[newstube] fix extraction

youtube_dl/extractor/newstube.py

@@ -1,12 +1,17 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
-import re
+import base64
+import hashlib
 
 from .common import InfoExtractor
+from ..aes import aes_cbc_decrypt
 from ..utils import (
-    ExtractorError,
+    bytes_to_intlist,
     int_or_none,
+    intlist_to_bytes,
+    parse_codecs,
+    parse_duration,
 )
 
 
@@ -14,7 +19,7 @@ class NewstubeIE(InfoExtractor):
     _VALID_URL = r'https?://(?:www\.)?newstube\.ru/media/(?P<id>.+)'
     _TEST = {
         'url': 'http://www.newstube.ru/media/telekanal-cnn-peremestil-gorod-slavyansk-v-krym',
-        'md5': '801eef0c2a9f4089fa04e4fe3533abdc',
+        'md5': '9d10320ad473444352f72f746ccb8b8c',
         'info_dict': {
             'id': '728e0ef2-e187-4012-bac0-5a081fdcb1f6',
             'ext': 'mp4',
@@ -25,84 +30,45 @@ class NewstubeIE(InfoExtractor):
     }
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
+        video_id = self._match_id(url)
-        video_id = mobj.group('id')
 
-        page = self._download_webpage(url, video_id, 'Downloading page')
+        page = self._download_webpage(url, video_id)
+        title = self._html_search_meta(['og:title', 'twitter:title'], page, fatal=True)
 
         video_guid = self._html_search_regex(
-            r'<meta property="og:video:url" content="https?://(?:www\.)?newstube\.ru/freshplayer\.swf\?guid=(?P<guid>[\da-f]{8}-[\da-f]{4}-[\da-f]{4}-[\da-f]{4}-[\da-f]{12})',
+            r'<meta\s+property="og:video(?::(?:(?:secure_)?url|iframe))?"\s+content="https?://(?:www\.)?newstube\.ru/embed/(?P<guid>[\da-f]{8}-[\da-f]{4}-[\da-f]{4}-[\da-f]{4}-[\da-f]{12})',
             page, 'video GUID')
 
-        player = self._download_xml(
+        enc_data = base64.b64decode(self._download_webpage(
-            'http://p.newstube.ru/v2/player.asmx/GetAutoPlayInfo6?state=&url=%s&sessionId=&id=%s&placement=profile&location=n2' % (url, video_guid),
+            'https://www.newstube.ru/embed/api/player/getsources2',
-            video_guid, 'Downloading player XML')
+            video_guid, query={
-
+                'guid': video_guid,
-        def ns(s):
+                'ff': 3,
-            return s.replace('/', '/%(ns)s') % {'ns': '{http://app1.newstube.ru/N2SiteWS/player.asmx}'}
+            }))
-
+        key = hashlib.pbkdf2_hmac(
-        error_message = player.find(ns('./ErrorMessage'))
+            'sha1', video_guid.replace('-', '').encode(), enc_data[:16], 1)[:16]
-        if error_message is not None:
+        dec_data = aes_cbc_decrypt(
-            raise ExtractorError('%s returned error: %s' % (self.IE_NAME, error_message.text), expected=True)
+            bytes_to_intlist(enc_data[32:]), bytes_to_intlist(key),
-
+            bytes_to_intlist(enc_data[16:32]))
-        session_id = player.find(ns('./SessionId')).text
+        sources = self._parse_json(intlist_to_bytes(dec_data[:-dec_data[-1]]), video_guid)
-        media_info = player.find(ns('./Medias/MediaInfo'))
-        title = media_info.find(ns('./Name')).text
-        description = self._og_search_description(page)
-        thumbnail = media_info.find(ns('./KeyFrame')).text
-        duration = int(media_info.find(ns('./Duration')).text) / 1000.0
 
         formats = []
-
+        for source in sources:
-        for stream_info in media_info.findall(ns('./Streams/StreamInfo')):
+            source_url = source.get('Src')
-            media_location = stream_info.find(ns('./MediaLocation'))
+            if not source_url:
-            if media_location is None:
                 continue
-
+            height = int_or_none(source.get('Height'))
-            server = media_location.find(ns('./Server')).text
+            f = {
-            app = media_location.find(ns('./App')).text
+                'format_id': 'http' + ('-%dp' % height if height else ''),
-            media_id = stream_info.find(ns('./Id')).text
+                'url': source_url,
-            name = stream_info.find(ns('./Name')).text
+                'width': int_or_none(source.get('Width')),
-            width = int(stream_info.find(ns('./Width')).text)
-            height = int(stream_info.find(ns('./Height')).text)
-
-            formats.append({
-                'url': 'rtmp://%s/%s' % (server, app),
-                'app': app,
-                'play_path': '01/%s' % video_guid.upper(),
-                'rtmp_conn': ['S:%s' % session_id, 'S:%s' % media_id, 'S:n2'],
-                'page_url': url,
-                'ext': 'flv',
-                'format_id': 'rtmp' + ('-%s' % name if name else ''),
-                'width': width,
                 'height': height,
-            })
+            }
-
+            source_type = source.get('Type')
-        sources_data = self._download_json(
+            if source_type:
-            'http://www.newstube.ru/player2/getsources?guid=%s' % video_guid,
+                f.update(parse_codecs(self._search_regex(
-            video_guid, fatal=False)
+                    r'codecs="([^"]+)"', source_type, 'codecs', fatal=False)))
-        if sources_data:
+            formats.append(f)
-            for source in sources_data.get('Sources', []):
-                source_url = source.get('Src')
-                if not source_url:
-                    continue
-                height = int_or_none(source.get('Height'))
-                f = {
-                    'format_id': 'http' + ('-%dp' % height if height else ''),
-                    'url': source_url,
-                    'width': int_or_none(source.get('Width')),
-                    'height': height,
-                }
-                source_type = source.get('Type')
-                if source_type:
-                    mobj = re.search(r'codecs="([^,]+),\s*([^"]+)"', source_type)
-                    if mobj:
-                        vcodec, acodec = mobj.groups()
-                        f.update({
-                            'vcodec': vcodec,
-                            'acodec': acodec,
-                        })
-                formats.append(f)
 
         self._check_formats(formats, video_guid)
         self._sort_formats(formats)
@@ -110,8 +76,8 @@ def ns(s):
         return {
             'id': video_guid,
             'title': title,
-            'description': description,
+            'description': self._html_search_meta(['description', 'og:description'], page),
-            'thumbnail': thumbnail,
+            'thumbnail': self._html_search_meta(['og:image:secure_url', 'og:image', 'twitter:image'], page),
-            'duration': duration,
+            'duration': parse_duration(self._html_search_meta('duration', page)),
             'formats': formats,
         }