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

[pbs] check all http formats and remove unnecessary request

Browse source 
- some of the quality that not reported in the documentation
are available(4500k, 6500k)
- the videoInfo request doesn't work for a long time
This commit is contained in:
Remita Amine 2016-08-12 08:38:06 +01:00
parent 990d533ee4
commit 3cddb8d6a7
1 changed files with 3 additions and 14 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

17
youtube_dl/extractor/pbs.py
View file

@ -448,17 +448,6 @@ class PBSIE(InfoExtractor):
redirects.append(redirect) redirects.append(redirect)
redirect_urls.add(redirect_url) redirect_urls.add(redirect_url)
try:
video_info = self._download_json(
'http://player.pbs.org/videoInfo/%s?format=json&type=partner' % video_id,
display_id, 'Downloading video info JSON')
extract_redirect_urls(video_info)
info = video_info
except ExtractorError as e:
# videoInfo API may not work for some videos
if not isinstance(e.cause, compat_HTTPError) or e.cause.code != 404:
raise
# Player pages may also serve different qualities # Player pages may also serve different qualities
for page in ('widget/partnerplayer', 'portalplayer'): for page in ('widget/partnerplayer', 'portalplayer'):
player = self._download_webpage( player = self._download_webpage(
@ -511,12 +500,12 @@ class PBSIE(InfoExtractor):
formats)) formats))
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. # extract only the formats that we know that they will be available as http format.
# https://projects.pbs.org/confluence/display/coveapi/COVE+Video+Specifications # https://projects.pbs.org/confluence/display/coveapi/COVE+Video+Specifications
if not bitrate or bitrate not in ('400k', '800k', '1200k', '2500k'): if not bitrate or int(bitrate) < 400:
continue continue
f_url = re.sub(r'\d+k|baseline', bitrate, 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-%s video' % bitrate):