From cb7cc448c0b7508215a45af0b81506403f61ef05 Mon Sep 17 00:00:00 2001 From: Ben Welsh Date: Sun, 14 Aug 2022 13:06:04 -0700 Subject: [PATCH] [extractor/truth] Add extractor (#4609) Closes #3865 Authored by: palewire --- yt_dlp/extractor/_extractors.py | 1 + yt_dlp/extractor/truth.py | 69 +++++++++++++++++++++++++++++++++ 2 files changed, 70 insertions(+) create mode 100644 yt_dlp/extractor/truth.py diff --git a/yt_dlp/extractor/_extractors.py b/yt_dlp/extractor/_extractors.py index 34f43cc1e..eb61ad386 100644 --- a/yt_dlp/extractor/_extractors.py +++ b/yt_dlp/extractor/_extractors.py @@ -1794,6 +1794,7 @@ from .trovo import ( ) from .trueid import TrueIDIE from .trunews import TruNewsIE +from .truth import TruthIE from .trutv import TruTVIE from .tube8 import Tube8IE from .tubetugraz import TubeTuGrazIE, TubeTuGrazSeriesIE diff --git a/yt_dlp/extractor/truth.py b/yt_dlp/extractor/truth.py new file mode 100644 index 000000000..1c6409ce2 --- /dev/null +++ b/yt_dlp/extractor/truth.py @@ -0,0 +1,69 @@ +from .common import InfoExtractor +from ..utils import ( + clean_html, + format_field, + int_or_none, + strip_or_none, + traverse_obj, + unified_timestamp, +) + + +class TruthIE(InfoExtractor): + _VALID_URL = r'https?://truthsocial\.com/@[^/]+/posts/(?P\d+)' + _TESTS = [ + { + 'url': 'https://truthsocial.com/@realDonaldTrump/posts/108779000807761862', + 'md5': '4a5fb1470c192e493d9efd6f19e514d3', + 'info_dict': { + 'id': '108779000807761862', + 'ext': 'qt', + 'title': 'Truth video #108779000807761862', + 'description': None, + 'timestamp': 1659835827, + 'upload_date': '20220807', + 'uploader': 'Donald J. Trump', + 'uploader_id': 'realDonaldTrump', + 'uploader_url': 'https://truthsocial.com/@realDonaldTrump', + 'repost_count': int, + 'comment_count': int, + 'like_count': int, + }, + }, + { + 'url': 'https://truthsocial.com/@ProjectVeritasAction/posts/108618228543962049', + 'md5': 'fd47ba68933f9dce27accc52275be9c3', + 'info_dict': { + 'id': '108618228543962049', + 'ext': 'mp4', + 'title': 'md5:debde7186cf83f60ff7b44dbb9444e35', + 'description': 'md5:de2fc49045bf92bb8dc97e56503b150f', + 'timestamp': 1657382637, + 'upload_date': '20220709', + 'uploader': 'Project Veritas Action', + 'uploader_id': 'ProjectVeritasAction', + 'uploader_url': 'https://truthsocial.com/@ProjectVeritasAction', + 'repost_count': int, + 'comment_count': int, + 'like_count': int, + }, + }, + ] + + def _real_extract(self, url): + video_id = self._match_id(url) + status = self._download_json(f'https://truthsocial.com/api/v1/statuses/{video_id}', video_id) + uploader_id = strip_or_none(traverse_obj(status, ('account', 'username'))) + return { + 'id': video_id, + 'url': status['media_attachments'][0]['url'], + 'title': '', + 'description': strip_or_none(clean_html(status.get('content'))) or None, + 'timestamp': unified_timestamp(status.get('created_at')), + 'uploader': strip_or_none(traverse_obj(status, ('account', 'display_name'))), + 'uploader_id': uploader_id, + 'uploader_url': format_field(uploader_id, None, 'https://truthsocial.com/@%s'), + 'repost_count': int_or_none(status.get('reblogs_count')), + 'like_count': int_or_none(status.get('favourites_count')), + 'comment_count': int_or_none(status.get('replies_count')), + }