diff options
author | Sergey M․ <dstftw@gmail.com> | 2015-06-12 17:49:07 +0600 |
---|---|---|
committer | Sergey M․ <dstftw@gmail.com> | 2015-06-12 17:49:07 +0600 |
commit | e5095f1198f9f7a7f01a6692bcdee0e92f577dd5 (patch) | |
tree | 3678808d2c8ab9d79a0c35f44b747b46e5ad39c8 | |
parent | 3507766bd0c1e753724ffe23c6d35b903bfc51a0 (diff) | |
parent | 499a077761b1577857952dc3b541c9f61a8bcade (diff) |
Merge branch 'hlintala-5tv'
-rw-r--r-- | youtube_dl/extractor/__init__.py | 1 | ||||
-rw-r--r-- | youtube_dl/extractor/fivetv.py | 88 |
2 files changed, 89 insertions, 0 deletions
diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py index 18b1c5e54..3bc62e9d5 100644 --- a/youtube_dl/extractor/__init__.py +++ b/youtube_dl/extractor/__init__.py @@ -152,6 +152,7 @@ from .fc2 import FC2IE from .firstpost import FirstpostIE from .firsttv import FirstTVIE from .fivemin import FiveMinIE +from .fivetv import FiveTVIE from .fktv import ( FKTVIE, FKTVPosteckeIE, diff --git a/youtube_dl/extractor/fivetv.py b/youtube_dl/extractor/fivetv.py new file mode 100644 index 000000000..13fbc4da2 --- /dev/null +++ b/youtube_dl/extractor/fivetv.py @@ -0,0 +1,88 @@ +# coding: utf-8 +from __future__ import unicode_literals + +import re + +from .common import InfoExtractor +from ..utils import int_or_none + + +class FiveTVIE(InfoExtractor): + _VALID_URL = r'''(?x) + http:// + (?:www\.)?5-tv\.ru/ + (?: + (?:[^/]+/)+(?P<id>\d+)| + (?P<path>[^/?#]+)(?:[/?#])? + ) + ''' + + _TESTS = [{ + 'url': 'http://5-tv.ru/news/96814/', + 'md5': 'bbff554ad415ecf5416a2f48c22d9283', + 'info_dict': { + 'id': '96814', + 'ext': 'mp4', + 'title': 'Россияне выбрали имя для общенациональной платежной системы', + 'description': 'md5:a8aa13e2b7ad36789e9f77a74b6de660', + 'thumbnail': 're:^https?://.*\.jpg$', + 'duration': 180, + }, + }, { + 'url': 'http://5-tv.ru/video/1021729/', + 'info_dict': { + 'id': '1021729', + 'ext': 'mp4', + 'title': '3D принтер', + 'description': 'md5:d76c736d29ef7ec5c0cf7d7c65ffcb41', + 'thumbnail': 're:^https?://.*\.jpg$', + 'duration': 180, + }, + }, { + 'url': 'http://www.5-tv.ru/glavnoe/#itemDetails', + 'info_dict': { + 'id': 'glavnoe', + 'ext': 'mp4', + 'title': 'Итоги недели с 8 по 14 июня 2015 года', + 'thumbnail': 're:^https?://.*\.jpg$', + }, + }, { + 'url': 'http://www.5-tv.ru/glavnoe/broadcasts/508645/', + 'only_matching': True, + }, { + 'url': 'http://5-tv.ru/films/1507502/', + 'only_matching': True, + }, { + 'url': 'http://5-tv.ru/programs/broadcast/508713/', + 'only_matching': True, + }, { + 'url': 'http://5-tv.ru/angel/', + 'only_matching': True, + }, { + 'url': 'http://www.5-tv.ru/schedule/?iframe=true&width=900&height=450', + 'only_matching': True, + }] + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + video_id = mobj.group('id') or mobj.group('path') + + webpage = self._download_webpage(url, video_id) + + video_url = self._search_regex( + r'<a[^>]+?href="([^"]+)"[^>]+?class="videoplayer"', + webpage, 'video url') + + title = self._og_search_title(webpage, default=None) or self._search_regex( + r'<title>([^<]+)</title>', webpage, 'title') + duration = int_or_none(self._og_search_property( + 'video:duration', webpage, 'duration', default=None)) + + return { + 'id': video_id, + 'url': video_url, + 'title': title, + 'description': self._og_search_description(webpage, default=None), + 'thumbnail': self._og_search_thumbnail(webpage, default=None), + 'duration': duration, + } |