GenericIE: Detect videos from Brightcove

Brightcove videos info is usually found in an <object class="BrightcoveExperience"></object> node, this is passed to a new method of BrightcoveIE that builds a url to extract the video.

youtube_dl/extractor/brightcove.py

@@ -1,17 +1,44 @@
 import re
 import json
+import xml.etree.ElementTree
 
 from .common import InfoExtractor
+from ..utils import (
+    compat_urllib_parse,
+)
 
 class BrightcoveIE(InfoExtractor):
     _VALID_URL = r'http://.*brightcove\.com/.*\?(?P<query>.*videoPlayer=(?P<id>\d*).*)'
+    _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
+    
+    # There is a test for Brigtcove in GenericIE, that way we test both the download
+    # and the detection of videos, and we don't have to find an URL that is always valid
+
+    @classmethod
+    def _build_brighcove_url(cls, object_str):
+        """
+        Build a Brightcove url from a xml string containing
+        <object class="BrightcoveExperience">{params}</object>
+        """
+        object_doc = xml.etree.ElementTree.fromstring(object_str)
+        assert object_doc.attrib['class'] == u'BrightcoveExperience'
+        params = {'flashID': object_doc.attrib['id'],
+                  'playerID': object_doc.find('./param[@name="playerID"]').attrib['value'],
+                  '@videoPlayer': object_doc.find('./param[@name="@videoPlayer"]').attrib['value'],
+                  }
+        playerKey = object_doc.find('./param[@name="playerKey"]')
+        # Not all pages define this value
+        if playerKey is not None:
+            params['playerKey'] = playerKey.attrib['value']
+        data = compat_urllib_parse.urlencode(params)
+        return cls._FEDERATED_URL_TEMPLATE % data
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
         query = mobj.group('query')
         video_id = mobj.group('id')
 
-        request_url = 'http://c.brightcove.com/services/viewer/htmlFederated?%s' % query
+        request_url = self._FEDERATED_URL_TEMPLATE % query
         webpage = self._download_webpage(request_url, video_id)
 
         self.report_extraction(video_id)

youtube_dl/extractor/generic.py

@@ -1,3 +1,5 @@
+# encoding: utf-8
+
 import os
 import re
 
@@ -9,12 +11,14 @@ from ..utils import (
 
     ExtractorError,
 )
+from .brightcove import BrightcoveIE
 
 class GenericIE(InfoExtractor):
     IE_DESC = u'Generic downloader that works on some sites'
     _VALID_URL = r'.*'
     IE_NAME = u'generic'
-    _TEST = {
+    _TESTS = [
+        {
             u'url': u'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
             u'file': u'13601338388002.mp4',
             u'md5': u'85b90ccc9d73b4acd9138d3af4c27f89',
@@ -22,7 +26,19 @@ class GenericIE(InfoExtractor):
                 u"uploader": u"www.hodiho.fr", 
                 u"title": u"R\u00e9gis plante sa Jeep"
             }
+        },
+        {
+            u'url': u'http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/',
+            u'file': u'2371591881001.mp4',
+            u'md5': u'9e80619e0a94663f0bdc849b4566af19',
+            u'note': u'Test Brightcove downloads and detection in GenericIE',
+            u'info_dict': {
+                u'title': u'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
+                u'uploader': u'8TV',
+                u'description': u'md5:a950cc4285c43e44d763d036710cd9cd',
             }
+        },
+    ]
 
     def report_download_webpage(self, video_id):
         """Report webpage download."""
@@ -103,6 +119,13 @@ class GenericIE(InfoExtractor):
             raise ExtractorError(u'Invalid URL: %s' % url)
 
         self.report_extraction(video_id)
+        # Look for BrigthCove:
+        m_brightcove = re.search(r'<object.+?class="BrightcoveExperience".+?</object>', webpage, re.DOTALL)
+        if m_brightcove is not None:
+            self.to_screen(u'Brightcove video detected.')
+            bc_url = BrightcoveIE._build_brighcove_url(m_brightcove.group())
+            return self.url_result(bc_url, 'Brightcove')
+
         # Start with something easy: JW Player in SWFObject
         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
         if mobj is None: