[extractor/generic] Improve ISM extraction

This commit is contained in:
Sergey M․ 2016-11-02 23:34:37 +07:00
parent 4f9cd4d36f
commit 26aae56690
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D

View file

@ -2453,8 +2453,21 @@ class GenericIE(InfoExtractor):
entry_info_dict['formats'] = self._extract_mpd_formats(video_url, video_id) entry_info_dict['formats'] = self._extract_mpd_formats(video_url, video_id)
elif ext == 'f4m': elif ext == 'f4m':
entry_info_dict['formats'] = self._extract_f4m_formats(video_url, video_id) entry_info_dict['formats'] = self._extract_f4m_formats(video_url, video_id)
elif re.search(r'(?i)\.ism/manifest', video_url): elif re.search(r'(?i)\.(?:ism|smil)/manifest', video_url):
entry_info_dict['formats'] = self._extract_ism_formats(video_url, video_id) # Just matching .ism/manifest is not enough to be reliably sure
# whether it's actually an ISM manifest or some other streaming
# manifest since there are various streaming URL formats
# possible (see [1]) as well as some other shenanigans like
# .smil/manifest URLs that actually serve an ISM (see [2]) and
# so on.
# Thus the most reasonable way to solve this is to delegate
# to generic extractor in order to look into the contents of
# the manifest itself.
# 1. https://azure.microsoft.com/en-us/documentation/articles/media-services-deliver-content-overview/#streaming-url-formats
# 2. https://svs.itworkscdn.net/lbcivod/smil:itwfcdn/lbci/170976.smil/Manifest
entry_info_dict = self.url_result(
smuggle_url(video_url, {'to_generic': True}),
GenericIE.ie_key())
else: else:
entry_info_dict['url'] = video_url entry_info_dict['url'] = video_url