[thehentaiworld] support '/gif-animated-hentai-images/' URLs (#8277)

generalize 'post' URL pattern
This commit is contained in:
Mike Fährmann
2025-09-23 19:06:23 +02:00
parent d03572c397
commit 889ecd7963
2 changed files with 15 additions and 10 deletions

View File

@@ -109,16 +109,6 @@ class ThehentaiworldExtractor(Extractor):
pnum += 1
class ThehentaiworldPostExtractor(ThehentaiworldExtractor):
subcategory = "post"
pattern = (rf"{BASE_PATTERN}"
rf"(/(?:(?:3d-cgi-)?hentai-image|video)s/([^/?#]+))")
example = "https://thehentaiworld.com/hentai-images/SLUG/"
def posts(self):
return (f"{self.root}{self.groups[0]}/",)
class ThehentaiworldTagExtractor(ThehentaiworldExtractor):
subcategory = "tag"
per_page = 24
@@ -137,3 +127,13 @@ class ThehentaiworldTagExtractor(ThehentaiworldExtractor):
self.page_start += pages
self.post_start += posts
return num
class ThehentaiworldPostExtractor(ThehentaiworldExtractor):
subcategory = "post"
pattern = (rf"{BASE_PATTERN}("
rf"/(?:video|(?:[\w-]+-)?hentai-image)s/([^/?#]+))")
example = "https://thehentaiworld.com/hentai-images/SLUG/"
def posts(self):
return (f"{self.root}{self.groups[0]}/",)

View File

@@ -179,6 +179,11 @@ __tests__ = (
"#results" : "https://thehentaiworld.com/wp-content/uploads/2025/08/Ciri-and-Shani-CEKC-The-Witcher-3-Hentai-3D-CGI.jpeg",
},
{
"#url" : "https://thehentaiworld.com/gif-animated-hentai-images/rangiku-matsumoto-sketchdrif-bleach/",
"#class" : thehentaiworld.ThehentaiworldPostExtractor,
},
{
"#url" : "https://thehentaiworld.com/tag/aurahack/",
"#class" : thehentaiworld.ThehentaiworldTagExtractor,