From 2c5e8a961e3f0fe6af870bdb420926ee208cb0bb Mon Sep 17 00:00:00 2001
From: u-spec-png <54671367+u-spec-png@users.noreply.github.com>
Date: Fri, 5 Nov 2021 22:12:16 +0000
Subject: [PATCH] [Newgrounds] Fix description (#1562)
Authored by: u-spec-png
---
yt_dlp/extractor/newgrounds.py | 19 ++++++++++++++-----
1 file changed, 14 insertions(+), 5 deletions(-)
diff --git a/yt_dlp/extractor/newgrounds.py b/yt_dlp/extractor/newgrounds.py
index bbbd9e8ee..1e1274ef0 100644
--- a/yt_dlp/extractor/newgrounds.py
+++ b/yt_dlp/extractor/newgrounds.py
@@ -6,7 +6,9 @@
from .common import InfoExtractor
from ..utils import (
+ clean_html,
extract_attributes,
+ get_element_by_id,
int_or_none,
parse_count,
parse_duration,
@@ -29,7 +31,8 @@ class NewgroundsIE(InfoExtractor):
'timestamp': 1378878540,
'upload_date': '20130911',
'duration': 143,
- 'description': 'md5:6d885138814015dfd656c2ddb00dacfc',
+ 'view_count': int,
+ 'description': 'md5:b8b3c2958875189f07d8e313462e8c4f',
},
}, {
'url': 'https://www.newgrounds.com/portal/view/1',
@@ -41,6 +44,7 @@ class NewgroundsIE(InfoExtractor):
'uploader': 'Brian-Beaton',
'timestamp': 955064100,
'upload_date': '20000406',
+ 'view_count': int,
'description': 'Scrotum plays "catch."',
'age_limit': 17,
},
@@ -54,7 +58,8 @@ class NewgroundsIE(InfoExtractor):
'uploader': 'ZONE-SAMA',
'timestamp': 1487965140,
'upload_date': '20170224',
- 'description': 'ZTV News Episode 8 (February 2017)',
+ 'view_count': int,
+ 'description': 'md5:aff9b330ec2e78ed93b1ad6d017accc6',
'age_limit': 17,
},
'params': {
@@ -70,7 +75,8 @@ class NewgroundsIE(InfoExtractor):
'uploader': 'Egoraptor',
'timestamp': 1140663240,
'upload_date': '20060223',
- 'description': 'Metal Gear is awesome is so is this movie.',
+ 'view_count': int,
+ 'description': 'md5:9246c181614e23754571995104da92e0',
'age_limit': 13,
}
}, {
@@ -80,7 +86,7 @@ class NewgroundsIE(InfoExtractor):
'id': '297383',
'ext': 'swf',
'title': 'Metal Gear Awesome',
- 'description': 'Metal Gear is awesome is so is this movie.',
+ 'description': 'Metal Gear Awesome',
'uploader': 'Egoraptor',
'upload_date': '20060223',
'timestamp': 1140663240,
@@ -145,10 +151,13 @@ def _real_extract(self, url):
(r'\s*Uploaded\s*\s*([^<]+\s*[^<]+)',
r'\s*Uploaded\s*\s*([^<]+)'), webpage, 'timestamp',
default=None))
+
duration = parse_duration(self._html_search_regex(
r'"duration"\s*:\s*["\']?(\d+)["\']?', webpage,
'duration', default=None))
+ description = clean_html(get_element_by_id('author_comments', webpage)) or self._og_search_description(webpage)
+
view_count = parse_count(self._html_search_regex(
r'(?s)\s*(?:Views|Listens)\s*\s*([\d\.,]+)', webpage,
'view count', default=None))
@@ -177,7 +186,7 @@ def _real_extract(self, url):
'duration': duration,
'formats': formats,
'thumbnail': self._og_search_thumbnail(webpage),
- 'description': self._og_search_description(webpage),
+ 'description': description,
'age_limit': age_limit,
'view_count': view_count,
}