ArteTvIE: support videos from videos.arte.tv

Each source of videos have a different extraction process, they are in different methods of the extractor.
Changed the extension of videos from mp4 to flv.

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index b061b9566168758465ad56f43b1f74b89b2cce10..183274eb75f0947d11434867f30ba3ef9f6fcc0c 100644 (file)
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -11,11 +11,21 @@ from ..utils import (
 )
 
 class ArteTvIE(InfoExtractor):
-    _VALID_URL = r'(?:http://)?www\.arte.tv/guide/(?:fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
+    """
+    There are two sources of video in arte.tv: videos.arte.tv and
+    www.arte.tv/guide, the extraction process is different for each one.
+    The videos expire in 7 days, so we can't add tests.
+    """
+    _EMISSION_URL = r'(?:http://)?www\.arte.tv/guide/(?:fr|de)/(?:(?:sendungen|emissions)/)?(?P<id>.*?)/(?P<name>.*?)(\?.*)?'
+    _VIDEOS_URL = r'(?:http://)?videos.arte.tv/(?:fr|de)/.*-(?P<id>.*?).html'
     _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))
+
     # TODO implement Live Stream
     # def extractLiveStream(self, url):
     #     video_lang = url.split('/')[-4]
@@ -44,17 +54,26 @@ class ArteTvIE(InfoExtractor):
     #     video_url = u'%s/%s' % (info.get('url'), info.get('path'))
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        name = mobj.group('name')
-        # This is not a real id, it can be for example AJT for the news
-        # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
-        video_id = mobj.group('id')
+        mobj = re.match(self._EMISSION_URL, url)
+        if mobj is not None:
+            name = mobj.group('name')
+            # This is not a real id, it can be for example AJT for the news
+            # http://www.arte.tv/guide/fr/emissions/AJT/arte-journal
+            video_id = mobj.group('id')
+            return self._extract_emission(url, video_id)
+
+        mobj = re.match(self._VIDEOS_URL, url)
+        if mobj is not None:
+            id = mobj.group('id')
+            return self._extract_video(url, id)
 
         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)
             # return
 
+    def _extract_emission(self, url, video_id):
+        """Extract from www.arte.tv/guide"""
         webpage = self._download_webpage(url, video_id)
         json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
 
@@ -68,6 +87,7 @@ class ArteTvIE(InfoExtractor):
                      'description': player_info['VDE'],
                      'upload_date': unified_strdate(player_info['VDA'].split(' ')[0]),
                      'thumbnail': player_info['programImage'],
+                     'ext': 'flv',
                      }
 
         formats = player_info['VSR'].values()
@@ -78,9 +98,36 @@ class ArteTvIE(InfoExtractor):
         if format_info['mediaType'] == u'rtmp':
             info_dict['url'] = format_info['streamer']
             info_dict['play_path'] = 'mp4:' + format_info['url']
-            info_dict['ext'] = 'mp4'
         else:
             info_dict['url'] = format_info['url']
-            info_dict['ext'] = 'mp4'
 
         return info_dict
+
+    def _extract_video(self, url, video_id):
+        """Extract from videos.arte.tv"""
+        config_xml_url = url.replace('/videos/', '/do_delegate/videos/')
+        config_xml_url = config_xml_url.replace('.html', ',view,asPlayerXml.xml')
+        config_xml = self._download_webpage(config_xml_url, video_id)
+        config_xml_url = self._html_search_regex(r'<video lang=".*?" ref="(.*?)"', config_xml, 'config xml url')
+        config_xml = self._download_webpage(config_xml_url, video_id)
+
+        video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
+        def _key(m):
+            quality = m.group('quality')
+            if quality == 'hd':
+                return 2
+            else:
+                return 1
+        # We pick the best quality
+        video_urls = sorted(video_urls, key=_key)
+        video_url = list(video_urls)[-1].group('url')
+        
+        title = self._html_search_regex(r'<name>(.*?)</name>', config_xml, 'title')
+        thumbnail = self._html_search_regex(r'<firstThumbnailUrl>(.*?)</firstThumbnailUrl>',
+                                            config_xml, 'thumbnail')
+        return {'id': video_id,
+                'title': title,
+                'thumbnail': thumbnail,
+                'url': video_url,
+                'ext': 'flv',
+                }