X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Farte.py;h=69b3b0ad7820600ef5107ad3d79230c0e4edcaac;hb=0f90943e455fd0e4f323fa6d43092060c8b1db61;hp=a030a28bb2da0c989fb2ea461c3ccc6279cb675e;hpb=9826925a207628deef3b5c8511ec29dffdfefda5;p=youtube-dl

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index a030a28bb..69b3b0ad7 100644
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -1,12 +1,11 @@
 import re
 import json
+import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from ..utils import (
-    # This is used by the not implemented extractLiveStream method
-    compat_urllib_parse,
-
     ExtractorError,
+    find_xpath_attr,
     unified_strdate,
 )
 
@@ -17,16 +16,18 @@ class ArteTvIE(InfoExtractor):
     The videos expire in 7 days, so we can't add tests.
     """
     _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/(?:fr|de)/.*-(?P<id>.*?).html'
+    _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
     # def extractLiveStream(self, url):
     #     video_lang = url.split('/')[-4]
     #     info = self.grep_webpage(
@@ -56,7 +57,6 @@ class ArteTvIE(InfoExtractor):
     def _real_extract(self, url):
         mobj = re.match(self._EMISSION_URL, url)
         if mobj is not None:
-            name = mobj.group('name')
             lang = mobj.group('lang')
             # 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
@@ -66,7 +66,14 @@ class ArteTvIE(InfoExtractor):
         mobj = re.match(self._VIDEOS_URL, url)
         if mobj is not None:
             id = mobj.group('id')
-            return self._extract_video(url, id)
+            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')
@@ -75,7 +82,8 @@ class ArteTvIE(InfoExtractor):
 
     def _extract_emission(self, url, video_id, lang):
         """Extract from www.arte.tv/guide"""
-        json_url = 'http://org-www.arte.tv/papi/tvguide/videos/stream/player/F/%s_PLUS7-F/ALL/ALL.json' % video_id
+        webpage = self._download_webpage(url, video_id)
+        json_url = self._html_search_regex(r'arte_vp_url="(.*?)"', webpage, 'json url')
 
         json_info = self._download_webpage(json_url, video_id, 'Downloading info json')
         self.report_extraction(video_id)
@@ -84,7 +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',
@@ -97,12 +105,14 @@ class ArteTvIE(InfoExtractor):
                 l = 'F'
             elif lang == 'de':
                 l = 'A'
-            regexes = [r'VO?%s' % l, r'V%s-ST.' % l]
+            regexes = [r'VO?%s' % l, r'VO?.-ST%s' % l]
             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)
         # 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':
@@ -113,13 +123,15 @@ class ArteTvIE(InfoExtractor):
 
         return info_dict
 
-    def _extract_video(self, url, video_id):
+    def _extract_video(self, url, video_id, lang):
         """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)
+        ref_xml_url = url.replace('/videos/', '/do_delegate/videos/')
+        ref_xml_url = ref_xml_url.replace('.html', ',view,asPlayerXml.xml')
+        ref_xml = self._download_webpage(ref_xml_url, video_id, note=u'Downloading metadata')
+        ref_xml_doc = xml.etree.ElementTree.fromstring(ref_xml)
+        config_node = find_xpath_attr(ref_xml_doc, './/video', 'lang', lang)
+        config_xml_url = config_node.attrib['ref']
+        config_xml = self._download_webpage(config_xml_url, video_id, note=u'Downloading configuration')
 
         video_urls = list(re.finditer(r'<url quality="(?P<quality>.*?)">(?P<url>.*?)</url>', config_xml))
         def _key(m):
@@ -141,3 +153,22 @@ 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),
+                }