Initiate new release.

[youtubedl] / youtube_dl / extractor / brightcove.py

diff --git a/youtube_dl/extractor/brightcove.py b/youtube_dl/extractor/brightcove.py
index 8ac38f4aa71e1fe63e51e1267f4a80919431621f..4721c22930f15cb51d0daaac294eeeca3a329092 100644 (file)
--- a/youtube_dl/extractor/brightcove.py
+++ b/youtube_dl/extractor/brightcove.py
@@ -6,31 +6,38 @@ import json
 import xml.etree.ElementTree
 
 from .common import InfoExtractor
 import xml.etree.ElementTree
 
 from .common import InfoExtractor

-from ..utils import (
-    compat_urllib_parse,
-    find_xpath_attr,
-    compat_urlparse,
+from ..compat import (
+    compat_parse_qs,

     compat_str,
     compat_str,

+    compat_urllib_parse,
+    compat_urllib_parse_urlparse,

     compat_urllib_request,
     compat_urllib_request,

-
+    compat_urlparse,
+    compat_xml_parse_error,
+)
+from ..utils import (
+    determine_ext,

     ExtractorError,
     ExtractorError,

+    find_xpath_attr,
+    fix_xml_ampersands,
+    unescapeHTML,

     unsmuggle_url,
 )
 
 
 class BrightcoveIE(InfoExtractor):
     unsmuggle_url,
 )
 
 
 class BrightcoveIE(InfoExtractor):

-    _VALID_URL = r'https?://.*brightcove\.com/(services|viewer).*\?(?P<query>.*)'
+    _VALID_URL = r'(?:https?://.*brightcove\.com/(services|viewer).*?\?|brightcove:)(?P<query>.*)'

     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'
     _FEDERATED_URL_TEMPLATE = 'http://c.brightcove.com/services/viewer/htmlFederated?%s'

-    _PLAYLIST_URL_TEMPLATE = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s'

 
     _TESTS = [
         {
             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',
 
     _TESTS = [
         {
             # From http://www.8tv.cat/8aldia/videos/xavier-sala-i-martin-aquesta-tarda-a-8-al-dia/
             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1654948606001&flashID=myExperience&%40videoPlayer=2371591881001',

-            'file': '2371591881001.mp4',

             'md5': '5423e113865d26e40624dce2e4b45d95',
             'note': 'Test Brightcove downloads and detection in GenericIE',
             'info_dict': {
             'md5': '5423e113865d26e40624dce2e4b45d95',
             'note': 'Test Brightcove downloads and detection in GenericIE',
             'info_dict': {

+                'id': '2371591881001',
+                'ext': 'mp4',

                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
                 'uploader': '8TV',
                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',
                 'title': 'Xavier Sala i Martín: “Un banc que no presta és un banc zombi que no serveix per a res”',
                 'uploader': '8TV',
                 'description': 'md5:a950cc4285c43e44d763d036710cd9cd',


@@ -39,8 +46,9 @@ class BrightcoveIE(InfoExtractor):
         {
             # From http://medianetwork.oracle.com/video/player/1785452137001
             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',
         {
             # From http://medianetwork.oracle.com/video/player/1785452137001
             'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=1217746023001&flashID=myPlayer&%40videoPlayer=1785452137001',

-            'file': '1785452137001.flv',

             'info_dict': {
             'info_dict': {

+                'id': '1785452137001',
+                'ext': 'flv',

                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
                 'uploader': 'Oracle',
                 'title': 'JVMLS 2012: Arrays 2.0 - Opportunities and Challenges',
                 'description': 'John Rose speaks at the JVM Language Summit, August 1, 2012.',
                 'uploader': 'Oracle',


@@ -69,6 +77,29 @@ class BrightcoveIE(InfoExtractor):
                 'uploader': 'National Ballet of Canada',
             },
         },
                 'uploader': 'National Ballet of Canada',
             },
         },

