利用selenium爬取企查查企业信息

根据excel中的企业列表数据爬取对应的营业范围信息，并写入excel文件。

from selenium import webdriver
from selenium.webdriver.chrome.service import Service  # 新增
from selenium.webdriver.common.by import By
from selenium.webdriver import Chrome, ChromeOptions
import openpyxl as op
from selenium.webdriver import *
import time
from openpyxl import load_workbook

# 声明调用哪个浏览器，本文使用的是Chrome，其他浏览器同理
#没有把Chromedriver放到python安装路径
service = Service(executable_path=r'C:\Users\Administrator\Desktop\chromedriver.exe')
driver = webdriver.Chrome(service=service)
#爬取网址
url = "https://www.qcc.com"

driver.get(url)
driver.page_source
print("已打开浏览器")
driver.implicitly_wait(60)
print("等待加载完成")
#30s的时间用来在网站打开时手动登陆
time.sleep(30)
print('开始主程序')
time_start = time.time()

#获取公司名称
wb = op.load_workbook(r'C:\Users\Administrator\Desktop\XXXXX.xlsx')
ws = wb['Sheet1']
m_row = ws.max_row
print('一共' + str(m_row) + '个公司需要爬取')
j = 0
for i in range(m_row):
    try:
        company_name = ws.cell(i+1, 1).value

        # 搜索框输入
        driver.find_elements(By.ID, "searchKey")[0].clear()
        driver.find_elements(By.ID, "searchKey")[0].send_keys(company_name)
        #模拟输入回车按键
        from selenium.webdriver.common.keys import Keys
        driver.find_elements(By.CLASS_NAME, "input-group-btn")[0].click()
        driver.implicitly_wait(10)
        time.sleep(2)
        # 获取主窗口
        main_window = driver.current_window_handle
        #点击【公司名称】
        company_title = driver.find_element(By.CLASS_NAME, "copy-value")
        company_title.click()
        #关闭旧窗口
        driver.close()
        # 切换到新窗口
        for window in driver.window_handles:
            if window != main_window:
                driver.switch_to.window(window)
                break
        time.sleep(2)
        spyder_text = driver.find_elements(By.CLASS_NAME, "copy-value")[-1].text
        print(spyder_text)
        ws.cell(i+3, 3).value = spyder_text
        print('第' + str(i+1) + '个爬取完成，还剩' + str(m_row-i) + '个')
    except Exception as e:
        j = j + 1
        print('报错')
        ws.cell(i + 1, 3).value = '报错'
        continue

wb.save(r'C:\Users\Administrator\Desktop\南京API展会企业名单_爬取完成.xlsx')
time_end = time.time()
print('全部爬取并写入完成,失败个数：' + str(j) + ' 总用时' + str('%.2f'%(time_end-time_start)) + '秒')