Update upstream source from tag 'upstream/2018.04.25'

[youtubedl] / youtube_dl / extractor / youtube.py

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 43051512bc1013640bc99d98437b3cf2a7b258a3..e7bd1f18fb504c8541d0fbdc301980cab07d1f54 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -87,7 +87,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
         (username, password) = self._get_login_info()
         # No authentication to be performed
         if username is None:
         (username, password) = self._get_login_info()
         # No authentication to be performed
         if username is None:

-            if self._LOGIN_REQUIRED:
+            if self._LOGIN_REQUIRED and self._downloader.params.get('cookiefile') is None:

                 raise ExtractorError('No login info available, needed for using %s.' % self.IE_NAME, expected=True)
             return True
 
                 raise ExtractorError('No login info available, needed for using %s.' % self.IE_NAME, expected=True)
             return True
 


@@ -1944,6 +1944,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                                     break
                             if codecs:
                                 dct.update(parse_codecs(codecs))
                                     break
                             if codecs:
                                 dct.update(parse_codecs(codecs))

+                if dct.get('acodec') == 'none' or dct.get('vcodec') == 'none':
+                    dct['downloader_options'] = {
+                        # Youtube throttles chunks >~10M
+                        'http_chunk_size': 10485760,
+                    }

                 formats.append(dct)
         elif video_info.get('hlsvp'):
             manifest_url = video_info['hlsvp'][0]
                 formats.append(dct)
         elif video_info.get('hlsvp'):
             manifest_url = video_info['hlsvp'][0]


@@ -2446,7 +2451,7 @@ class YoutubeChannelIE(YoutubePlaylistBaseInfoExtractor):
 
 class YoutubeUserIE(YoutubeChannelIE):
     IE_DESC = 'YouTube.com user videos (URL or "ytuser" keyword)'
 
 class YoutubeUserIE(YoutubeChannelIE):
     IE_DESC = 'YouTube.com user videos (URL or "ytuser" keyword)'

-    _VALID_URL = r'(?:(?:https?://(?:\w+\.)?youtube\.com/(?:(?P<user>user|c)/)?(?!(?:attribution_link|watch|results)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)(?P<id>[A-Za-z0-9_-]+)'
+    _VALID_URL = r'(?:(?:https?://(?:\w+\.)?youtube\.com/(?:(?P<user>user|c)/)?(?!(?:attribution_link|watch|results|shared)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)(?P<id>[A-Za-z0-9_-]+)'

     _TEMPLATE_URL = 'https://www.youtube.com/%s/%s/videos'
     IE_NAME = 'youtube:user'
 
     _TEMPLATE_URL = 'https://www.youtube.com/%s/%s/videos'
     IE_NAME = 'youtube:user'
 


@@ -2578,7 +2583,11 @@ class YoutubePlaylistsIE(YoutubePlaylistsBaseInfoExtractor):
     }]
 
 
     }]
 
 

-class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE):
+class YoutubeSearchBaseInfoExtractor(YoutubePlaylistBaseInfoExtractor):
+    _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})(?:[^"]*"[^>]+\btitle="(?P<title>[^"]+))?'
+
+
+class YoutubeSearchIE(SearchInfoExtractor, YoutubeSearchBaseInfoExtractor):

     IE_DESC = 'YouTube.com searches'
     # there doesn't appear to be a real limit, for example if you search for
     # 'python' you get more than 8.000.000 results
     IE_DESC = 'YouTube.com searches'
     # there doesn't appear to be a real limit, for example if you search for
     # 'python' you get more than 8.000.000 results


@@ -2612,8 +2621,7 @@ class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE):
                 raise ExtractorError(
                     '[youtube] No video results', expected=True)
 
                 raise ExtractorError(
                     '[youtube] No video results', expected=True)
 

-            new_videos = self._ids_to_results(orderedSet(re.findall(
-                r'href="/watch\?v=(.{11})', html_content)))
+            new_videos = list(self._process_page(html_content))

             videos += new_videos
             if not new_videos or len(videos) > limit:
                 break
             videos += new_videos
             if not new_videos or len(videos) > limit:
                 break


