From 9afb1afcc6e3b4fbd6c46e2b6d2d37b5c1c9109f Mon Sep 17 00:00:00 2001 From: Philipp Hagemeister Date: Sat, 29 Jun 2013 17:29:40 +0200 Subject: [PATCH] [tutv] Add IE (Fixes #965) --- youtube_dl/extractor/__init__.py | 1 + youtube_dl/extractor/tutv.py | 41 ++++++++++++++++++++++++++++++++ 2 files changed, 42 insertions(+) create mode 100644 youtube_dl/extractor/tutv.py diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py index 35b619cca..98733e394 100644 --- a/youtube_dl/extractor/__init__.py +++ b/youtube_dl/extractor/__init__.py @@ -50,6 +50,7 @@ from .tf1 import TF1IE from .tudou import TudouIE from .tumblr import TumblrIE +from .tutv import TutvIE from .ustream import UstreamIE from .vbox7 import Vbox7IE from .vevo import VevoIE diff --git a/youtube_dl/extractor/tutv.py b/youtube_dl/extractor/tutv.py new file mode 100644 index 000000000..8afe48348 --- /dev/null +++ b/youtube_dl/extractor/tutv.py @@ -0,0 +1,41 @@ +import base64 +import re + +from .common import InfoExtractor +from ..utils import ( + compat_parse_qs, +) + +class TutvIE(InfoExtractor): + _VALID_URL=r'https?://(?:www\.)?tu\.tv/videos/(?P[^/?]+)' + _TEST = { + u'url': u'http://tu.tv/videos/noah-en-pabellon-cuahutemoc', + u'file': u'2742556.flv', + u'md5': u'5eb766671f69b82e528dc1e7769c5cb2', + u'info_dict': { + u"title": u"Noah en pabellon cuahutemoc" + } + } + + def _real_extract(self, url): + mobj = re.match(self._VALID_URL, url) + video_id = mobj.group('id') + + webpage = self._download_webpage(url, video_id) + title = self._html_search_regex( + r'', webpage, u'title') + internal_id = self._search_regex(r'codVideo=([0-9]+)', webpage, u'internal video ID') + + data_url = u'http://tu.tv/flvurl.php?codVideo=' + str(internal_id) + data_content = self._download_webpage(data_url, video_id, note=u'Downloading video info') + data = compat_parse_qs(data_content) + video_url = base64.b64decode(data['kpt'][0]) + ext = video_url.partition('?')[0].rpartition('.')[2] + + info = { + 'id': internal_id, + 'url': video_url, + 'ext': ext, + 'title': title, + } + return [info]