微信公众号 06/28

comData/weixin_solo/get_tokenCookies.py

@@ -11,8 +11,9 @@ baseCore = BaseCore()
 log = baseCore.getLogger()
 cnx_ = baseCore.cnx
 cursor_ = baseCore.cursor
-def flushAndGetToken(browser):
 
+
+def flushAndGetToken(browser):
     log.info('======刷新浏览器=====')
     browser.refresh()
     cookie_list = browser.get_cookies()
@@ -28,52 +29,56 @@ def flushAndGetToken(browser):
     soup = BeautifulSoup(info, 'html.parser')
     user_name = soup.find('div', class_='weui-desktop_name').text
 
-    return token,cookies,user_name
+    return token, cookies, user_name
 
 
-if __name__=="__main__":
-    requests.DEFAULT_RETRIES = 5
-    time_start = time.strftime("%Y-%m-%d %H:%M:%S", time.localtime())
-    log.info(f'开始时间为：{time_start}')
+if __name__ == "__main__":
+    while True:
+        requests.DEFAULT_RETRIES = 5
+        time_start = time.strftime("%Y-%m-%d %H:%M:%S", time.localtime())
+        log.info(f'开始时间为：{time_start}')
 
-    requests.adapters.DEFAULT_RETRIES = 3
-    headers = {
-        'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/106.0.0.0 Safari/537.36',
-    }
+        requests.adapters.DEFAULT_RETRIES = 3
+        headers = {
+            'user-agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/106.0.0.0 Safari/537.36',
+        }
 
-    opt = webdriver.ChromeOptions()
-    opt.add_argument(
-        'user-agent=Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.198 Safari/537.36')
+        opt = webdriver.ChromeOptions()
+        opt.add_argument(
+            'user-agent=Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.198 Safari/537.36')
 
-    opt.add_argument("--ignore-certificate-errors")
-    opt.add_argument("--ignore-ssl-errors")
-    opt.add_experimental_option("excludeSwitches", ["enable-automation"])
-    opt.add_experimental_option('excludeSwitches', ['enable-logging'])
-    opt.add_experimental_option('useAutomationExtension', False)
-    opt.binary_location = r'D:\Google\Chrome\Application\chrome.exe'
-    chromedriver = r'D:\cmd100\chromedriver.exe'
-    browser = webdriver.Chrome(chrome_options=opt, executable_path=chromedriver)
-    url = "https://mp.weixin.qq.com/"
-    browser.get(url)
-    # 可改动
-    while True:
-        if browser.current_url != url:
-            break
-        else:
-            time.sleep(10)
+        opt.add_argument("--ignore-certificate-errors")
+        opt.add_argument("--ignore-ssl-errors")
+        opt.add_experimental_option("excludeSwitches", ["enable-automation"])
+        opt.add_experimental_option('excludeSwitches', ['enable-logging'])
+        opt.add_experimental_option('useAutomationExtension', False)
+        opt.binary_location = r'F:\spider\Google\Chrome\Application\chrome.exe'
+        chromedriver = r'F:\spider\cmd100\chromedriver.exe'
+        browser = webdriver.Chrome(chrome_options=opt, executable_path=chromedriver)
+        url = "https://mp.weixin.qq.com/"
+        browser.get(url)
+        # 可改动
+        while True:
+            if browser.current_url != url:
+                break
+            else:
+                time.sleep(10)
 
-    s = requests.session()
-    #获取到token和cookies
-    token, cookies,user_name = flushAndGetToken(browser)
-    print(token,cookies)
-    cookies = json.dumps(cookies)
+        s = requests.session()
+        # 获取到token和cookies
+        token, cookies, user_name = flushAndGetToken(browser)
+        print(token, cookies)
+        cookies = json.dumps(cookies)
 
-    # loadinfo = [token,cookies]
-    #保存到数据库中
-    insert = f"insert into weixin_tokenCookies (token,cookies,create_time,fenghao_time,user_name,update_time) values ('{token}','{escape_string(cookies)}',now(),DATE_SUB(NOW(), INTERVAL 1 DAY),'{user_name}',now())"
-    cursor_.execute(insert)
-    cnx_.commit()
-    browser.close()
+        # loadinfo = [token,cookies]
+        # 保存到数据库中
+        insert = f"insert into weixin_tokenCookies (token,cookies,create_time,fenghao_time,user_name,update_time) values ('{token}','{escape_string(cookies)}',now(),DATE_SUB(NOW(), INTERVAL 1 DAY),'{user_name}',now())"
+        cursor_.execute(insert)
+        cnx_.commit()
+        browser.close()
+        flg = input('是否继续登录')
+        if flg == 'n' or flg == 'N':
+            break
     baseCore.close()
     # s.cookies.update(cookies)
     # s.keep_alive = False

comData/weixin_solo/wxList_.py

@@ -276,6 +276,7 @@ def spider(dic_url, dic_user_count):
 
         total_count = data_json['total_count']
         publish_list = data_json['publish_list']
+        num = 0
         for publish_data in publish_list:
             publish_info = publish_data['publish_info']
             publish_info = json.loads(publish_info)
@@ -293,10 +294,12 @@ def spider(dic_url, dic_user_count):
                         if flgA == 0:
                             flgB = True
                             break
+                        num += 1
                     except:
                         pass
             if flgB:
                 break
+        log.info(f"---{dic_url['name']}--第{int(begin / 5 + 1)}页----新增数：{num}-----使用账号{user_name}--------")
         if flgB:
             break
         if len(linke_list) >= total_count: