spider/js_reverse
1
0
Fork 0
mirror of https://github.com/luzhisheng/js_reverse.git synced 2025-04-20 21:55:07 +08:00
Code Issues Packages Projects Releases Wiki Activity

修复代码

Browse Source
This commit is contained in:
aiguigu 2021-12-16 07:53:36 +08:00
parent 6c7cf247d4
commit 43edf90f95
1 changed files with 5 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
1688/spider/1688企业产品详情内容.py → 1688/spider/1688企业产品商品内容描述.py
View File

@ -6,19 +6,19 @@ import json
import re import re
class 企业产品详情内容(Baes): class 企业产品商品内容描述(Baes):
def __init__(self): def __init__(self):
self.client = MyMongodb().db self.client = MyMongodb().db
self.col = MongoDao() self.col = MongoDao()
super(企业产品详情内容, self).__init__() super(企业产品商品内容描述, self).__init__()
def get_detail(self, url): def get_detail(self, url):
res = requests.get(url) res = requests.get(url)
return res return res
def run(self): def run(self):
res = self.client['CLEAN_CONTENT'].find({"detail_url_status": 0}).batch_size(500) res = self.client['CLEAN_CONTENT'].find({"detail_url_status": 0}).batch_size(1)
for s in res: for s in res:
sign = s.get('sign') sign = s.get('sign')
id = s.get('id') id = s.get('id')
@ -29,7 +29,7 @@ class 企业产品详情内容(Baes):
try: try:
offer_details = re.findall(r'offer_details=(.*);', res.text)[0] offer_details = re.findall(r'offer_details=(.*);', res.text)[0]
offer_details_dict = json.loads(offer_details).get('content') offer_details_dict = eval(offer_details).get('content')
except IndexError: except IndexError:
offer_details_dict = re.findall(r'desc=\'(.*)\';', res.text)[0] offer_details_dict = re.findall(r'desc=\'(.*)\';', res.text)[0]
@ -45,5 +45,5 @@ class 企业产品详情内容(Baes):
if __name__ == '__main__': if __name__ == '__main__':
img = 企业产品详情内容() img = 企业产品商品内容描述()
img.run() img.run()