+        {
+            # test flv videos served by akamaihd.net
+            # From http://www.redbull.com/en/bike/stories/1331655643987/replay-uci-dh-world-cup-2014-from-fort-william
+            'url': 'http://c.brightcove.com/services/viewer/htmlFederated?%40videoPlayer=ref%3ABC2996102916001&linkBaseURL=http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fvideos%2F1331655630249%2Freplay-uci-fort-william-2014-dh&playerKey=AQ%7E%7E%2CAAAApYJ7UqE%7E%2Cxqr_zXk0I-zzNndy8NlHogrCb5QdyZRf&playerID=1398061561001#__youtubedl_smuggle=%7B%22Referer%22%3A+%22http%3A%2F%2Fwww.redbull.com%2Fen%2Fbike%2Fstories%2F1331655643987%2Freplay-uci-dh-world-cup-2014-from-fort-william%22%7D',
+            # The md5 checksum changes on each download
+            'info_dict': {
+                'id': '2996102916001',
+                'ext': 'flv',
+                'title': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
+                'uploader': 'Red Bull TV',
+                'description': 'UCI MTB World Cup 2014: Fort William, UK - Downhill Finals',
+            },
+        },
+        {
+            # playlist test
+            # from http://support.brightcove.com/en/video-cloud/docs/playlist-support-single-video-players
+            'url': 'http://c.brightcove.com/services/viewer/htmlFederated?playerID=3550052898001&playerKey=AQ%7E%7E%2CAAABmA9XpXk%7E%2C-Kp7jNgisre1fG5OdqpAFUTcs0lP_ZoL',
+            'info_dict': {
+                'title': 'Sealife',
+                'id': '3550319591001',
+            },
+            'playlist_mincount': 7,
+        },

     ]
 
     @classmethod
     ]
 
     @classmethod


@@ -79,21 +110,42 @@ class BrightcoveIE(InfoExtractor):
         """
 
         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553
         """
 
         # Fix up some stupid HTML, see https://github.com/rg3/youtube-dl/issues/1553

-        object_str = re.sub(r'(<param name="[^"]+" value="[^"]+")>',
+        object_str = re.sub(r'(<param(?:\s+[a-zA-Z0-9_]+="[^"]*")*)>',

                             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('<--', '<!--')
                             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('<--', '<!--')

+        # remove namespace to simplify extraction
+        object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
+        object_str = fix_xml_ampersands(object_str)
+
+        try:
+            object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
+        except compat_xml_parse_error:
+            return
+
+        fv_el = find_xpath_attr(object_doc, './param', 'name', 'flashVars')
+        if fv_el is not None:
+            flashvars = dict(
+                (k, v[0])
+                for k, v in compat_parse_qs(fv_el.attrib['value']).items())
+        else:
+            flashvars = {}

 
 

-        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'],
-        }

         def find_param(name):
         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
             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:
         playerKey = find_param('playerKey')
         # Not all pages define this value
         if playerKey is not None:


@@ -105,21 +157,66 @@ class BrightcoveIE(InfoExtractor):
         linkBase = find_param('linkBaseURL')
         if linkBase is not None:
             params['linkBaseURL'] = linkBase
         linkBase = find_param('linkBaseURL')
         if linkBase is not None:
             params['linkBaseURL'] = linkBase

+        return cls._make_brightcove_url(params)
+
+    @classmethod
+    def _build_brighcove_url_from_js(cls, object_js):
+        # The layout of JS is as follows:
+        # customBC.createVideo = function (width, height, playerID, playerKey, videoPlayer, VideoRandomID) {
+        #   // build Brightcove <object /> XML
+        # }
+        m = re.search(
+            r'''(?x)customBC.\createVideo\(
+                .*?                                                  # skipping width and height
+                ["\'](?P<playerID>\d+)["\']\s*,\s*                   # playerID
+                ["\'](?P<playerKey>AQ[^"\']{48})[^"\']*["\']\s*,\s*  # playerKey begins with AQ and is 50 characters
+                                                                     # in length, however it's appended to itself
+                                                                     # in places, so truncate
+                ["\'](?P<videoID>\d+)["\']                           # @videoPlayer
+            ''', object_js)
+        if m:
+            return cls._make_brightcove_url(m.groupdict())
+
+    @classmethod
+    def _make_brightcove_url(cls, params):

         data = compat_urllib_parse.urlencode(params)
         return cls._FEDERATED_URL_TEMPLATE % data
 
