模拟网页点击爬虫交管12123违章数据！违章信息及时处理！(1)

2401_83703835

于 2024-04-22 22:46:59 发布

阅读量399

点赞数 4

分类专栏：程序员文章标签：爬虫

本文链接：https://blog.csdn.net/2401_83703835/article/details/138095961

版权

程序员专栏收录该内容

451 篇文章 1 订阅

订阅专栏

driver.find_element_by_xpath(“/html/body/div[6]/div[4]/table/tbody/tr/td/span[1]”).click()#点击

time.sleep(0.5) #加载等待

driver.find_element_by_xpath(“/html/body/div[6]/div[3]/table/tbody/tr[2]/td[1]”).click()#点击

循环依次查询每个车牌违章信息，每次都需要清空上次输入，填写本次查询车牌，识别有多少条数据，共多少页，每页最多展示10条，最后一页有多少条数据

for ii in range(0,nrows):

rowValues= table.row_values(ii) #某一行数据

print(‘正在读取第’+str(ii+1)+‘辆车’)

填写车牌

time.sleep(0.5) #加载等待

elem = driver.find_element_by_xpath(“/html/body/div[3]/div/div[2]/div[1]/div[2]/form/div[3]/div/input”)

elem.clear()

elem.send_keys(rowValues)#输入车牌

time.sleep(0.1) #加载等待

driver.find_element_by_xpath(“/html/body/div[3]/div/div[2]/div[1]/div[2]/form/div[4]/button”).click()#点击查询

time.sleep(0.5) #加载等待

result=driver.find_element_by_xpath(“/html/body/div[3]/div/div[2]/div[2]/div[1]/div/p/span”).text#总违章条数

result=int(result)

a=math.ceil(result/10)#总页数

b=result%10 #除余

读取列表中的数据，其中扣分和罚款需要点击"查看详情"，从弹窗中读取数据’

result1=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[1]”))).text

result2=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[2]”))).text

result3=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[3]”))).text

result4=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[4]”))).text

result5=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[5]”))).text

result6=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[6]”))).text

result7=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[7]”))).text

WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[8]/a”))).click()#查看详情，打开弹窗

time.sleep(1) #加载等待

result8=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//form[@class=‘form-horizontal’]/div[7]/span[2]”))).text

result9=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//form[@class=‘form-horizontal’]/div[8]/span[2]”))).text

result=[result1,result2,result3,result4,result5,result6,result7,result8,result9]

R.append(result)

WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//div[@class=‘modal-footer ui_modal’]/button”))).click()#关闭弹窗

time.sleep(0.5) #加载等待

每读取一辆车的数据就写入表格中

with open(wenjian,‘w’,encoding=‘utf-8’,newline=‘’) as fp:

writer = csv.writer(fp)

writer.writerows® #写入数据

完整代码

from selenium import webdriver

import time

import csv

import datetime

from selenium.webdriver.common.by import By

from selenium.webdriver.support import expected_conditions as EC

from selenium.webdriver.support.wait import WebDriverWait

import math

import xlrd

data = xlrd.open_workbook(‘cheliang.xlsx’)

table = data.sheets()[0]

nrows = table.nrows #行数

ncols = table.ncols #列数

opt = webdriver.ChromeOptions() #创建浏览

#opt.set_headless() #无窗口模式

driver = webdriver.Chrome(options=opt) #创建浏览器对象

driver.maximize_window() #最大化窗口

print(“正在打开网页”)

driver.get(‘’) #打开网页

time.sleep(3) #加载等待

print(“点击单位登录”)

time.sleep(3) #加载等待

driver.find_element_by_xpath(“/html/body/div[1]/div[2]/div/div[2]/div[2]/button”).click()#点击单位登录

time.sleep(3) #加载等待

print(“正在填写账号”)

elem = driver.find_element_by_xpath(“/html/body/div[4]/div/div[1]/div/div[2]/form[1]/div[1]/div/input”)

清空原有内容

elem.clear()

填入账号

elem.send_keys(“”)

time.sleep(1) #加载等待

print(“正在填写密码”)

elem = driver.find_element_by_xpath(“/html/body/div[4]/div/div[1]/div/div[2]/form[1]/div[2]/div/input”)

清空原有内容

elem.clear()

