Merge branch 'master' into rtmpdump

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

diff --git a/youtube_dl/extractor/arte.py b/youtube_dl/extractor/arte.py
index 5ee8a67b14699a330914cd4f0e0f627ca9fca5a5..44d0b5d708aa31aef6d9997321142cb8638ab56b 100644 (file)
--- a/youtube_dl/extractor/arte.py
+++ b/youtube_dl/extractor/arte.py
@@ -10,6 +10,7 @@ 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 
@@ -68,7 +69,7 @@ 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
@@ -114,7 +115,7 @@ class ArteTvIE(InfoExtractor):
         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,
@@ -158,7 +159,9 @@ class ArteTVPlus7IE(InfoExtractor):
             'thumbnail': player_info.get('programImage') or player_info.get('VTU', {}).get('IUR'),
         }
 
-        formats = player_info['VSR'].values()
+        all_formats = player_info['VSR'].values()
+        # Some formats use the m3u8 protocol
+        all_formats = list(filter(lambda f: f.get('videoFormat') != 'M3U8', all_formats))
         def _match_lang(f):
             if f.get('versionCode') is None:
                 return True
@@ -170,18 +173,48 @@ class ArteTVPlus7IE(InfoExtractor):
             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)
-        # Some formats use the m3u8 protocol
-        formats = filter(lambda f: f.get('videoFormat') != 'M3U8', formats)
-        # We order the formats by quality
-        formats = sorted(formats, key=lambda f: int(f.get('height',-1)))
-        # 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
+        formats = filter(_match_lang, all_formats)
+        formats = list(formats) # in python3 filter returns an iterator
+        if not formats:
+            # Some videos are only available in the 'Originalversion'
+            # they aren't tagged as being in French or German
+            if all(f['versionCode'] == 'VO' for f in all_formats):
+                formats = all_formats
+            else:
+                raise ExtractorError(u'The formats list is empty')
+
+        if re.match(r'[A-Z]Q', formats[0]['quality']) is not None:
+            def sort_key(f):
+                return ['HQ', 'MQ', 'EQ', 'SQ'].index(f['quality'])
+        else:
+            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)
         def _format(format_info):
+            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:
+                format_id = quality
             info = {
+                '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']
@@ -192,8 +225,6 @@ class ArteTVPlus7IE(InfoExtractor):
                 info['ext'] = determine_ext(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
 
@@ -207,7 +238,7 @@ class ArteTVCreativeIE(ArteTVPlus7IE):
         u'url': u'http://creative.arte.tv/de/magazin/agentur-amateur-corporate-design',
         u'file': u'050489-002.mp4',
         u'info_dict': {
-            u'title': u'Agentur Amateur #2 - Corporate Design',
+            u'title': u'Agentur Amateur / Agence Amateur #2 : Corporate Design',
         },
     }