[xasiat] fix 'IndexError' for albums without category (#8569)

This commit is contained in:
Mike Fährmann
2025-11-18 18:03:55 +01:00
parent 0e0f446368
commit e60007add6
2 changed files with 23 additions and 4 deletions

View File

@@ -38,7 +38,7 @@ class XasiatExtractor(Extractor):
"block_id": "list_albums_common_albums_list",
"sort_by": "post_date",
"from": pnum,
"_": int(time.time() * 1000)
"_": int(time.time() * 1000),
}
page = self.request(url, params=params).text
@@ -66,15 +66,14 @@ class XasiatAlbumExtractor(XasiatExtractor):
images = extr('class="images"', "</div>")
urls = list(text.extract_iter(images, 'href="', '"'))
categories = text.re(r'categories/[^"]+\">\s*(.+)\s*</a').findall(info)
data = {
"title": text.unescape(title),
"model": text.re(
r'top_models1"></i>\s*(.+)\s*</span').findall(info),
"tags": text.re(
r'tags/[^"]+\">\s*(.+)\s*</a').findall(info),
"album_category": text.re(
r'categories/[^"]+\">\s*(.+)\s*</a').findall(info)[0],
"album_category": categories[0] if categories else "",
"album_url": response.url,
"album_id": text.parse_int(album_id),
"count": len(urls),

View File

@@ -58,6 +58,26 @@ __tests__ = (
],
},
{
"#url" : "https://www.xasiat.com/albums/30478/gekkan-young-magazine-2025-no-11/",
"#comment" : "no 'album_category' (#8569)",
"#class" : xasiat.XasiatAlbumExtractor,
"#pattern" : r"https://www\.xasiat\.com/get_image/\d+/\w+",
"#count" : 13,
"album_category": "",
"album_id" : 30478,
"album_url" : "https://www.xasiat.com/albums/30478/gekkan-young-magazine-2025-no-11/",
"count" : 13,
"extension" : "jpg",
"model" : [],
"title" : "[Gekkan Young Magazine] 2025 No.11",
"tags" : [
"Young Magazine",
"Teen",
],
},
{
"#url" : "https://www.xasiat.com/albums/categories/gravure-idols/",
"#class" : xasiat.XasiatCategoryExtractor,