X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/9dc487f48b50767cf540fa36c3de2c386fd74c04..1d3fd83f473663fce3e0a10303473a38d80cc3d0:/youtube_dl/extractor/vidzi.py?ds=inline diff --git a/youtube_dl/extractor/vidzi.py b/youtube_dl/extractor/vidzi.py index 3c78fb3..d055629 100644 --- a/youtube_dl/extractor/vidzi.py +++ b/youtube_dl/extractor/vidzi.py @@ -1,16 +1,20 @@ # coding: utf-8 from __future__ import unicode_literals -from .jwplatform import JWPlatformBaseIE +import re + +from .common import InfoExtractor from ..utils import ( decode_packed_codes, js_to_json, + NO_DEFAULT, + PACKED_CODES_RE, ) -class VidziIE(JWPlatformBaseIE): - _VALID_URL = r'https?://(?:www\.)?vidzi\.tv/(?P\w+)' - _TEST = { +class VidziIE(InfoExtractor): + _VALID_URL = r'https?://(?:www\.)?vidzi\.(?:tv|cc)/(?:embed-)?(?P[0-9a-zA-Z]+)' + _TESTS = [{ 'url': 'http://vidzi.tv/cghql9yq6emu.html', 'md5': '4f16c71ca0c8c8635ab6932b5f3f1660', 'info_dict': { @@ -22,19 +26,33 @@ class VidziIE(JWPlatformBaseIE): # m3u8 download 'skip_download': True, }, - } + }, { + 'url': 'http://vidzi.tv/embed-4z2yb0rzphe9-600x338.html', + 'skip_download': True, + }, { + 'url': 'http://vidzi.cc/cghql9yq6emu.html', + 'skip_download': True, + }] def _real_extract(self, url): video_id = self._match_id(url) - webpage = self._download_webpage(url, video_id) + webpage = self._download_webpage( + 'http://vidzi.tv/%s' % video_id, video_id) title = self._html_search_regex( r'(?s)

(.*?)

', webpage, 'title') - code = decode_packed_codes(webpage).replace('\\\'', '\'') - jwplayer_data = self._parse_json( - self._search_regex(r'setup\(([^)]+)\)', code, 'jwplayer data'), - video_id, transform_source=js_to_json) + packed_codes = [mobj.group(0) for mobj in re.finditer( + PACKED_CODES_RE, webpage)] + for num, pc in enumerate(packed_codes, 1): + code = decode_packed_codes(pc).replace('\\\'', '\'') + jwplayer_data = self._parse_json( + self._search_regex( + r'setup\(([^)]+)\)', code, 'jwplayer data', + default=NO_DEFAULT if num == len(packed_codes) else '{}'), + video_id, transform_source=js_to_json) + if jwplayer_data: + break info_dict = self._parse_jwplayer_data(jwplayer_data, video_id, require_title=False) info_dict['title'] = title