[TikTok] Extract captions (#2185)

Closes #2184
Authored by: MinePlayersPE
This commit is contained in:
MinePlayersPE 2022-01-20 05:35:27 +07:00 committed by GitHub
parent 426764371f
commit e0585e6562
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 27 additions and 0 deletions

View File

@ -17,6 +17,7 @@ from ..utils import (
int_or_none, int_or_none,
join_nonempty, join_nonempty,
LazyList, LazyList,
srt_subtitles_timecode,
str_or_none, str_or_none,
traverse_obj, traverse_obj,
try_get, try_get,
@ -83,6 +84,27 @@ class TikTokBaseIE(InfoExtractor):
'Accept': 'application/json', 'Accept': 'application/json',
}, query=real_query) }, query=real_query)
def _get_subtitles(self, aweme_detail, aweme_id):
# TODO: Extract text positioning info
subtitles = {}
captions_info = traverse_obj(
aweme_detail, ('interaction_stickers', ..., 'auto_video_caption_info', 'auto_captions', ...), expected_type=dict, default=[])
for caption in captions_info:
caption_url = traverse_obj(caption, ('url', 'url_list', ...), expected_type=url_or_none, get_all=False)
if not caption_url:
continue
caption_json = self._download_json(
caption_url, aweme_id, note='Downloading captions', errnote='Unable to download captions', fatal=False)
if not caption_json:
continue
subtitles.setdefault(caption.get('language', 'en'), []).append({
'ext': 'srt',
'data': '\n\n'.join(
f'{i + 1}\n{srt_subtitles_timecode(line["start_time"] / 1000)} --> {srt_subtitles_timecode(line["end_time"] / 1000)}\n{line["text"]}'
for i, line in enumerate(caption_json['utterances']) if line.get('text'))
})
return subtitles
def _parse_aweme_video_app(self, aweme_detail): def _parse_aweme_video_app(self, aweme_detail):
aweme_id = aweme_detail['aweme_id'] aweme_id = aweme_detail['aweme_id']
video_info = aweme_detail['video'] video_info = aweme_detail['video']
@ -218,6 +240,7 @@ class TikTokBaseIE(InfoExtractor):
'artist': music_author, 'artist': music_author,
'timestamp': int_or_none(aweme_detail.get('create_time')), 'timestamp': int_or_none(aweme_detail.get('create_time')),
'formats': formats, 'formats': formats,
'subtitles': self.extract_subtitles(aweme_detail, aweme_id),
'thumbnails': thumbnails, 'thumbnails': thumbnails,
'duration': int_or_none(traverse_obj(video_info, 'duration', ('download_addr', 'duration')), scale=1000), 'duration': int_or_none(traverse_obj(video_info, 'duration', ('download_addr', 'duration')), scale=1000),
'availability': self._availability( 'availability': self._availability(
@ -396,6 +419,10 @@ class TikTokIE(TikTokBaseIE):
'comment_count': int, 'comment_count': int,
}, },
'expected_warnings': ['Video not available'] 'expected_warnings': ['Video not available']
}, {
# Auto-captions available
'url': 'https://www.tiktok.com/@hankgreen1/video/7047596209028074758',
'only_matching': True
}] }]
def _extract_aweme_app(self, aweme_id): def _extract_aweme_app(self, aweme_id):