     @classmethod
     def _extract_brightcove_url(cls, webpage):
         data = compat_urllib_parse.urlencode(params)
         return cls._FEDERATED_URL_TEMPLATE % data
 
     @classmethod
     def _extract_brightcove_url(cls, webpage):

-        """Try to extract the brightcove url from the wepbage, returns None
+        """Try to extract the brightcove url from the webpage, returns None

         if it can't be found
         """
         if it can't be found
         """

-        m_brightcove = re.search(
-            r'<object[^>]+?class=([\'"])[^>]*?BrightcoveExperience.*?\1.+?</object>',
-            webpage, re.DOTALL)
-        if m_brightcove is not None:
-            return cls._build_brighcove_url(m_brightcove.group())
-        else:
-            return None
+        urls = cls._extract_brightcove_urls(webpage)
+        return urls[0] if urls else None
+
+    @classmethod
+    def _extract_brightcove_urls(cls, webpage):
+        """Return a list of all Brightcove URLs from the webpage """
+
+        url_m = re.search(
+            r'<meta\s+property=[\'"]og:video[\'"]\s+content=[\'"](https?://(?:secure|c)\.brightcove.com/[^\'"]+)[\'"]',
+            webpage)
+        if url_m:
+            url = unescapeHTML(url_m.group(1))
+            # Some sites don't add it, we can't download with this url, for example:
+            # http://www.ktvu.com/videos/news/raw-video-caltrain-releases-video-of-man-almost/vCTZdY/
+            if 'playerKey' in url or 'videoId' in url:
+                return [url]
+
+        matches = re.findall(
+            r'''(?sx)<object
+            (?:
+                [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
+                [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
+            ).+?>\s*</object>''',
+            webpage)
+        if matches:
+            return list(filter(None, [cls._build_brighcove_url(m) for m in matches]))
+
+        return list(filter(None, [
+            cls._build_brighcove_url_from_js(custom_bc)
+            for custom_bc in re.findall(r'(customBC\.createVideo\(.+?\);)', webpage)]))

 
     def _real_extract(self, url):
         url, smuggled_data = unsmuggle_url(url, {})
 
     def _real_extract(self, url):
         url, smuggled_data = unsmuggle_url(url, {})


@@ -138,9 +235,13 @@ class BrightcoveIE(InfoExtractor):
             referer = smuggled_data.get('Referer', url)
             return self._get_video_info(
                 videoPlayer[0], query_str, query, referer=referer)
             referer = smuggled_data.get('Referer', url)
             return self._get_video_info(
                 videoPlayer[0], query_str, query, referer=referer)

-        else:
+        elif 'playerKey' in query:

             player_key = query['playerKey']
             return self._get_playlist_info(player_key[0])
             player_key = query['playerKey']
             return self._get_playlist_info(player_key[0])

+        else:
+            raise ExtractorError(
+                'Cannot find playerKey= variable. Did you forget quotes in a shell invocation?',
+                expected=True)

 
     def _get_video_info(self, video_id, query_str, query, referer=None):
         request_url = self._FEDERATED_URL_TEMPLATE % query_str
 
     def _get_video_info(self, video_id, query_str, query, referer=None):
         request_url = self._FEDERATED_URL_TEMPLATE % query_str


@@ -152,16 +253,25 @@ class BrightcoveIE(InfoExtractor):
             req.add_header('Referer', referer)
         webpage = self._download_webpage(req, video_id)
 
             req.add_header('Referer', referer)
         webpage = self._download_webpage(req, video_id)
 

+        error_msg = self._html_search_regex(
+            r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
+            'error message', default=None)
+        if error_msg is not None:
+            raise ExtractorError(
+                'brightcove said: %s' % error_msg, expected=True)
+

         self.report_extraction(video_id)
         self.report_extraction(video_id)

