river_tao/OneForAll-mirror
1
0
Fork 0
mirror of https://github.com/shmilylty/OneForAll.git synced 2026-08-25 20:37:48 +08:00
Code Issues Packages Projects Releases Wiki Activity

统一调整子域匹配函数

Browse Source
This commit is contained in:
Jing Ling
2020-05-11 21:17:25 +08:00
parent ebd55faa9e
commit 74d5354ead
57 changed files with 94 additions and 93 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/crawl/archivecrawl.py
+2 -2
View File
@@ -26,8 +26,8 @@ class ArchiveCrawl(Crawl):
for resp in cdx.iter(url, limit=limit):
if resp.data.get('status') not in ['301', '302']:
url = resp.data.get('url')
subdomains = self.match(self.register(domain),
url + resp.text)
subdomains = self.match_subdomains(self.register(domain),
url + resp.text)
# 合并搜索子域名搜索结果
self.subdomains = self.subdomains.union(subdomains)
modules/crawl/commoncrawl.py
+1 -1
View File
@@ -26,7 +26,7 @@ class CommonCrawl(Crawl):
for resp in tqdm(cdx.iter(url, limit=limit), total=limit):
if resp.data.get('status') not in ['301', '302']:
subdomains = self.match(self.register(domain), resp.text)
subdomains = self.match_subdomains(self.register(domain), resp.text)
# 合并搜索子域名搜索结果
self.subdomains = self.subdomains.union(subdomains)