审计法规 02/06

shenji/sclx.py

+import time
+
+import pandas as pd
 import requests
 from bs4 import BeautifulSoup
 from retry import retry
@@ -69,11 +72,13 @@ def getHref(Keywords):
         href = 'https://sclx.pkulaw.com' + tag.get('url')
     except:
         href = ''
+    time.sleep(1)
     return href
 
 
 @retry(tries=2, delay=5)
 def getData(href):
+    data = []
     ip = baseCore.get_proxy()
     req = requests.get(href, headers=headers, proxies=ip, verify=False)
     req.encoding = req.apparent_encoding
@@ -89,15 +94,18 @@ def getData(href):
             relevance = li.find('div', class_='relevance').text.strip()
         except:
             relevance = ''
-        log.info(f'{publishDate}==={theme}==={relevance}')
-
+        data.append([publishDate,theme,relevance])
+    time.sleep(1)
+    return data
 
 def doJob():
+    data = []
     Keywords = '中华人民共和国公司法(2023修订)'
     href = getHref(Keywords)
     if href:
-        getData(href)
-
+        data += getData(href)
+    df = pd.DataFrame(data)
+    print(df)
 
 if __name__ == '__main__':
     doJob()