[ie/twitter] Work around API rate-limit (#8825)

Closes #8762
Authored by: bashonly

yt_dlp/extractor/instagram.py

@@ -10,6 +10,7 @@ from ..utils import (
     ExtractorError,
     decode_base_n,
     encode_base_n,
+    filter_dict,
     float_or_none,
     format_field,
     get_element_by_attribute,
@@ -703,28 +704,31 @@ class InstagramStoryIE(InstagramBaseIE):
         user_info = self._search_json(r'"user":', story_info, 'user info', story_id, fatal=False)
         if not user_info:
             self.raise_login_required('This content is unreachable')
-        user_id = user_info.get('id')
 
+        user_id = traverse_obj(user_info, 'pk', 'id', expected_type=str)
         story_info_url = user_id if username != 'highlights' else f'highlight:{story_id}'
+        if not story_info_url:  # user id is only mandatory for non-highlights
+            raise ExtractorError('Unable to extract user id')
+
         videos = traverse_obj(self._download_json(
             f'{self._API_BASE_URL}/feed/reels_media/?reel_ids={story_info_url}',
             story_id, errnote=False, fatal=False, headers=self._API_HEADERS), 'reels')
         if not videos:
             self.raise_login_required('You need to log in to access this content')
 
-        full_name = traverse_obj(videos, (f'highlight:{story_id}', 'user', 'full_name'), (str(user_id), 'user', 'full_name'))
+        full_name = traverse_obj(videos, (f'highlight:{story_id}', 'user', 'full_name'), (user_id, 'user', 'full_name'))
         story_title = traverse_obj(videos, (f'highlight:{story_id}', 'title'))
         if not story_title:
             story_title = f'Story by {username}'
 
-        highlights = traverse_obj(videos, (f'highlight:{story_id}', 'items'), (str(user_id), 'items'))
+        highlights = traverse_obj(videos, (f'highlight:{story_id}', 'items'), (user_id, 'items'))
         info_data = []
         for highlight in highlights:
             highlight_data = self._extract_product(highlight)
             if highlight_data.get('formats'):
                 info_data.append({
-                    **highlight_data,
                     'uploader': full_name,
                     'uploader_id': user_id,
+                    **filter_dict(highlight_data),
                 })
         return self.playlist_result(info_data, playlist_id=story_id, playlist_title=story_title)

yt_dlp/extractor/litv.py

@@ -6,6 +6,7 @@ from ..utils import (
     int_or_none,
     smuggle_url,
     traverse_obj,
+    try_call,
     unsmuggle_url,
 )
 
@@ -96,13 +97,22 @@ class LiTVIE(InfoExtractor):
             r'uiHlsUrl\s*=\s*testBackendData\(([^;]+)\);',
             webpage, 'video data', default='{}'), video_id)
         if not video_data:
-            payload = {
+            payload = {'assetId': program_info['assetId']}
-                'assetId': program_info['assetId'],
+            puid = try_call(lambda: self._get_cookies('https://www.litv.tv/')['PUID'].value)
-                'watchDevices': program_info['watchDevices'],
+            if puid:
-                'contentType': program_info['contentType'],
+                payload.update({
-            }
+                    'type': 'auth',
+                    'puid': puid,
+                })
+                endpoint = 'getUrl'
+            else:
+                payload.update({
+                    'watchDevices': program_info['watchDevices'],
+                    'contentType': program_info['contentType'],
+                })
+                endpoint = 'getMainUrlNoAuth'
             video_data = self._download_json(
-                'https://www.litv.tv/vod/ajax/getMainUrlNoAuth', video_id,
+                f'https://www.litv.tv/vod/ajax/{endpoint}', video_id,
                 data=json.dumps(payload).encode('utf-8'),
                 headers={'Content-Type': 'application/json'})
 

yt_dlp/extractor/twitter.py

@@ -10,6 +10,7 @@ from ..compat import (
     compat_urllib_parse_unquote,
     compat_urllib_parse_urlparse,
 )
+from ..networking.exceptions import HTTPError
 from ..utils import (
     ExtractorError,
     dict_get,
@@ -1317,41 +1318,51 @@ class TwitterIE(TwitterBaseIE):
             }
         }
 
