From 055e6f36577497d807d4f474db2489f2e0ef1d4e Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jaime=20Marqui=CC=81nez=20Ferra=CC=81ndiz?= Date: Wed, 11 Sep 2013 19:02:01 +0200 Subject: [youtube] Support automatic captions with original language different from English (fixes #1225) and download in multiple languages. --- youtube_dl/extractor/subtitles.py | 51 +++++++++++++++++---------------------- 1 file changed, 22 insertions(+), 29 deletions(-) (limited to 'youtube_dl/extractor/subtitles.py') diff --git a/youtube_dl/extractor/subtitles.py b/youtube_dl/extractor/subtitles.py index 9a3c54b65..a6780f176 100644 --- a/youtube_dl/extractor/subtitles.py +++ b/youtube_dl/extractor/subtitles.py @@ -15,28 +15,33 @@ class SubtitlesInfoExtractor(InfoExtractor): self.to_screen(u'%s: Available subtitles for video: %s' % (video_id, sub_lang)) - def _extract_subtitles(self, video_id): + def extract_subtitles(self, video_id, video_webpage=None): """ returns {sub_lang: sub} or {} if subtitles not found """ - available_subs_list = self._get_available_subtitles(video_id) + if self._downloader.params.get('writesubtitles', False) or self._downloader.params.get('allsubtitles', False): + available_subs_list = self._get_available_subtitles(video_id) + elif self._downloader.params.get('writeautomaticsub', False): + available_subs_list = self._get_available_automatic_caption(video_id, video_webpage) + else: + return None + 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('writesubtitles', False): - if self._downloader.params.get('subtitleslangs', False): - requested_langs = self._downloader.params.get('subtitleslangs') - elif 'en' in available_subs_list: - requested_langs = ['en'] - else: - requested_langs = [list(available_subs_list.keys())[0]] + if self._downloader.params.get('subtitleslangs', False): + requested_langs = self._downloader.params.get('subtitleslangs') + elif 'en' in available_subs_list: + requested_langs = ['en'] + else: + 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] + 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.items(): @@ -64,23 +69,11 @@ class SubtitlesInfoExtractor(InfoExtractor): """ pass - def _request_automatic_caption(self, video_id, webpage): + def _get_available_automatic_caption(self, video_id, webpage): """ - returns {sub_lang: sub} or {} if not available + 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 {} - - def extract_subtitles(self, video_id, video_webpage=None): - """ - Extract the subtitles and/or the automatic captions if requested. - Returns None or a dictionary in the format {sub_lang: sub} - """ - video_subtitles = None - 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) - return video_subtitles -- cgit v1.2.3