aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authornulloz <sebastien@mauroy.net>2014-11-22 13:34:29 +0100
committernulloz <sebastien@mauroy.net>2014-11-22 13:34:29 +0100
commit2cead7e7bca9643a2f7c92cf5654ac5da512d6bc (patch)
tree075a38f49f1bd57c8f99446db1510faa79fc9c13
parente07e931375df3984e4f92610413a2f52764bf432 (diff)
downloadyoutube-dl-2cead7e7bca9643a2f7c92cf5654ac5da512d6bc.tar.xz
telebruxelles Add new extractor
-rw-r--r--youtube_dl/extractor/__init__.py1
-rw-r--r--youtube_dl/extractor/telebruxelles.py52
2 files changed, 53 insertions, 0 deletions
diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py
index f45ce05ab..7275d247a 100644
--- a/youtube_dl/extractor/__init__.py
+++ b/youtube_dl/extractor/__init__.py
@@ -379,6 +379,7 @@ from .teachingchannel import TeachingChannelIE
from .teamcoco import TeamcocoIE
from .techtalks import TechTalksIE
from .ted import TEDIE
+from .telebruxelles import TeleBruxellesIE
from .telecinco import TelecincoIE
from .telemb import TeleMBIE
from .tenplay import TenPlayIE
diff --git a/youtube_dl/extractor/telebruxelles.py b/youtube_dl/extractor/telebruxelles.py
new file mode 100644
index 000000000..39852dd0e
--- /dev/null
+++ b/youtube_dl/extractor/telebruxelles.py
@@ -0,0 +1,52 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+import re
+import json
+
+from .common import InfoExtractor
+
+
+class TeleBruxellesIE(InfoExtractor):
+ _VALID_URL = r'https?://(?:www\.)?telebruxelles\.be/(news|sport|dernier-jt)/?(?P<title>[^\?]+)'
+ _TESTS = [{
+ 'url': r'http://www.telebruxelles.be/news/auditions-devant-parlement-francken-galant-tres-attendus/',
+ 'md5': '59439e568c9ee42fb77588b2096b214f',
+ 'info_dict': {
+ 'id': '11942',
+ 'ext': 'flv',
+ 'title': 're:Parlement : Francken et Galant répondent aux interpellations*',
+ 'description': 're:Les auditions des ministres se poursuivent*'
+ }
+ }, {
+ 'url': r'http://www.telebruxelles.be/sport/basket-brussels-bat-mons-80-74/',
+ 'md5': '181d3fbdcf20b909309e5aef5c6c6047',
+ 'info_dict': {
+ 'id': '10091',
+ 'ext': 'flv',
+ 'title': 'Basket : le Brussels bat Mons 80-74',
+ 'description': 're:Ils l\u2019on fait ! En basket, le B*'
+ }
+ }]
+
+ def _real_extract(self, url):
+ mobj = re.match(self._VALID_URL, url)
+ title = mobj.group('title')
+
+ webpage = self._download_webpage(url, title)
+
+ article_id = self._html_search_regex(r"<article id=\"post-(\d+)\"", webpage, '0')
+ title = self._html_search_regex(r'<h1 class=\"entry-title\">(.*?)</h1>', webpage, 'title')
+ description = self._html_search_regex(r"property=\"og:description\" content=\"(.*?)\"", webpage, 'description', fatal=False)
+
+ rtmp_url = self._html_search_regex(r"file: \"(rtmp://\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}:\d{1,5}/vod/mp4:\" \+ \"\w+\" \+ \".mp4)\"", webpage, 'url')
+ rtmp_url = rtmp_url.replace("\" + \"", "")
+
+ return {
+ 'id': article_id,
+ 'title': title,
+ 'description': description,
+ 'url': rtmp_url,
+ 'ext': 'flv',
+ 'rtmp_live': True # if rtmpdump is not called with "--live" argument, the download is blocked and can be completed
+ } \ No newline at end of file