[facebook] Fix support for untitled videos (Fixes #3757)

This commit is contained in:
Philipp Hagemeister 2014-09-15 15:10:24 +02:00
parent b04c8f7358
commit a020a0dc20
3 changed files with 28 additions and 3 deletions

View File

@ -40,6 +40,7 @@
parse_iso8601, parse_iso8601,
strip_jsonp, strip_jsonp,
uppercase_escape, uppercase_escape,
limit_length,
) )
@ -286,5 +287,12 @@ def test_uppercase_escape(self):
self.assertEqual(uppercase_escape(''), '') self.assertEqual(uppercase_escape(''), '')
self.assertEqual(uppercase_escape('\\U0001d550'), '𝕐') self.assertEqual(uppercase_escape('\\U0001d550'), '𝕐')
def test_limit_length(self):
self.assertEqual(limit_length(None, 12), None)
self.assertEqual(limit_length('foo', 12), 'foo')
self.assertTrue(
limit_length('foo bar baz asd', 12).startswith('foo bar'))
self.assertTrue('...' in limit_length('foo bar baz asd', 12))
if __name__ == '__main__': if __name__ == '__main__':
unittest.main() unittest.main()

View File

@ -12,8 +12,8 @@
compat_urllib_parse, compat_urllib_parse,
compat_urllib_request, compat_urllib_request,
urlencode_postdata, urlencode_postdata,
ExtractorError, ExtractorError,
limit_length,
) )
@ -37,6 +37,14 @@ class FacebookIE(InfoExtractor):
'duration': 38, 'duration': 38,
'title': 'Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam fin...', 'title': 'Did you know Kei Nishikori is the first Asian man to ever reach a Grand Slam fin...',
} }
}, {
'note': 'Video without discernible title',
'url': 'https://www.facebook.com/video.php?v=274175099429670',
'info_dict': {
'id': '274175099429670',
'ext': 'mp4',
'title': 'Facebook video #274175099429670',
}
}, { }, {
'url': 'https://www.facebook.com/video.php?v=10204634152394104', 'url': 'https://www.facebook.com/video.php?v=10204634152394104',
'only_matching': True, 'only_matching': True,
@ -131,8 +139,7 @@ def _real_extract(self, url):
video_title = self._html_search_regex( video_title = self._html_search_regex(
r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>', r'(?s)<span class="fbPhotosPhotoCaption".*?id="fbPhotoPageCaption"><span class="hasCaption">(.*?)</span>',
webpage, 'alternative title', default=None) webpage, 'alternative title', default=None)
if len(video_title) > 80 + 3: video_title = limit_length(video_title, 80)
video_title = video_title[:80] + '...'
if not video_title: if not video_title:
video_title = 'Facebook video #%s' % video_id video_title = 'Facebook video #%s' % video_id

View File

@ -1571,3 +1571,13 @@ def subprocess_check_output(*args, **kwargs):
if ret: if ret:
raise subprocess.CalledProcessError(ret, p.args, output=output) raise subprocess.CalledProcessError(ret, p.args, output=output)
return output return output
def limit_length(s, length):
""" Add ellipses to overly long strings """
if s is None:
return None
ELLIPSES = '...'
if len(s) > length:
return s[:length - len(ELLIPSES)] + ELLIPSES
return s