gitlab-com-dstftw/youtube-dl
1
0
Fork 0
mirror of https://gitlab.com/dstftw/youtube-dl.git synced 2020-11-16 09:42:26 +00:00
Code Issues Releases Wiki Activity

[pbs] add test for videos with undocumented http formats and remove unused import

Browse source
This commit is contained in:
Remita Amine 2016-08-13 09:08:57 +01:00
parent c366f8d30a
commit f50365e91c
1 changed files with 16 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

19
youtube_dl/extractor/pbs.py
View file

@ -4,7 +4,6 @@ from __future__ import unicode_literals
import re import re
from .common import InfoExtractor from .common import InfoExtractor
from ..compat import compat_HTTPError
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
determine_ext, determine_ext,
@ -334,6 +333,16 @@ class PBSIE(InfoExtractor):
'formats': 'mincount:8', 'formats': 'mincount:8',
}, },
}, },
{
# has undocumented http formats(4500k and 6500k)
'url': 'http://www.pbs.org/video/2365815229/',
'md5': '94635cd06b7133688e23f4b94e6637a5',
'info_dict': {
'id': '2365815229',
'ext': 'mp4',
'title': 'FRONTLINE - Mosquito Hunter',
},
},
{ {
'url': 'http://player.pbs.org/widget/partnerplayer/2365297708/?start=0&end=0&chapterbar=false&endscreen=false&topbar=true', 'url': 'http://player.pbs.org/widget/partnerplayer/2365297708/?start=0&end=0&chapterbar=false&endscreen=false&topbar=true',
'only_matching': True, 'only_matching': True,
@ -501,14 +510,18 @@ class PBSIE(InfoExtractor):
if http_url: if http_url:
for m3u8_format in m3u8_formats: for m3u8_format in m3u8_formats:
bitrate = self._search_regex(r'(\d+)k', m3u8_format['url'], 'bitrate', default=None) bitrate = self._search_regex(r'(\d+)k', m3u8_format['url'], 'bitrate', default=None)
# extract only the formats that we know that they will be available as http format. # lower qualities(150k and 192k) are not available as http formats
# https://github.com/rg3/youtube-dl/commit/cbc032c8b70a038a69259378c92b4ba97b42d491#commitcomment-17313656
# we will try to extract any http format higher than than the lowest quality documented in
# https://projects.pbs.org/confluence/display/coveapi/COVE+Video+Specifications # https://projects.pbs.org/confluence/display/coveapi/COVE+Video+Specifications
# as there also undocumented http formats formats(4500k and 6500k)
# http://www.pbs.org/video/2365815229/
if not bitrate or int(bitrate) < 400: if not bitrate or int(bitrate) < 400:
continue continue
f_url = re.sub(r'\d+k|baseline', bitrate + 'k', http_url) f_url = re.sub(r'\d+k|baseline', bitrate + 'k', http_url)
# This may produce invalid links sometimes (e.g. # This may produce invalid links sometimes (e.g.
# http://www.pbs.org/wgbh/frontline/film/suicide-plan) # http://www.pbs.org/wgbh/frontline/film/suicide-plan)
if not self._is_valid_url(f_url, display_id, 'http-%s video' % bitrate): if not self._is_valid_url(f_url, display_id, 'http-%sk video' % bitrate):
continue continue
f = m3u8_format.copy() f = m3u8_format.copy()
f.update({ f.update({