Merge branch 'master' of github.com:rg3/youtube-dl

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

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 75044d71a3fd9f81fa5d89ab8283eb13e5d8191d..78f3b7e7bddb1b085c96ae41696837178c2f42aa 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -316,6 +316,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
                 u"upload_date": u"20121002",
                 u"description": u"test chars:  \"'/\\ä↭𝕐\ntest URL: https://github.com/rg3/youtube-dl/issues/1892\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de .",
                 u"categories": [u'Science & Technology'],
+                'like_count': int,
+                'dislike_count': int,
             }
         },
         {
@@ -784,7 +786,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
             upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split())
             upload_date = unified_strdate(upload_date)
 
-        m_cat_container = get_element_by_id("eow-category", video_webpage)
+        m_cat_container = self._search_regex(
+            r'(?s)<h4[^>]*>\s*Category\s*</h4>\s*<ul[^>]*>(.*?)</ul>',
+            video_webpage, 'categories', fatal=False)
         if m_cat_container:
             category = self._html_search_regex(
                 r'(?s)<a[^<]+>(.*?)</a>', m_cat_container, 'category',
@@ -813,15 +817,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
             else:
                 video_description = u''
 
-        def _extract_count(klass):
+        def _extract_count(count_name):
             count = self._search_regex(
-                r'class="%s">([\d,]+)</span>' % re.escape(klass),
-                video_webpage, klass, default=None)
+                r'id="watch-%s"[^>]*>.*?([\d,]+)\s*</span>' % re.escape(count_name),
+                video_webpage, count_name, default=None)
             if count is not None:
                 return int(count.replace(',', ''))
             return None
-        like_count = _extract_count(u'likes-count')
-        dislike_count = _extract_count(u'dislikes-count')
+        like_count = _extract_count(u'like')
+        dislike_count = _extract_count(u'dislike')
 
         # subtitles
         video_subtitles = self.extract_subtitles(video_id, video_webpage)
@@ -1430,12 +1434,6 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
             paging = mobj.group('paging')
         return self.playlist_result(feed_entries, playlist_title=self._PLAYLIST_TITLE)
 
-class YoutubeSubscriptionsIE(YoutubeFeedsInfoExtractor):
-    IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
-    _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
-    _FEED_NAME = 'subscriptions'
-    _PLAYLIST_TITLE = u'Youtube Subscriptions'
-
 class YoutubeRecommendedIE(YoutubeFeedsInfoExtractor):
     IE_DESC = u'YouTube.com recommended videos, "ytrec" keyword (requires authentication)'
     _VALID_URL = r'https?://www\.youtube\.com/feed/recommended|:ytrec(?:ommended)?'
@@ -1468,6 +1466,43 @@ class YoutubeFavouritesIE(YoutubeBaseInfoExtractor):
         return self.url_result(playlist_id, 'YoutubePlaylist')
 
 
+class YoutubeSubscriptionsIE(YoutubePlaylistIE):
+    IE_NAME = u'youtube:subscriptions'
+    IE_DESC = u'YouTube.com subscriptions feed, "ytsubs" keyword (requires authentication)'
+    _VALID_URL = r'https?://www\.youtube\.com/feed/subscriptions|:ytsubs(?:criptions)?'
+
+    def _real_extract(self, url):
+        title = u'Youtube Subscriptions'
+        page = self._download_webpage('https://www.youtube.com/feed/subscriptions', title)
+
+        # The extraction process is the same as for playlists, but the regex
+        # for the video ids doesn't contain an index
+        ids = []
+        more_widget_html = content_html = page
+
+        for page_num in itertools.count(1):
+            matches = re.findall(r'href="\s*/watch\?v=([0-9A-Za-z_-]{11})', content_html)
+            new_ids = orderedSet(matches)
+            ids.extend(new_ids)
+
+            mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
+            if not mobj:
+                break
+
+            more = self._download_json(
+                'https://youtube.com/%s' % mobj.group('more'), title,
+                'Downloading page #%s' % page_num,
+                transform_source=uppercase_escape)
+            content_html = more['content_html']
+            more_widget_html = more['load_more_widget_html']
+
+        return {
+            '_type': 'playlist',
+            'title': title,
+            'entries': self._ids_to_results(ids),
+        }
+
+
 class YoutubeTruncatedURLIE(InfoExtractor):
     IE_NAME = 'youtube:truncated_url'
     IE_DESC = False  # Do not list