[pyvideo] add support for videos that don't come from Youtube

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

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index e10c74c112a0b7bbb9335e76cdff88524c4cbafb..56a5d009fc0678a8a80d9fed5e4d0cff770fba06 100644 (file)
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -1,7 +1,6 @@
 # encoding: utf-8
 import re
 import json
-import xml.etree.ElementTree
 
 from .common import InfoExtractor
 from ..utils import (
@@ -10,6 +9,7 @@ from ..utils import (
     unified_strdate,
     determine_ext,
     get_element_by_id,
+    compat_str,
 )
 
 # There are different sources of video in arte.tv, the extraction process 
@@ -17,8 +17,8 @@ from ..utils import (
 # add tests.
 
 class ArteTvIE(InfoExtractor):
-    _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>.+)'
+    _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'
@@ -68,7 +68,7 @@ class ArteTvIE(InfoExtractor):
             lang = mobj.group('lang')
             return self._extract_liveweb(url, name, lang)
 
-        if re.search(self._LIVE_URL, video_id) is not None:
+        if re.search(self._LIVE_URL, url) is not None:
             raise ExtractorError(u'Arte live streams are not yet supported, sorry')
             # self.extractLiveStream(url)
             # return
@@ -77,8 +77,7 @@ class ArteTvIE(InfoExtractor):
         """Extract from videos.arte.tv"""
         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)
+        ref_xml_doc = self._download_xml(ref_xml_url, video_id, note=u'Downloading metadata')
         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')
@@ -108,13 +107,12 @@ class ArteTvIE(InfoExtractor):
         """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,
+        config_doc = self._download_xml('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')
+            url_node = event_doc.find('urlSd')
 
         return {'id': video_id,
                 'title': event_doc.find('name%s' % lang.capitalize()).text,
@@ -181,20 +179,30 @@ class ArteTVPlus7IE(InfoExtractor):
                 formats = all_formats
             else:
                 raise ExtractorError(u'The formats list is empty')
-        # We order the formats by quality
+
         if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
-            sort_key = lambda f: ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
+            def sort_key(f):
+                return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
         else:
-            sort_key = lambda f: int(f.get('height',-1))
+            def sort_key(f):
+                return (
+                    # Sort first by quality
+                    int(f.get('height',-1)),
+                    int(f.get('bitrate',-1)),
+                    # The original version with subtitles has lower relevance
+                    re.match(r'VO-ST(F|A)', f.get('versionCode', '')) is None,
+                    # The version with sourds/mal subtitles has also lower relevance
+                    re.match(r'VO?(F|A)-STM\1', f.get('versionCode', '')) is None,
+                )
         formats = sorted(formats, key=sort_key)
-        # Prefer videos without subtitles in the same language
-        formats = sorted(formats, key=lambda f: re.match(r'VO(F|A)-STM\1', f.get('versionCode', '')) is None)
-        # Pick the best quality
         def _format(format_info):
-            quality = format_info['quality']
-            m_quality = re.match(r'\w*? - (\d*)p', quality)
-            if m_quality is not None:
-                quality = m_quality.group(1)
+            quality = ''
+            height = format_info.get('height')
+            if height is not None:
+                quality = compat_str(height)
+            bitrate = format_info.get('bitrate')
+            if bitrate is not None:
+                quality += '-%d' % bitrate
             if format_info.get('versionCode') is not None:
                 format_id = u'%s-%s' % (quality, format_info['versionCode'])
             else:
@@ -203,7 +211,7 @@ class ArteTVPlus7IE(InfoExtractor):
                 'format_id': format_id,
                 'format_note': format_info.get('versionLibelle'),
                 'width': format_info.get('width'),
-                'height': format_info.get('height'),
+                'height': height,
             }
             if format_info['mediaType'] == u'rtmp':
                 info['url'] = format_info['streamer']