-        info = self._search_regex(r'var experienceJSON = ({.*?});', webpage, 'json')
+        info = self._search_regex(r'var experienceJSON = ({.*});', webpage, 'json')

         info = json.loads(info)['data']
         video_info = info['programmedContent']['videoPlayer']['mediaDTO']
         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)
 
     def _get_playlist_info(self, player_key):
 
         return self._extract_video_info(video_info)
 
     def _get_playlist_info(self, player_key):

-        playlist_info = self._download_webpage(self._PLAYLIST_URL_TEMPLATE % player_key,
-                                               player_key, 'Downloading playlist information')
+        info_url = 'http://c.brightcove.com/services/json/experience/runtime/?command=get_programming_for_experience&playerKey=%s' % player_key
+        playlist_info = self._download_webpage(
+            info_url, player_key, 'Downloading playlist information')

 
         json_data = json.loads(playlist_info)
         if 'videoList' not in json_data:
 
         json_data = json.loads(playlist_info)
         if 'videoList' not in json_data:


@@ -169,13 +279,13 @@ class BrightcoveIE(InfoExtractor):
         playlist_info = json_data['videoList']
         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 
         playlist_info = json_data['videoList']
         videos = [self._extract_video_info(video_info) for video_info in playlist_info['mediaCollectionDTO']['videoDTOs']]
 

-        return self.playlist_result(videos, playlist_id=playlist_info['id'],
+        return self.playlist_result(videos, playlist_id='%s' % playlist_info['id'],

                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 
     def _extract_video_info(self, video_info):
         info = {
             'id': compat_str(video_info['id']),
                                     playlist_title=playlist_info['mediaCollectionDTO']['displayName'])
 
     def _extract_video_info(self, video_info):
         info = {
             'id': compat_str(video_info['id']),

-            'title': video_info['displayName'],
+            'title': video_info['displayName'].strip(),

             'description': video_info.get('shortDescription'),
             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
             'uploader': video_info.get('publisherName'),
             'description': video_info.get('shortDescription'),
             'thumbnail': video_info.get('videoStillURL') or video_info.get('thumbnailURL'),
             'uploader': video_info.get('publisherName'),


@@ -183,16 +293,56 @@ class BrightcoveIE(InfoExtractor):
 
         renditions = video_info.get('renditions')
         if renditions:
 
         renditions = video_info.get('renditions')
         if renditions:

-            renditions = sorted(renditions, key=lambda r: r['size'])
-            info['formats'] = [{
-                'url': rend['defaultURL'],
-                'height': rend.get('frameHeight'),
-                'width': rend.get('frameWidth'),
-            } for rend in renditions]
+            formats = []
+            for rend in renditions:
+                url = rend['defaultURL']
+                if not url:
+                    continue
+                ext = None
+                if rend['remote']:
+                    url_comp = compat_urllib_parse_urlparse(url)
+                    if url_comp.path.endswith('.m3u8'):
+                        formats.extend(
+                            self._extract_m3u8_formats(url, info['id'], 'mp4'))
+                        continue
+                    elif 'akamaihd.net' in url_comp.netloc:
+                        # This type of renditions are served through
+                        # akamaihd.net, but they don't use f4m manifests
+                        url = url.replace('control/', '') + '?&v=3.3.0&fp=13&r=FEEFJ&g=RTSJIMBMPFPB'
+                        ext = 'flv'
+                if ext is None:
+                    ext = determine_ext(url)
+                size = rend.get('size')
+                formats.append({
+                    'url': url,
+                    'ext': ext,
+                    'height': rend.get('frameHeight'),
+                    'width': rend.get('frameWidth'),
+                    'filesize': size if size != 0 else None,
+                })
+            self._sort_formats(formats)
+            info['formats'] = formats

         elif video_info.get('FLVFullLengthURL') is not None:
             info.update({
                 'url': video_info['FLVFullLengthURL'],
             })
         elif video_info.get('FLVFullLengthURL') is not None:
             info.update({
                 'url': video_info['FLVFullLengthURL'],
             })

-        else:
+
+        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 and not info.get('formats'):

             raise ExtractorError('Unable to extract video url for %s' % info['id'])
         return info
             raise ExtractorError('Unable to extract video url for %s' % info['id'])
         return info