Merge remote-tracking branch 'Rudloff/websurg'

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

diff --git a/youtube_dl/extractor/vevo.py b/youtube_dl/extractor/vevo.py
index 49a249ae32a901ea3bc5be431b2a510a09fc95bc..1c1cc418d29a8897e2a2825492ed7becab75af6b 100644 (file)
--- a/youtube_dl/extractor/vevo.py
+++ b/youtube_dl/extractor/vevo.py
@@ -1,43 +1,77 @@
 import re
 import json
+import xml.etree.ElementTree
+import datetime
 
 from .common import InfoExtractor
 from ..utils import (
+    determine_ext,
     ExtractorError,
 )
 
+
 class VevoIE(InfoExtractor):
     """
-    Accecps urls from vevo.com or in the format 'vevo:{id}'
+    Accepts urls from vevo.com or in the format 'vevo:{id}'
     (currently used by MTVIE)
     """
-    _VALID_URL = r'((http://www.vevo.com/watch/.*?/.*?/)|(vevo:))(?P<id>.*)$'
+    _VALID_URL = r'((http://www.vevo.com/watch/.*?/.*?/)|(vevo:))(?P<id>.*?)(\?|$)'
+    _TEST = {
+        u'url': u'http://www.vevo.com/watch/hurts/somebody-to-die-for/GB1101300280',
+        u'file': u'GB1101300280.mp4',
+        u'info_dict': {
+            u"upload_date": u"20130624",
+            u"uploader": u"Hurts",
+            u"title": u"Somebody to Die For",
+            u'duration': 230,
+        }
+    }
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
         video_id = mobj.group('id')
 
-        json_url = 'http://www.vevo.com/data/video/%s' % video_id
-        base_url = 'http://smil.lvl3.vevo.com'
-        videos_url = '%s/Video/V2/VFILE/%s/%sr.smil' % (base_url, video_id, video_id.lower())
+        json_url = 'http://videoplayer.vevo.com/VideoService/AuthenticateVideo?isrc=%s' % video_id
         info_json = self._download_webpage(json_url, video_id, u'Downloading json info')
-        links_webpage = self._download_webpage(videos_url, video_id, u'Downloading videos urls')
 
         self.report_extraction(video_id)
-        video_info = json.loads(info_json)
-        m_urls = list(re.finditer(r'<video src="(?P<ext>.*?):(?P<url>.*?)"', links_webpage))
-        if m_urls is None or len(m_urls) == 0:
-            raise ExtractorError(u'Unable to extract video url')
-        # They are sorted from worst to best quality
-        m_url = m_urls[-1]
-        video_url = base_url + m_url.group('url')
-        ext = m_url.group('ext')
-
-        return {'url': video_url,
-                'ext': ext,
-                'id': video_id,
-                'title': video_info['title'],
-                'thumbnail': video_info['img'],
-                'upload_date': video_info['launchDate'].replace('/',''),
-                'uploader': video_info['Artists'][0]['title'],
-                }
+        video_info = json.loads(info_json)['video']
+        last_version = {'version': -1}
+        for version in video_info['videoVersions']:
+            # These are the HTTP downloads, other types are for different manifests
+            if version['sourceType'] == 2:
+                if version['version'] > last_version['version']:
+                    last_version = version
+        if last_version['version'] == -1:
+            raise ExtractorError(u'Unable to extract last version of the video')
+
+        renditions = xml.etree.ElementTree.fromstring(last_version['data'])
+        formats = []
+        # Already sorted from worst to best quality
+        for rend in renditions.findall('rendition'):
+            attr = rend.attrib
+            f_url = attr['url']
+            formats.append({
+                'url': f_url,
+                'ext': determine_ext(f_url),
+                'height': int(attr['frameheight']),
+                'width': int(attr['frameWidth']),
+            })
+
+        date_epoch = int(self._search_regex(
+            r'/Date\((\d+)\)/', video_info['launchDate'], u'launch date'))/1000
+        upload_date = datetime.datetime.fromtimestamp(date_epoch)
+        info = {
+            'id': video_id,
+            'title': video_info['title'],
+            'formats': formats,
+            'thumbnail': video_info['imageUrl'],
+            'upload_date': upload_date.strftime('%Y%m%d'),
+            'uploader': video_info['mainArtists'][0]['artistName'],
+            'duration': video_info['duration'],
+        }
+
+        # TODO: Remove when #980 has been merged
+        info.update(formats[-1])
+
+        return info