[animeondemand] Add support for lazy playlist extraction (closes #27829)

youtube_dl/YoutubeDL.py

@@ -338,6 +338,8 @@ class YoutubeDL(object):
     _pps = []
     _download_retcode = None
     _num_downloads = None
+    _playlist_level = 0
+    _playlist_urls = set()
     _screen_file = None
 
     def __init__(self, params=None, auto_init=True):
@@ -906,8 +908,51 @@ class YoutubeDL(object):
             return self.process_ie_result(
                 new_result, download=download, extra_info=extra_info)
         elif result_type in ('playlist', 'multi_video'):
+            # Protect from infinite recursion due to recursively nested playlists
+            # (see https://github.com/ytdl-org/youtube-dl/issues/27833)
+            webpage_url = ie_result['webpage_url']
+            if webpage_url in self._playlist_urls:
+                self.to_screen(
+                    '[download] Skipping already downloaded playlist: %s'
+                    % ie_result.get('title') or ie_result.get('id'))
+                return
+
+            self._playlist_level += 1
+            self._playlist_urls.add(webpage_url)
+            try:
+                return self.__process_playlist(ie_result, download)
+            finally:
+                self._playlist_level -= 1
+                if not self._playlist_level:
+                    self._playlist_urls.clear()
+        elif result_type == 'compat_list':
+            self.report_warning(
+                'Extractor %s returned a compat_list result. '
+                'It needs to be updated.' % ie_result.get('extractor'))
+
+            def _fixup(r):
+                self.add_extra_info(
+                    r,
+                    {
+                        'extractor': ie_result['extractor'],
+                        'webpage_url': ie_result['webpage_url'],
+                        'webpage_url_basename': url_basename(ie_result['webpage_url']),
+                        'extractor_key': ie_result['extractor_key'],
+                    }
+                )
+                return r
+            ie_result['entries'] = [
+                self.process_ie_result(_fixup(r), download, extra_info)
+                for r in ie_result['entries']
+            ]
+            return ie_result
+        else:
+            raise Exception('Invalid result type: %s' % result_type)
+
+    def __process_playlist(self, ie_result, download):
         # We process each entry in the playlist
         playlist = ie_result.get('title') or ie_result.get('id')
+
         self.to_screen('[download] Downloading playlist: %s' % playlist)
 
         playlist_results = []
@@ -1015,29 +1060,6 @@ class YoutubeDL(object):
         ie_result['entries'] = playlist_results
         self.to_screen('[download] Finished downloading playlist: %s' % playlist)
         return ie_result
-        elif result_type == 'compat_list':
-            self.report_warning(
-                'Extractor %s returned a compat_list result. '
-                'It needs to be updated.' % ie_result.get('extractor'))
-
-            def _fixup(r):
-                self.add_extra_info(
-                    r,
-                    {
-                        'extractor': ie_result['extractor'],
-                        'webpage_url': ie_result['webpage_url'],
-                        'webpage_url_basename': url_basename(ie_result['webpage_url']),
-                        'extractor_key': ie_result['extractor_key'],
-                    }
-                )
-                return r
-            ie_result['entries'] = [
-                self.process_ie_result(_fixup(r), download, extra_info)
-                for r in ie_result['entries']
-            ]
-            return ie_result
-        else:
-            raise Exception('Invalid result type: %s' % result_type)
 
     @__handle_extraction_exceptions
     def __process_iterable_entry(self, entry, download, extra_info):

youtube_dl/extractor/animeondemand.py

@@ -116,8 +116,6 @@ class AnimeOnDemandIE(InfoExtractor):
             r'(?s)<div[^>]+itemprop="description"[^>]*>(.+?)</div>',
             webpage, 'anime description', default=None)
 
-        entries = []
-
         def extract_info(html, video_id, num=None):
             title, description = [None] * 2
             formats = []
@@ -233,7 +231,7 @@ class AnimeOnDemandIE(InfoExtractor):
                 self._sort_formats(info['formats'])
                 f = common_info.copy()
                 f.update(info)
-                entries.append(f)
+                yield f
 
             # Extract teaser/trailer only when full episode is not available
             if not info['formats']:
@@ -247,7 +245,7 @@ class AnimeOnDemandIE(InfoExtractor):
                         'title': m.group('title'),
                         'url': urljoin(url, m.group('href')),
                     })
-                    entries.append(f)
+                    yield f
 
         def extract_episodes(html):
             for num, episode_html in enumerate(re.findall(
@@ -275,7 +273,8 @@ class AnimeOnDemandIE(InfoExtractor):
                     'episode_number': episode_number,
                 }
 
-                extract_entries(episode_html, video_id, common_info)
+                for e in extract_entries(episode_html, video_id, common_info):
+                    yield e
 
         def extract_film(html, video_id):
             common_info = {
@@ -283,11 +282,18 @@ class AnimeOnDemandIE(InfoExtractor):
                 'title': anime_title,
                 'description': anime_description,
             }
-            extract_entries(html, video_id, common_info)
+            for e in extract_entries(html, video_id, common_info):
+                yield e
 
-        extract_episodes(webpage)
+        def entries():
+            has_episodes = False
+            for e in extract_episodes(webpage):
+                has_episodes = True
+                yield e
 
-        if not entries:
-            extract_film(webpage, anime_id)
+            if not has_episodes:
+                for e in extract_film(webpage, anime_id):
+                    yield e
 
-        return self.playlist_result(entries, anime_id, anime_title, anime_description)
+        return self.playlist_result(
+            entries(), anime_id, anime_title, anime_description)