@@ -2636,11 +2644,10 @@ class YoutubeSearchDateIE(YoutubeSearchIE):
     _EXTRA_QUERY_ARGS = {'search_sort': 'video_date_uploaded'}
 
 
     _EXTRA_QUERY_ARGS = {'search_sort': 'video_date_uploaded'}
 
 

-class YoutubeSearchURLIE(YoutubePlaylistBaseInfoExtractor):
+class YoutubeSearchURLIE(YoutubeSearchBaseInfoExtractor):

     IE_DESC = 'YouTube.com search URLs'
     IE_NAME = 'youtube:search_url'
     _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?(?:search_query|q)=(?P<query>[^&]+)(?:[&]|$)'
     IE_DESC = 'YouTube.com search URLs'
     IE_NAME = 'youtube:search_url'
     _VALID_URL = r'https?://(?:www\.)?youtube\.com/results\?(.*?&)?(?:search_query|q)=(?P<query>[^&]+)(?:[&]|$)'

-    _VIDEO_RE = r'href="\s*/watch\?v=(?P<id>[0-9A-Za-z_-]{11})(?:[^"]*"[^>]+\btitle="(?P<title>[^"]+))?'

     _TESTS = [{
         'url': 'https://www.youtube.com/results?baz=bar&search_query=youtube-dl+test+video&filters=video&lclk=video',
         'playlist_mincount': 5,
     _TESTS = [{
         'url': 'https://www.youtube.com/results?baz=bar&search_query=youtube-dl+test+video&filters=video&lclk=video',
         'playlist_mincount': 5,


@@ -2692,10 +2699,7 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
     def _real_initialize(self):
         self._login()
 
     def _real_initialize(self):
         self._login()
 

-    def _real_extract(self, url):
-        page = self._download_webpage(
-            'https://www.youtube.com/feed/%s' % self._FEED_NAME, self._PLAYLIST_TITLE)
-
+    def _entries(self, page):

         # The extraction process is the same as for playlists, but the regex
         # for the video ids doesn't contain an index
         ids = []
         # The extraction process is the same as for playlists, but the regex
         # for the video ids doesn't contain an index
         ids = []


@@ -2706,12 +2710,15 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
             # 'recommended' feed has infinite 'load more' and each new portion spins
             # the same videos in (sometimes) slightly different order, so we'll check
             # for unicity and break when portion has no new videos
             # 'recommended' feed has infinite 'load more' and each new portion spins
             # the same videos in (sometimes) slightly different order, so we'll check
             # for unicity and break when portion has no new videos

-            new_ids = filter(lambda video_id: video_id not in ids, orderedSet(matches))
+            new_ids = list(filter(lambda video_id: video_id not in ids, orderedSet(matches)))

             if not new_ids:
                 break
 
             ids.extend(new_ids)
 
             if not new_ids:
                 break
 
             ids.extend(new_ids)
 

+            for entry in self._ids_to_results(new_ids):
+                yield entry
+

             mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
             if not mobj:
                 break
             mobj = re.search(r'data-uix-load-more-href="/?(?P<more>[^"]+)"', more_widget_html)
             if not mobj:
                 break


@@ -2723,8 +2730,12 @@ class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):
             content_html = more['content_html']
             more_widget_html = more['load_more_widget_html']
 
             content_html = more['content_html']
             more_widget_html = more['load_more_widget_html']
 

+    def _real_extract(self, url):
+        page = self._download_webpage(
+            'https://www.youtube.com/feed/%s' % self._FEED_NAME,
+            self._PLAYLIST_TITLE)

         return self.playlist_result(
         return self.playlist_result(

-            self._ids_to_results(ids), playlist_title=self._PLAYLIST_TITLE)
+            self._entries(page), playlist_title=self._PLAYLIST_TITLE)

 
 
 class YoutubeWatchLaterIE(YoutubePlaylistIE):
 
 
 class YoutubeWatchLaterIE(YoutubePlaylistIE):