From 7543c9c99bcb116b085fdb1f41b84a0ead04c05d Mon Sep 17 00:00:00 2001 From: "lauren n. liberda" Date: Thu, 2 Feb 2023 14:32:14 +0100 Subject: [PATCH] [extractor/twitter] Fix graphql extraction on some tweets (#6075) Authored by: selfisekai --- yt_dlp/extractor/twitter.py | 28 +++++++++++++++++++++++++++- 1 file changed, 27 insertions(+), 1 deletion(-) diff --git a/yt_dlp/extractor/twitter.py b/yt_dlp/extractor/twitter.py index a4e280c82..d3e52f392 100644 --- a/yt_dlp/extractor/twitter.py +++ b/yt_dlp/extractor/twitter.py @@ -769,6 +769,29 @@ class TwitterIE(TwitterBaseIE): 'age_limit': 0, }, 'params': {'noplaylist': True}, + }, { + # id pointing to TweetWithVisibilityResults type entity which wraps the actual Tweet over + # note the id different between extraction and url + 'url': 'https://twitter.com/s2FAKER/status/1621117700482416640', + 'info_dict': { + 'id': '1621117577354424321', + 'display_id': '1621117700482416640', + 'ext': 'mp4', + 'title': '뽀 - 아 최우제 이동속도 봐', + 'description': '아 최우제 이동속도 봐 https://t.co/dxu2U5vXXB', + 'duration': 24.598, + 'uploader': '뽀', + 'uploader_id': 's2FAKER', + 'uploader_url': 'https://twitter.com/s2FAKER', + 'upload_date': '20230202', + 'timestamp': 1675339553.0, + 'thumbnail': r're:https?://pbs\.twimg\.com/.+', + 'age_limit': 18, + 'tags': [], + 'like_count': int, + 'repost_count': int, + 'comment_count': int, + }, }, { # onion route 'url': 'https://twitter3e4tixl4xyajtrzo62zg5vztmjuricljdp2c5kshju4avyoid.onion/TwitterBlue/status/1484226494708662273', @@ -811,9 +834,12 @@ def _graphql_to_legacy(self, data, twid): result = traverse_obj(data, ( 'threaded_conversation_with_injections_v2', 'instructions', 0, 'entries', lambda _, v: v['entryId'] == f'tweet-{twid}', 'content', 'itemContent', - 'tweet_results', 'result' + 'tweet_results', 'result', ('tweet', None), ), expected_type=dict, default={}, get_all=False) + if result.get('__typename') not in ('Tweet', None): + self.report_warning(f'Unknown typename: {result.get("__typename")}', twid, only_once=True) + if 'tombstone' in result: cause = traverse_obj(result, ('tombstone', 'text', 'text'), expected_type=str) raise ExtractorError(f'Twitter API says: {cause or "Unknown error"}', expected=True)