爬取基金持有股票并存进数据库

#!/usr/bin/python
# -*- coding: utf-8 -*-
# @Author : He who seize the right moment, is the right man
# @Time   : 2021/3/5 10:40

import re
import sys
import time
import datetime
import pymysql
from selenium import webdriver
from selenium.webdriver import Chrome
from selenium.webdriver.chrome.options import Options

opt = Options()
opt.add_argument('--no-sandbox')                        # 解决DevToolsActivePort文件不存在的报错
opt.add_argument('window-size=1920x3000')               # 设置浏览器分辨率
opt.add_argument('--disable-gpu')                       # 谷歌文档提到需要加上这个属性来规避bug
opt.add_argument('--hide-scrollbars')                   # 隐藏滚动条，应对一些特殊页面
opt.add_argument('blink-settings=imagesEnabled=false')  # 不加载图片，提升运行速度
opt.add_argument('--headless')                          # 浏览器不提供可视化界面。Linux下如果系统不支持可视化不加这条会启动失败
driver = Chrome(options=opt)                            # 创建无界面对象
# driver = Chrome(executable_path='/opt/google/chrome/chromedriver',options=opt)

def dbconnections():
    conn = pymysql.connect(host='localhost', user='root', password='rkm2018', port=3306, db='stock', charset='utf8')
    dbs = conn.cursor()

    dbs.execute("SELECT DISTINCT(codes) from tb_fund_sim;")
    results = dbs.fetchall()
    str = ''
    for fundcode in results:
        u = 'http://fund.eastmoney.com/' + str.join(fundcode) + '.html'
        url = u.replace(' ', '')

        doc = open('py.log', 'a+')
        print(url,file=doc)

        driver.get(url)
        data = driver.find_element_by_xpath('//*[@id="position_shares"]/div[1]/table').text
        datas = "'" + data + "'"
        datass = datas.replace('\n','<br/>').replace(' ','&nbsp;&nbsp;&nbsp;&nbsp;').replace('-','<font size=5 color="green">下跌</font>').replace('相关资讯','').replace('股吧','')

        sql = "update tb_fund_sim set holdstock=" + datass + " where codes=" + str.join(fundcode)
        dbs.execute(sql)
        conn.commit()
    dbs.close()
    conn.close()
    print(datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')+"，完成刷新。",file=doc)
    doc.close()

if __name__ == '__main__':
    while True:
        today = datetime.datetime.now().weekday() + 1
        if today <= 5:
            # 范围时间
            d_time = datetime.datetime.strptime(str(datetime.datetime.now().date()) + '9:00', '%Y-%m-%d%H:%M')
            d_time1 = datetime.datetime.strptime(str(datetime.datetime.now().date()) + '15:00', '%Y-%m-%d%H:%M')
            # 当前时间
            n_time = datetime.datetime.now()
            # 判断当前时间是否在范围时间内
            if n_time > d_time and n_time < d_time1:
                dbconnections()
                time.sleep(20)
                # driver.close()
            else:
                print(False)
        else:
            doc = open('py.log', 'a+')
            print("今天是星期："+str(today)+"，不进行持有股获取，休眠Python Script.",file=doc)
            doc.close()
            time.sleep(480)