[servingsys] Add support

This also adds support for brightcove advertisements.
Fixes #2181

youtube_dl/YoutubeDL.py

@@ -151,6 +151,7 @@ class YoutubeDL(object):
     bidi_workaround:   Work around buggy terminals without bidirectional text
                        support, using fridibi
     debug_printtraffic:Print out sent and received HTTP traffic
+    include_ads:       Download ads as well
 
     The following parameters are not used by YoutubeDL itself, they are used by
     the FileDownloader:

youtube_dl/__init__.py

@@ -238,7 +238,10 @@ def _hide_login_info(opts):
     selection.add_option('--download-archive', metavar='FILE',
                          dest='download_archive',
                          help='Download only videos not listed in the archive file. Record the IDs of all downloaded videos in it.')
-
+    selection.add_option(
+        '--include-ads', dest='include_ads',
+        action='store_true',
+        help='Download advertisements as well (experimental)')
 
     authentication.add_option('-u', '--username',
             dest='username', metavar='USERNAME', help='account username')
@@ -716,6 +719,7 @@ def _real_main(argv=None):
         'bidi_workaround': opts.bidi_workaround,
         'debug_printtraffic': opts.debug_printtraffic,
         'prefer_ffmpeg': opts.prefer_ffmpeg,
+        'include_ads': opts.include_ads,
     }
 
     with YoutubeDL(ydl_opts) as ydl:

youtube_dl/extractor/__init__.py

@@ -152,6 +152,7 @@
 from .roxwel import RoxwelIE
 from .rtlnow import RTLnowIE
 from .rutube import RutubeIE
+from .servingsys import ServingSysIE
 from .sina import SinaIE
 from .slashdot import SlashdotIE
 from .slideshare import SlideshareIE

youtube_dl/extractor/brightcove.py

@@ -9,9 +9,11 @@
 from ..utils import (
     compat_urllib_parse,
     find_xpath_attr,
+    fix_xml_ampersands,
     compat_urlparse,
     compat_str,
     compat_urllib_request,
+    compat_parse_qs,
 
     ExtractorError,
     unsmuggle_url,
@@ -83,17 +85,30 @@ def _build_brighcove_url(cls, object_str):
                             lambda m: m.group(1) + '/>', object_str)
         # Fix up some stupid XML, see https://github.com/rg3/youtube-dl/issues/1608
         object_str = object_str.replace('<--', '<!--')
+        object_str = fix_xml_ampersands(object_str)
 
         object_doc = xml.etree.ElementTree.fromstring(object_str)
-        assert 'BrightcoveExperience' in object_doc.attrib['class']
-        params = {
-            'playerID': find_xpath_attr(object_doc, './param', 'name', 'playerID').attrib['value'],
-        }
+
+        fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
+        flashvars = dict(
+            (k, v[0])
+            for k, v in compat_parse_qs(fv_el.attrib['value']).items())
+
         def find_param(name):
+            if name in flashvars:
+                return flashvars[name]
             node = find_xpath_attr(object_doc, './param', 'name', name)
             if node is not None:
                 return node.attrib['value']
             return None
+
+        params = {}
+
+        playerID = find_param('playerID')
+        if playerID is None:
+            raise ExtractorError('Cannot find player ID')
+        params['playerID'] = playerID
+
         playerKey = find_param('playerKey')
         # Not all pages define this value
         if playerKey is not None:
@@ -114,8 +129,12 @@ def _extract_brightcove_url(cls, webpage):
         if it can't be found
         """
         m_brightcove = re.search(
-            r'<object[^>]+?class=([\'"])[^>]*?BrightcoveExperience.*?\1.+?</object>',
-            webpage, re.DOTALL)
+            r'''(?sx)<object
+            (?:
+                :[^>]+?class=([\'"])[^>]*?BrightcoveExperience.*?\1 |
+                [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
+            ).+?</object>''',
+            webpage)
         if m_brightcove is not None:
             return cls._build_brighcove_url(m_brightcove.group())
         else:
@@ -156,6 +175,7 @@ def _get_video_info(self, video_id, query_str, query, referer=None):
         info = self._search_regex(r'var experienceJSON = ({.*?});', webpage, 'json')
         info = json.loads(info)['data']
         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
+        video_info['_youtubedl_adServerURL'] = info.get('adServerURL')
 
         return self._extract_video_info(video_info)
 
@@ -193,6 +213,23 @@ def _extract_video_info(self, video_info):
             info.update({
                 'url': video_info['FLVFullLengthURL'],
             })
+
+        if self._downloader.params.get('include_ads', False):
+            adServerURL = video_info.get('_youtubedl_adServerURL')
+            if adServerURL:
+                ad_info = {
+                    '_type': 'url',
+                    'url': adServerURL,
+                }
+                if 'url' in info:
+                    return {
+                        '_type': 'playlist',
+                        'title': info['title'],
+                        'entries': [ad_info, info],
+                    }
                 else:
+                    return ad_info
+
+        if 'url' not in info:
             raise ExtractorError('Unable to extract video url for %s' % info['id'])
         return info

youtube_dl/extractor/servingsys.py

@@ -0,0 +1,70 @@
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+    int_or_none,
+)
+
+
+class ServingSysIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:[^.]+\.)?serving-sys\.com/BurstingPipe/adServer\.bs\?.*?&pli=(?P<id>[0-9]+)'
+
+    _TEST = {
+        'url': 'http://bs.serving-sys.com/BurstingPipe/adServer.bs?cn=is&c=23&pl=VAST&pli=5349193&PluID=0&pos=7135&ord=[timestamp]&cim=1?',
+        'playlist': [{
+            'file': '29955898.flv',
+            'md5': 'baed851342df6846eb8677a60a011a0f',
+            'info_dict': {
+                'title': 'AdAPPter_Hyundai_demo (1)',
+                'duration': 74,
+                'tbr': 1378,
+                'width': 640,
+                'height': 400,
+            },
+        }, {
+            'file': '29907998.flv',
+            'md5': '979b4da2655c4bc2d81aeb915a8c5014',
+            'info_dict': {
+                'title': 'AdAPPter_Hyundai_demo (2)',
+                'duration': 34,
+                'width': 854,
+                'height': 480,
+                'tbr': 516,
+            },
+        }],
+        'params': {
+            'playlistend': 2,
+        }
+    }
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        pl_id = mobj.group('id')
+
+        vast_doc = self._download_xml(url, pl_id)
+        title = vast_doc.find('.//AdTitle').text
+        media = vast_doc.find('.//MediaFile').text
+        info_url = self._search_regex(r'&adData=([^&]+)&', media, 'info URL')
+
+        doc = self._download_xml(info_url, pl_id, 'Downloading video info')
+        entries = [{
+            '_type': 'video',
+            'id': a.attrib['id'],
+            'title': '%s (%s)' % (title, a.attrib['assetID']),
+            'url': a.attrib['URL'],
+            'duration': int_or_none(a.attrib.get('length')),
+            'tbr': int_or_none(a.attrib.get('bitrate')),
+            'height': int_or_none(a.attrib.get('height')),
+            'width': int_or_none(a.attrib.get('width')),
+        } for a in doc.findall('.//AdditionalAssets/asset')]
+
+        return {
+            '_type': 'playlist',
+            'id': pl_id,
+            'title': title,
+            'entries': entries,
+        }
+
+