w/youtube-dl
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

[nrktv] Workaround subtitles conversion issues on python 2.6 (Closes #5036)

Browse source
This commit is contained in:
Sergey M․ 2015-02-24 23:06:44 +06:00
parent df4bd0d53f
commit 41b264e77c
1 changed files with 5 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
youtube_dl/extractor/nrk.py
View file

@ -4,6 +4,7 @@ from __future__ import unicode_literals
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import compat_str
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
float_or_none, float_or_none,
@ -158,7 +159,9 @@ class NRKTVIE(InfoExtractor):
def _get_subtitles(self, subtitlesurl, video_id, baseurl): def _get_subtitles(self, subtitlesurl, video_id, baseurl):
url = "%s%s" % (baseurl, subtitlesurl) url = "%s%s" % (baseurl, subtitlesurl)
self._debug_print('%s: Subtitle url: %s' % (video_id, url)) self._debug_print('%s: Subtitle url: %s' % (video_id, url))
captions = self._download_xml(url, video_id, 'Downloading subtitles') captions = self._download_xml(
url, video_id, 'Downloading subtitles',
transform_source=lambda s: s.replace(r'<br />', '\r\n'))
lang = captions.get('lang', 'no') lang = captions.get('lang', 'no')
ps = captions.findall('./{0}body/{0}div/{0}p'.format('{http://www.w3.org/ns/ttml}')) ps = captions.findall('./{0}body/{0}div/{0}p'.format('{http://www.w3.org/ns/ttml}'))
srt = '' srt = ''
@ -167,8 +170,7 @@ class NRKTVIE(InfoExtractor):
duration = parse_duration(p.get('dur')) duration = parse_duration(p.get('dur'))
starttime = self._seconds2str(begin) starttime = self._seconds2str(begin)
endtime = self._seconds2str(begin + duration) endtime = self._seconds2str(begin + duration)
text = '\n'.join(p.itertext()) srt += '%s\r\n%s --> %s\r\n%s\r\n\r\n' % (compat_str(pos), starttime, endtime, p.text)
srt += '%s\r\n%s --> %s\r\n%s\r\n\r\n' % (str(pos), starttime, endtime, text)
return {lang: [ return {lang: [
{'ext': 'ttml', 'url': url}, {'ext': 'ttml', 'url': url},
{'ext': 'srt', 'data': srt}, {'ext': 'srt', 'data': srt},