mirror of
https://github.com/l1ving/youtube-dl
synced 2024-12-27 10:05:33 +08:00
69 lines
2.2 KiB
Python
69 lines
2.2 KiB
Python
from __future__ import unicode_literals
|
|
|
|
from .common import InfoExtractor
|
|
from ..compat import compat_str
|
|
from ..utils import (
|
|
clean_html,
|
|
float_or_none,
|
|
int_or_none,
|
|
try_get,
|
|
)
|
|
|
|
|
|
class HitRecordIE(InfoExtractor):
|
|
_VALID_URL = r'https?://(?:www\.)?hitrecord\.org/records/(?P<id>\d+)'
|
|
_TEST = {
|
|
'url': 'https://hitrecord.org/records/2954362',
|
|
'md5': 'fe1cdc2023bce0bbb95c39c57426aa71',
|
|
'info_dict': {
|
|
'id': '2954362',
|
|
'ext': 'mp4',
|
|
'title': 'A Very Different World (HITRECORD x ACLU)',
|
|
'description': 'md5:e62defaffab5075a5277736bead95a3d',
|
|
'duration': 139.327,
|
|
'timestamp': 1471557582,
|
|
'upload_date': '20160818',
|
|
'uploader': 'Zuzi.C12',
|
|
'uploader_id': '362811',
|
|
'view_count': int,
|
|
'like_count': int,
|
|
'comment_count': int,
|
|
'tags': list,
|
|
}
|
|
}
|
|
|
|
def _real_extract(self, url):
|
|
video_id = self._match_id(url)
|
|
|
|
video = self._download_json(
|
|
'https://hitrecord.org/api/web/records/%s' % video_id, video_id)
|
|
|
|
title = video['title']
|
|
video_url = video['source_url']['mp4_url']
|
|
|
|
tags = None
|
|
tags_list = try_get(video, lambda x: x['tags'], list)
|
|
if tags_list:
|
|
tags = [
|
|
t['text']
|
|
for t in tags_list
|
|
if isinstance(t, dict) and t.get('text') and
|
|
isinstance(t['text'], compat_str)]
|
|
|
|
return {
|
|
'id': video_id,
|
|
'url': video_url,
|
|
'title': title,
|
|
'description': clean_html(video.get('body')),
|
|
'duration': float_or_none(video.get('duration'), 1000),
|
|
'timestamp': int_or_none(video.get('created_at_i')),
|
|
'uploader': try_get(
|
|
video, lambda x: x['user']['username'], compat_str),
|
|
'uploader_id': try_get(
|
|
video, lambda x: compat_str(x['user']['id'])),
|
|
'view_count': int_or_none(video.get('total_views_count')),
|
|
'like_count': int_or_none(video.get('hearts_count')),
|
|
'comment_count': int_or_none(video.get('comments_count')),
|
|
'tags': tags,
|
|
}
|