0
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-04 23:35:04 +00:00
yt-dlp/youtube_dl/extractor/subtitles.py
Jaime Marquínez Ferrándiz 0b7f31184d Now --all-sub is a modifier to --write-sub and --write-auto-sub (closes #1412)
For keeping backwards compatibility --all-sub sets --write-sub if --write-auto-sub is not given
2013-09-14 11:14:40 +02:00

91 lines
3.7 KiB
Python

from .common import InfoExtractor
from ..utils import (
compat_str,
ExtractorError,
)
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, webpage=None):
""" outputs the available subtitles for the video """
sub_lang_list = self._get_available_subtitles(video_id)
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, video_webpage=None):
"""
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, video_webpage))
if self._downloader.params.get('writesubtitles', False):
available_subs_list.update(self._get_available_subtitles(video_id))
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:
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.items():
subtitle = self._request_subtitle_url(sub_lang, url)
if subtitle:
subtitles[sub_lang] = subtitle
return subtitles
def _request_subtitle_url(self, sub_lang, url):
""" makes the http request for the subtitle """
try:
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:
self._downloader.report_warning(u'Did not fetch video subtitles')
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 _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 {}