[ustream] Add an alternative approach to extract title (fixes #5128)

[youtube-dl.git] / youtube_dl / extractor / ustream.py

diff --git a/youtube_dl/extractor/ustream.py b/youtube_dl/extractor/ustream.py
index 488b10df96e298c683cd02287e2da0c49f21a1cc..a395dc5dac85d0dabe5aea59e25275942f7e28a8 100644 (file)
--- a/youtube_dl/extractor/ustream.py
+++ b/youtube_dl/extractor/ustream.py
@@ -4,10 +4,10 @@ import json
 import re
 
 from .common import InfoExtractor
-from ..utils import (
+from ..compat import (
     compat_urlparse,
-    get_meta_content,
 )
+from ..utils import ExtractorError
 
 
 class UstreamIE(InfoExtractor):
@@ -41,19 +41,39 @@ class UstreamIE(InfoExtractor):
             desktop_url = 'http://www.ustream.tv/recorded/' + desktop_video_id
             return self.url_result(desktop_url, 'Ustream')
 
-        video_url = 'http://tcdn.ustream.tv/video/%s' % video_id
+        params = self._download_json(
+            'http://cdngw.ustream.tv/rgwjson/Viewer.getVideo/' + json.dumps({
+                'brandId': 1,
+                'videoId': int(video_id),
+                'autoplay': False,
+            }), video_id)
+
+        if 'error' in params:
+            raise ExtractorError(params['error']['message'], expected=True)
+
+        video_url = params['flv']
+
         webpage = self._download_webpage(url, video_id)
 
         self.report_extraction(video_id)
 
         video_title = self._html_search_regex(r'data-title="(?P<title>.+)"',
-            webpage, 'title')
+                                              webpage, 'title', default=None)
+
+        if not video_title:
+            try:
+                video_title = params['moduleConfig']['meta']['title']
+            except KeyError:
+                pass
+
+        if not video_title:
+            video_title = 'Ustream video ' + video_id
 
         uploader = self._html_search_regex(r'data-content-type="channel".*?>(?P<uploader>.*?)</a>',
-            webpage, 'uploader', fatal=False, flags=re.DOTALL)
+                                           webpage, 'uploader', fatal=False, flags=re.DOTALL)
 
         thumbnail = self._html_search_regex(r'<link rel="image_src" href="(?P<thumb>.*?)"',
-            webpage, 'thumbnail', fatal=False)
+                                            webpage, 'thumbnail', fatal=False)
 
         return {
             'id': video_id,
@@ -68,21 +88,36 @@ class UstreamIE(InfoExtractor):
 class UstreamChannelIE(InfoExtractor):
     _VALID_URL = r'https?://www\.ustream\.tv/channel/(?P<slug>.+)'
     IE_NAME = 'ustream:channel'
+    _TEST = {
+        'url': 'http://www.ustream.tv/channel/channeljapan',
+        'info_dict': {
+            'id': '10874166',
+        },
+        'playlist_mincount': 17,
+    }
 
     def _real_extract(self, url):
         m = re.match(self._VALID_URL, url)
-        slug = m.group('slug')
-        webpage = self._download_webpage(url, slug)
-        channel_id = get_meta_content('ustream:channel_id', webpage)
+        display_id = m.group('slug')
+        webpage = self._download_webpage(url, display_id)
+        channel_id = self._html_search_meta('ustream:channel_id', webpage)
 
         BASE = 'http://www.ustream.tv'
         next_url = '/ajax/socialstream/videos/%s/1.json' % channel_id
         video_ids = []
         while next_url:
-            reply = json.loads(self._download_webpage(compat_urlparse.urljoin(BASE, next_url), channel_id))
+            reply = self._download_json(
+                compat_urlparse.urljoin(BASE, next_url), display_id,
+                note='Downloading video information (next: %d)' % (len(video_ids) + 1))
             video_ids.extend(re.findall(r'data-content-id="(\d.*)"', reply['data']))
             next_url = reply['nextUrl']
 
-        urls = ['http://www.ustream.tv/recorded/' + vid for vid in video_ids]
-        url_entries = [self.url_result(eurl, 'Ustream') for eurl in urls]
-        return self.playlist_result(url_entries, channel_id)
+        entries = [
+            self.url_result('http://www.ustream.tv/recorded/' + vid, 'Ustream')
+            for vid in video_ids]
+        return {
+            '_type': 'playlist',
+            'id': channel_id,
+            'display_id': display_id,
+            'entries': entries,
+        }