Merge branch 'lecture2go' of https://github.com/nichdu/youtube-dl into nichdu-lecture2go

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

diff --git a/youtube_dl/extractor/brightcove.py b/youtube_dl/extractor/brightcove.py
index 8d5acef14bd9556b25abceb7df4d4d32131f8624..4721c22930f15cb51d0daaac294eeeca3a329092 100644 (file)
--- a/youtube_dl/extractor/brightcove.py
+++ b/youtube_dl/extractor/brightcove.py
@@ -6,25 +6,27 @@ import json
 import xml.etree.ElementTree
 
 from .common import InfoExtractor
-from ..utils import (
-    compat_urllib_parse,
-    find_xpath_attr,
-    fix_xml_ampersands,
-    compat_urlparse,
-    compat_str,
-    compat_urllib_request,
+from ..compat import (
     compat_parse_qs,
+    compat_str,
+    compat_urllib_parse,
     compat_urllib_parse_urlparse,
-
+    compat_urllib_request,
+    compat_urlparse,
+    compat_xml_parse_error,
+)
+from ..utils import (
     determine_ext,
     ExtractorError,
-    unsmuggle_url,
+    find_xpath_attr,
+    fix_xml_ampersands,
     unescapeHTML,
+    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'
 
     _TESTS = [
@@ -94,6 +96,7 @@ class BrightcoveIE(InfoExtractor):
             '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,
         },
@@ -107,7 +110,7 @@ class BrightcoveIE(InfoExtractor):
         """
 
         # 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('<--', '<!--')
@@ -115,7 +118,10 @@ class BrightcoveIE(InfoExtractor):
         object_str = re.sub(r'(<object[^>]*)(xmlns=".*?")', r'\1', object_str)
         object_str = fix_xml_ampersands(object_str)
 
-        object_doc = xml.etree.ElementTree.fromstring(object_str.encode('utf-8'))
+        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:
@@ -151,6 +157,28 @@ class BrightcoveIE(InfoExtractor):
         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
 
@@ -167,7 +195,7 @@ class BrightcoveIE(InfoExtractor):
         """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/[^"]+)"',
+            r'<meta\s+property=[\'"]og:video[\'"]\s+content=[\'"](https?://(?:secure|c)\.brightcove.com/[^\'"]+)[\'"]',
             webpage)
         if url_m:
             url = unescapeHTML(url_m.group(1))
@@ -181,9 +209,14 @@ class BrightcoveIE(InfoExtractor):
             (?:
                 [^>]+?class=[\'"][^>]*?BrightcoveExperience.*?[\'"] |
                 [^>]*?>\s*<param\s+name="movie"\s+value="https?://[^/]*brightcove\.com/
-            ).+?</object>''',
+            ).+?>\s*</object>''',
             webpage)
-        return [cls._build_brighcove_url(m) for m in matches]
+        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, {})
@@ -221,7 +254,7 @@ class BrightcoveIE(InfoExtractor):
         webpage = self._download_webpage(req, video_id)
 
         error_msg = self._html_search_regex(
-            r"<h1>We're sorry.</h1>\s*<p>(.*?)</p>", webpage,
+            r"<h1>We're sorry.</h1>([\s\n]*<p>.*?</p>)+", webpage,
             'error message', default=None)
         if error_msg is not None:
             raise ExtractorError(
@@ -246,7 +279,7 @@ class BrightcoveIE(InfoExtractor):
         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):
@@ -265,6 +298,7 @@ class BrightcoveIE(InfoExtractor):
                 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'):
@@ -276,7 +310,7 @@ class BrightcoveIE(InfoExtractor):
                         # 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'
-                else:
+                if ext is None:
                     ext = determine_ext(url)
                 size = rend.get('size')
                 formats.append({