江西省人民政府

REITs专题数据/policy-jiangxi.py

-import time
+import time
@@ -66,6 +66,7 @@ def getDataJson():
 def getContent(url, num, publishDate):
     id_list = []
     soup = getSoup(url)
+    policy.paserUrl(soup, url)
     contentWithTag = soup.find('div', attrs={'id': 'zoom'})
     img_list = contentWithTag.find_all('img')
     num_ = 1
@@ -108,6 +109,7 @@ def doJob():
     for data_ in data_json:
         data_ = data_.replace('\\', '')
         soup = BeautifulSoup(data_, 'lxml')
+
         title = soup.select('body > div > div:nth-of-type(1) > span:nth-of-type(2) > a')[0].text.lstrip().strip()
         pub_hao = soup.find('table', class_='jcse-service-table').find_all('tr')[0].find_all('td')[
             -1].text.lstrip().strip()