填入密码

elem.send_keys(“”)

time.sleep(1) #加载等待

print(“正在查看验证码”)

driver.find_element_by_xpath(“/html/body/div[4]/div/div[1]/div/div[2]/form[1]/div[3]/div/input”).click()#查看验证码

print(“请输入验证码”)

yanzhengma=input()

time.sleep(1) #加载等待

driver.find_element_by_xpath(“/html/body/div[4]/div/div[1]/div/div[2]/form[1]/div[4]/div/label/input”).click()#勾选

time.sleep(1) #加载等待

填入验证码

elem = driver.find_element_by_xpath(“/html/body/div[4]/div/div[1]/div/div[2]/form[1]/div[3]/div/input”)

elem.clear()

elem.send_keys(str(yanzhengma))

time.sleep(1) #加载等待

print(“正在登陆”)

driver.find_element_by_xpath(“/html/body/div[4]/div/div[1]/div/div[2]/form[1]/div[5]/button”).click()#点击

time.sleep(3) #加载等待

driver.find_element_by_xpath(“/html/body/div[4]/div/div[1]/ul/li[5]/a”).click()#点击违法查询

time.sleep(1) #加载等待

driver.find_element_by_xpath(“/html/body/div[3]/div/div[2]/div[1]/div[2]/form/div[1]/div/div[1]/span/i”).click()#点击选择日期

for i in range(3):

time.sleep(0.5) #加载等待

driver.find_element_by_xpath(“/html/body/div[6]/div[4]/table/thead/tr/th[1]/i”).click()#点击

time.sleep(0.5) #加载等待

driver.find_element_by_xpath(“/html/body/div[6]/div[4]/table/tbody/tr/td/span[1]”).click()#点击

time.sleep(0.5) #加载等待

driver.find_element_by_xpath(“/html/body/div[6]/div[3]/table/tbody/tr[2]/td[1]”).click()#点击

wenjian=datetime.datetime.now().strftime(‘%Y-%m-%d-%H%M%S’) #以开始时间作为数据导出的表格文件名

wenjian=wenjian+‘.csv’

R=[]

for ii in range(0,nrows):

rowValues= table.row_values(ii) #某一行数据

print(‘正在读取第’+str(ii+1)+‘辆车’)

填写车牌

time.sleep(0.5) #加载等待

elem = driver.find_element_by_xpath(“/html/body/div[3]/div/div[2]/div[1]/div[2]/form/div[3]/div/input”)

elem.clear()

elem.send_keys(rowValues)#输入车牌

time.sleep(0.1) #加载等待

driver.find_element_by_xpath(“/html/body/div[3]/div/div[2]/div[1]/div[2]/form/div[4]/button”).click()#点击查询

time.sleep(0.5) #加载等待

result=driver.find_element_by_xpath(“/html/body/div[3]/div/div[2]/div[2]/div[1]/div/p/span”).text#总违章条数

result=int(result)

a=math.ceil(result/10)#总页数

b=result%10 #除余

for i in range(1,a):

for j in range(1,11):

result1=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[1]”))).text

result2=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[2]”))).text

result3=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[3]”))).text

result4=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[4]”))).text

result5=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[5]”))).text

result6=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[6]”))).text

result7=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[7]”))).text

#result1=driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[1]”).text

#result2=driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[2]”).text

#result3=driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[3]”).text

#result4=driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[4]”).text

#result5=driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[5]”).text

#result6=driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[6]”).text

#result7=driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[7]”).text

WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[8]/a”))).click()#查看详情，打开弹窗

time.sleep(1) #加载等待

#driver.find_element_by_xpath(“//table[@id=‘my-msg-list’]/tbody/tr[”+str(j)+“]/td[8]/a”).click()#点击列表中的元素

#time.sleep(0.5) #加载等待

result8=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//form[@class=‘form-horizontal’]/div[7]/span[2]”))).text

result9=WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//form[@class=‘form-horizontal’]/div[8]/span[2]”))).text

#result8=driver.find_element_by_xpath(“//form[@class=‘form-horizontal’]/div[7]/span[2]”).text

#result9=driver.find_element_by_xpath(“//form[@class=‘form-horizontal’]/div[8]/span[2]”).text

result=[result1,result2,result3,result4,result5,result6,result7,result8,result9]

R.append(result)

WebDriverWait(driver,10).until(EC.element_to_be_clickable((By.XPATH,“//div[@class=‘modal-footer ui_modal’]/button”))).click()#关闭弹窗

time.sleep(0.5) #加载等待

#driver.find_element_by_xpath(“//div[@class=‘modal-footer ui_modal’]/button”).click()#点击列表中的元素

#time.sleep(0.5) #加载等待

driver.find_element_by_link_text(“下一页”).click()#翻页

time.sleep(0.5) #加载等待

if b>0:

for j in range(1,b+1):