github/gallery-dl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[deviantart:search] fix regex (#8083)

Browse Source 
use custom regex for extracting search results from HTML
instead of the 'deviation' pattern
This commit is contained in:
Mike Fährmann
2025-08-25 20:59:09 +02:00
parent 2c035e184b
commit 84dc8116a4
1 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
gallery_dl/extractor/deviantart.py
View File

@@ -1356,7 +1356,8 @@ class DeviantartSearchExtractor(DeviantartExtractor):
def _search_html(self, params): def _search_html(self, params):
url = self.root + "/search" url = self.root + "/search"
find = text.re(r'''href="https://www.deviantart.com/([^/?#]+)'''
r'''/(art|journal)/(?:[^"]+-)?(\d+)''').findall
while True: while True:
response = self.request(url, params=params) response = self.request(url, params=params)
@@ -1364,12 +1365,11 @@ class DeviantartSearchExtractor(DeviantartExtractor):
raise exception.AbortExtraction("HTTP redirect to login page") raise exception.AbortExtraction("HTTP redirect to login page")
page = response.text page = response.text
for dev in DeviantartDeviationExtractor.pattern.findall( for user, type, did in find(page)[:-3:3]:
page)[2::3]:
yield { yield {
"deviationId": dev[3], "deviationId": did,
"author": {"username": dev[0]}, "author": {"username": user},
"isJournal": dev[2] == "journal", "isJournal": type == "journal",
} }
cursor = text.extr(page, r'\"cursor\":\"', '\\',) cursor = text.extr(page, r'\"cursor\":\"', '\\',)