[arte] Prepare for generic format support (#980)

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

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index 18d5916589b239c9cf12a7caa629efeea784ce86..4707d7ccab51502dadf787ab2a2fb1558a1c9d45 100644 (file)
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -17,13 +17,14 @@ class ArteTvIE(InfoExtractor):
     """
     _EMISSION_URL = r'(?:http://)?www\.arte.tv/guide/(?P<lang>fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
     _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?P<lang>fr|de)/.*-(?P<id>.*?).html'
+    _LIVEWEB_URL = r'(?:http://)?liveweb.arte.tv/(?P<lang>fr|de)/(?P<subpage>.+?)/(?P<name>.+)'
     _LIVE_URL = r'index-[0-9]+\.html$'
 
     IE_NAME = u'arte.tv'
 
     @classmethod
     def suitable(cls, url):
-        return any(re.match(regex, url) for regex in (cls._EMISSION_URL, cls._VIDEOS_URL))
+        return any(re.match(regex, url) for regex in (cls._EMISSION_URL, cls._VIDEOS_URL, cls._LIVEWEB_URL))
 
     # TODO implement Live Stream
     # from ..utils import compat_urllib_parse
@@ -68,6 +69,12 @@ class ArteTvIE(InfoExtractor):
             lang = mobj.group('lang')
             return self._extract_video(url, id, lang)
 
+        mobj = re.match(self._LIVEWEB_URL, url)
+        if mobj is not None:
+            name = mobj.group('name')
+            lang = mobj.group('lang')
+            return self._extract_liveweb(url, name, lang)
+
         if re.search(self._LIVE_URL, video_id) is not None:
             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
             # self.extractLiveStream(url)
@@ -85,7 +92,7 @@ class ArteTvIE(InfoExtractor):
 
         info_dict = {'id': player_info['VID'],
                      'title': player_info['VTI'],
-                     'description': player_info['VDE'],
+                     'description': player_info.get('VDE'),
                      'upload_date': unified_strdate(player_info['VDA'].split(' ')[0]),
                      'thumbnail': player_info['programImage'],
                      'ext': 'flv',
@@ -102,15 +109,27 @@ class ArteTvIE(InfoExtractor):
             return any(re.match(r, f['versionCode']) for r in regexes)
         # Some formats may not be in the same language as the url
         formats = filter(_match_lang, formats)
+        # Some formats use the m3u8 protocol
+        formats = filter(lambda f: f['videoFormat'] != 'M3U8', formats)
         # We order the formats by quality
         formats = sorted(formats, key=lambda f: int(f['height']))
+        # Prefer videos without subtitles in the same language
+        formats = sorted(formats, key=lambda f: re.match(r'VO(F|A)-STM\1', f['versionCode']) is None)
         # Pick the best quality
-        format_info = formats[-1]
-        if format_info['mediaType'] == u'rtmp':
-            info_dict['url'] = format_info['streamer']
-            info_dict['play_path'] = 'mp4:' + format_info['url']
-        else:
-            info_dict['url'] = format_info['url']
+        def _format(format_info):
+            info = {'ext': 'flv',
+                    'width': format_info.get('width'),
+                    'height': format_info.get('height'),
+                    }
+            if format_info['mediaType'] == u'rtmp':
+                info['url'] = format_info['streamer']
+                info['play_path'] = 'mp4:' + format_info['url']
+            else:
+                info_dict['url'] = format_info['url']
+            return info
+        info_dict['formats'] = [_format(f) for f in formats]
+        # TODO: Remove when #980 has been merged 
+        info_dict.update(info_dict['formats'][-1])
 
         return info_dict
 
@@ -144,3 +163,22 @@ class ArteTvIE(InfoExtractor):
                 'url': video_url,
                 'ext': 'flv',
                 }
+
+    def _extract_liveweb(self, url, name, lang):
+        """Extract form http://liveweb.arte.tv/"""
+        webpage = self._download_webpage(url, name)
+        video_id = self._search_regex(r'eventId=(\d+?)("|&)', webpage, u'event id')
+        config_xml = self._download_webpage('http://download.liveweb.arte.tv/o21/liveweb/events/event-%s.xml' % video_id,
+                                            video_id, u'Downloading information')
+        config_doc = xml.etree.ElementTree.fromstring(config_xml.encode('utf-8'))
+        event_doc = config_doc.find('event')
+        url_node = event_doc.find('video').find('urlHd')
+        if url_node is None:
+            url_node = video_doc.find('urlSd')
+
+        return {'id': video_id,
+                'title': event_doc.find('name%s' % lang.capitalize()).text,
+                'url': url_node.text.replace('MP4', 'mp4'),
+                'ext': 'flv',
+                'thumbnail': self._og_search_thumbnail(webpage),
+                }