[voicerepublic] Make more robust and extract more metadata

youtube_dl/extractor/voicerepublic.py

@@ -1,52 +1,99 @@
-# coding: utf-8
 from __future__ import unicode_literals
 
 import re
 
 from .common import InfoExtractor
-from ..compat import compat_urllib_request
-from ..utils import ExtractorError
+from ..compat import (
+    compat_urllib_request,
+    compat_urlparse,
+)
+from ..utils import (
+    ExtractorError,
+    determine_ext,
+    int_or_none,
+)
 
 
 class VoiceRepublicIE(InfoExtractor):
-    _VALID_URL = r'https?://voicerepublic\.com/talks/(?P<id>[0-9a-z-]+)'
-    _TEST = {
-        'url': 'https://voicerepublic.com/talks/watching-the-watchers-building-a-sousveillance-state',
+    _VALID_URL = r'https?://voicerepublic\.com/(?:talks|embed)/(?P<id>[0-9a-z-]+)'
+    _TESTS = [{
+        'url': 'http://voicerepublic.com/talks/watching-the-watchers-building-a-sousveillance-state',
         'md5': '0554a24d1657915aa8e8f84e15dc9353',
         'info_dict': {
             'id': '2296',
+            'display_id': 'watching-the-watchers-building-a-sousveillance-state',
             'ext': 'm4a',
             'title': 'Watching the Watchers: Building a Sousveillance State',
-            'thumbnail': 'https://voicerepublic.com/system/flyer/2296.png',
             'description': 'md5:715ba964958afa2398df615809cfecb1',
+            'thumbnail': 're:^https?://.*\.(?:png|jpg)$',
+            'duration': 1800,
+            'view_count': int,
         }
-    }
+    }, {
+        'url': 'http://voicerepublic.com/embed/watching-the-watchers-building-a-sousveillance-state',
+        'only_matching': True,
+    }]
 
     def _real_extract(self, url):
         display_id = self._match_id(url)
-        req = compat_urllib_request.Request(url)
+
+        req = compat_urllib_request.Request(
+            compat_urlparse.urljoin(url, '/talks/%s' % display_id))
         # Older versions of Firefox get redirected to an "upgrade browser" page
         req.add_header('User-Agent', 'youtube-dl')
         webpage = self._download_webpage(req, display_id)
-        thumbnail = self._og_search_thumbnail(webpage)
-        video_id = self._search_regex(r'/(\d+)\.png', thumbnail, 'id')
 
-        if '<a>Queued for processing, please stand by...</a>' in webpage:
-            raise ExtractorError('Audio is still queued for processing')
+        if '>Queued for processing, please stand by...<' in webpage:
+            raise ExtractorError(
+                'Audio is still queued for processing', expected=True)
 
+        data = self._parse_json(
+            self._search_regex(
+                r'(?s)return ({.+?});\s*\n', webpage,
+                'data', default=None),
+            display_id, fatal=False)
+
+        if data:
+            title = data['title']
+            description = data.get('teaser')
+            talk_id = data.get('talk_id') or display_id
+            talk = data['talk']
+            duration = int_or_none(talk.get('duration'))
             formats = [{
-            'url': 'https://voicerepublic.com' + path,
-            'ext': ext,
-            'format_id': ext,
+                'url': compat_urlparse.urljoin(url, talk_url),
+                'format_id': format_id,
+                'ext': determine_ext(talk_url) or format_id,
                 'vcodec': 'none',
-        } for ext, path in re.findall(r"data-([^=]+)='(/[^']+\.\1)'", webpage)]
+            } for format_id, talk_url in talk['links'].items()]
+        else:
+            title = self._og_search_title(webpage)
+            description = self._html_search_regex(
+                r"(?s)<div class='talk-teaser'[^>]*>(.+?)</div>",
+                webpage, 'description', fatal=False)
+            talk_id = self._search_regex(
+                [r"id='jc-(\d+)'", r"data-shareable-id='(\d+)'"],
+                webpage, 'talk id', default=None) or display_id
+            duration = None
+            formats = [{
+                'url': compat_urlparse.urljoin(url, talk_url),
+                'format_id': format_id,
+                'ext': determine_ext(talk_url) or format_id,
+                'vcodec': 'none',
+            } for format_id, talk_url in re.findall(r"data-([^=]+)='([^']+)'", webpage)]
         self._sort_formats(formats)
 
+        thumbnail = self._og_search_thumbnail(webpage)
+        view_count = int_or_none(self._search_regex(
+            r"class='play-count[^']*'>\s*(\d+) plays",
+            webpage, 'play count', fatal=False))
+
         return {
-            'id': video_id,
-            'title': self._og_search_title(webpage),
-            'formats': formats,
-            'url': self._og_search_url(webpage),
+            'id': talk_id,
+            'display_id': display_id,
+            'title': title,
+            'description': description,
             'thumbnail': thumbnail,
-            'description': self._og_search_description(webpage),
+            'duration': duration,
+            'view_count': view_count,
+            'formats': formats,
         }