[OraTV] update extractor

"current" is now "video"
"hls_stream" is now hls_stream without quotes
video_id is now id
duration for current video is not present(for other videos it is)

modified regex to find hls_stream variable to work reguardless of whether it is quoted or not.

[ora] Improve (Closes #8273)
pull/8/head
CeruleanSky 9 years ago committed by Sergey M․
parent 7dde358adc
commit 1f78ed189a

@ -21,7 +21,6 @@ class OraTVIE(InfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'Vine & YouTube Stars Zach King & King Bach On Their Viral Videos!', 'title': 'Vine & YouTube Stars Zach King & King Bach On Their Viral Videos!',
'description': 'md5:ebbc5b1424dd5dba7be7538148287ac1', 'description': 'md5:ebbc5b1424dd5dba7be7538148287ac1',
'duration': 1477,
} }
} }
@ -30,9 +29,9 @@ class OraTVIE(InfoExtractor):
webpage = self._download_webpage(url, display_id) webpage = self._download_webpage(url, display_id)
video_data = self._search_regex( video_data = self._search_regex(
r'"current"\s*:\s*({[^}]+?})', webpage, 'current video') r'"(?:video|current)"\s*:\s*({[^}]+?})', webpage, 'current video')
m3u8_url = self._search_regex( m3u8_url = self._search_regex(
r'"hls_stream"\s*:\s*"([^"]+)', video_data, 'm3u8 url', None) r'hls_stream"?\s*:\s*"([^"]+)', video_data, 'm3u8 url', None)
if m3u8_url: if m3u8_url:
formats = self._extract_m3u8_formats( formats = self._extract_m3u8_formats(
m3u8_url, display_id, 'mp4', 'm3u8_native', m3u8_url, display_id, 'mp4', 'm3u8_native',
@ -62,14 +61,12 @@ class OraTVIE(InfoExtractor):
return { return {
'id': self._search_regex( 'id': self._search_regex(
r'"video_id"\s*:\s*(\d+)', video_data, 'video id'), r'"id"\s*:\s*(\d+)', video_data, 'video id', default=display_id),
'display_id': display_id, 'display_id': display_id,
'title': unescapeHTML(self._og_search_title(webpage)), 'title': unescapeHTML(self._og_search_title(webpage)),
'description': get_element_by_attribute( 'description': get_element_by_attribute(
'class', 'video_txt_decription', webpage), 'class', 'video_txt_decription', webpage),
'thumbnail': self._proto_relative_url(self._search_regex( 'thumbnail': self._proto_relative_url(self._search_regex(
r'"thumb"\s*:\s*"([^"]+)', video_data, 'thumbnail', None)), r'"thumb"\s*:\s*"([^"]+)', video_data, 'thumbnail', None)),
'duration': int(self._search_regex(
r'"duration"\s*:\s*(\d+)', video_data, 'duration')),
'formats': formats, 'formats': formats,
} }

Loading…
Cancel
Save