub - def _get_available_subtitles(self, video_id): - """ returns {sub_lang: url} or {} if not available """ - """ Must be redefined by the subclasses """ - pass - - def _request_automatic_caption(self, video_id, webpage): - """ returns {sub_lang: sub} or {} if not available """ - """ Must be redefined by the subclasses """ + def _get_available_subtitles(self, video_id, webpage): + """ + returns {sub_lang: url} or {} if not available + Must be redefined by the subclasses + """ pass - -class NoAutoSubtitlesIE(SubtitlesIE): - """ A subtitle class for the servers that don't support auto-captions""" - - def _request_automatic_caption(self, video_id, webpage): + def _get_available_automatic_caption(self, video_id, webpage): + """ + returns {sub_lang: url} or {} if not available + Must be redefined by the subclasses that support automatic captions, + otherwise it will return {} + """ self._downloader.report_warning(u'Automatic Captions not supported by this server') return {}