Merge pull request #1310 from rzhxeo/rtlnow

[youtube-dl] / youtube_dl / extractor / rtlnow.py

diff --git a/youtube_dl/extractor/rtlnow.py b/youtube_dl/extractor/rtlnow.py
index 15f01a2e2bd522987dcb45b01c144db9194b8155..7bb236c2b9fe02e998a9cdee38c640110f22b19c 100644 (file)
--- a/youtube_dl/extractor/rtlnow.py
+++ b/youtube_dl/extractor/rtlnow.py
@@ -2,11 +2,14 @@
 import re
 
 from .common import InfoExtractor
-from ..utils import ExtractorError
+from ..utils import (
+    clean_html,
+    ExtractorError,
+)
 
 class RTLnowIE(InfoExtractor):
-    """Information Extractor for RTL(2)now"""
-    _VALID_URL = r'(?:http://)?(?P<url>(?P<base_url>rtl(?:(?P<is_rtl2>2)|-)now\.rtl(?(is_rtl2)2|)\.de/)[a-zA-Z0-9-]+/[a-zA-Z0-9-]+\.php\?(?:container_id|film_id)=(?P<video_id>[0-9]+)&player=1(?:&season=[0-9]+)?(?:&.*)?)'
+    """Information Extractor for RTL NOW, RTL2 NOW, SUPER RTL NOW and VOX NOW"""
+    _VALID_URL = r'(?:http://)?(?P<url>(?P<base_url>rtl-now\.rtl\.de/|rtl2now\.rtl2\.de/|(?:www\.)?voxnow\.de/|(?:www\.)?superrtlnow\.de/)[a-zA-Z0-9-]+/[a-zA-Z0-9-]+\.php\?(?:container_id|film_id)=(?P<video_id>[0-9]+)&player=1(?:&season=[0-9]+)?(?:&.*)?)'
     _TESTS = [{
         u'url': u'http://rtl-now.rtl.de/ahornallee/folge-1.php?film_id=90419&player=1&season=1',
         u'file': u'90419.flv',
@@ -18,6 +21,7 @@ class RTLnowIE(InfoExtractor):
         u'params': {
             u'skip_download': True,
         },
+        u'skip': u'Only works from Germany',
     },
     {
         u'url': u'http://rtl2now.rtl2.de/aerger-im-revier/episode-15-teil-1.php?film_id=69756&player=1&season=2&index=5',
@@ -31,7 +35,33 @@ class RTLnowIE(InfoExtractor):
         u'params': {
             u'skip_download': True,
         },
-    },]
+        u'skip': u'Only works from Germany',
+    },
+    {
+        u'url': u'www.voxnow.de/voxtours/suedafrika-reporter-ii.php?film_id=13883&player=1&season=17',
+        u'file': u'13883.flv',
+        u'info_dict': {
+            u'upload_date': u'20090627', 
+            u'title': u'Voxtours - Südafrika-Reporter II',
+            u'description': u'Südafrika-Reporter II',
+        },
+        u'params': {
+            u'skip_download': True,
+        },
+    },
+    {
+        u'url': u'http://superrtlnow.de/medicopter-117/angst.php?film_id=99205&player=1',
+        u'file': u'99205.flv',
+        u'info_dict': {
+            u'upload_date': u'20080928', 
+            u'title': u'Medicopter 117 - Angst!',
+            u'description': u'Angst!',
+            u'thumbnail': u'http://autoimg.static-fra.de/superrtlnow/287529/1500x1500/image2.jpg'
+        },
+        u'params': {
+            u'skip_download': True,
+        },
+    }]
 
     def _real_extract(self,url):
         mobj = re.match(self._VALID_URL, url)
@@ -41,6 +71,14 @@ class RTLnowIE(InfoExtractor):
         video_id = mobj.group(u'video_id')
 
         webpage = self._download_webpage(webpage_url, video_id)
+
+        note_m = re.search(r'''(?sx)
+            <div[ ]style="margin-left:[ ]20px;[ ]font-size:[ ]13px;">(.*?)
+            <div[ ]id="playerteaser">''', webpage)
+        if note_m:
+            msg = clean_html(note_m.group(1))
+            raise ExtractorError(msg)
+
         video_title = self._html_search_regex(r'<title>(?P<title>[^<]+)</title>',
             webpage, u'title')
         playerdata_url = self._html_search_regex(r'\'playerdata\': \'(?P<playerdata_url>[^\']+)\'',