-    def _extract_status(self, twid):
+    def _call_syndication_api(self, twid):
-        if self.is_logged_in or self._selected_api == 'graphql':
+        self.report_warning(
-            status = self._graphql_to_legacy(self._call_graphql_api(self._GRAPHQL_ENDPOINT, twid), twid)
+            'Not all metadata or media is available via syndication endpoint', twid, only_once=True)
-
+        status = self._download_json(
-        elif self._selected_api == 'legacy':
+            'https://cdn.syndication.twimg.com/tweet-result', twid, 'Downloading syndication JSON',
-            status = self._call_api(f'statuses/show/{twid}.json', twid, {
+            headers={'User-Agent': 'Googlebot'}, query={
-                'cards_platform': 'Web-12',
+                'id': twid,
-                'include_cards': 1,
+                # TODO: token = ((Number(twid) / 1e15) * Math.PI).toString(36).replace(/(0+|\.)/g, '')
-                'include_reply_count': 1,
+                'token': ''.join(random.choices('123456789abcdefghijklmnopqrstuvwxyz', k=10)),
-                'include_user_entities': 0,
-                'tweet_mode': 'extended',
             })
+        if not status:
+            raise ExtractorError('Syndication endpoint returned empty JSON response')
+        # Transform the result so its structure matches that of legacy/graphql
+        media = []
+        for detail in traverse_obj(status, ((None, 'quoted_tweet'), 'mediaDetails', ..., {dict})):
+            detail['id_str'] = traverse_obj(detail, (
+                'video_info', 'variants', ..., 'url', {self._MEDIA_ID_RE.search}, 1), get_all=False) or twid
+            media.append(detail)
+        status['extended_entities'] = {'media': media}
 
-        elif self._selected_api == 'syndication':
+        return status
-            self.report_warning(
+
-                'Not all metadata or media is available via syndication endpoint', twid, only_once=True)
+    def _extract_status(self, twid):
-            status = self._download_json(
+        if self._selected_api not in ('graphql', 'legacy', 'syndication'):
-                'https://cdn.syndication.twimg.com/tweet-result', twid, 'Downloading syndication JSON',
+            raise ExtractorError(f'{self._selected_api!r} is not a valid API selection', expected=True)
-                headers={'User-Agent': 'Googlebot'}, query={
+
-                    'id': twid,
+        try:
-                    # TODO: token = ((Number(twid) / 1e15) * Math.PI).toString(36).replace(/(0+|\.)/g, '')
+            if self.is_logged_in or self._selected_api == 'graphql':
-                    'token': ''.join(random.choices('123456789abcdefghijklmnopqrstuvwxyz', k=10)),
+                status = self._graphql_to_legacy(self._call_graphql_api(self._GRAPHQL_ENDPOINT, twid), twid)
+            elif self._selected_api == 'legacy':
+                status = self._call_api(f'statuses/show/{twid}.json', twid, {
+                    'cards_platform': 'Web-12',
+                    'include_cards': 1,
+                    'include_reply_count': 1,
+                    'include_user_entities': 0,
+                    'tweet_mode': 'extended',
                 })
-            if not status:
+        except ExtractorError as e:
-                raise ExtractorError('Syndication endpoint returned empty JSON response')
+            if not isinstance(e.cause, HTTPError) or not e.cause.status == 429:
-            # Transform the result so its structure matches that of legacy/graphql
+                raise
-            media = []
+            self.report_warning('Rate-limit exceeded; falling back to syndication endpoint')
-            for detail in traverse_obj(status, ((None, 'quoted_tweet'), 'mediaDetails', ..., {dict})):
+            status = self._call_syndication_api(twid)
-                detail['id_str'] = traverse_obj(detail, (
-                    'video_info', 'variants', ..., 'url', {self._MEDIA_ID_RE.search}, 1), get_all=False) or twid
-                media.append(detail)
-            status['extended_entities'] = {'media': media}
 
-        else:
+        if self._selected_api == 'syndication':
-            raise ExtractorError(f'"{self._selected_api}" is not a valid API selection', expected=True)
+            status = self._call_syndication_api(twid)
 
         return traverse_obj(status, 'retweeted_status', None, expected_type=dict) or {}
 
@@ -1416,8 +1427,8 @@ class TwitterIE(TwitterBaseIE):
                 'thumbnails': thumbnails,
                 'view_count': traverse_obj(media, ('mediaStats', 'viewCount', {int_or_none})),  # No longer available
                 'duration': float_or_none(traverse_obj(media, ('video_info', 'duration_millis')), 1000),
-                # The codec of http formats are unknown
+                # Prioritize m3u8 formats for compat, see https://github.com/yt-dlp/yt-dlp/issues/8117
-                '_format_sort_fields': ('res', 'br', 'size', 'proto'),
+                '_format_sort_fields': ('res', 'proto:m3u8', 'br', 'size'),  # http format codec is unknown
             }
 
         def extract_from_card_info(card):