python爬虫--仅供学习，请勿违规使用

from selenium import webdriver

import time

import csv

#专门负责定位

f = open('./淘宝商品.csv','w',encoding='utf-8',newline='')

mywrite = csv.writer(f)

mywrite.writerow(['商品名','价格'])

print('*' * 200)

from selenium.webdriver.common.by import By

#Keys调用键盘事件

from selenium.webdriver.common.keys import Keys

import getpass

#等待

from selenium.webdriver.support.wait import WebDriverWait

#期望

from selenium.webdriver.support import expected_conditions as EC

options = webdriver.ChromeOptions()

# 避免终端下执行代码报错

options.add_experimental_option("excludeSwitches", ['enable-automation',

'enable-logging'])

#不加载图片

#options.add_argument('blink-settings=imagesEnabled=false')

#创建浏览器对象

browser = webdriver.Chrome('./chromedriver.exe',options=options)

browser.maximize_window()

URL = "https://www.taobao.com/"

browser.get(url=URL)

# print(browser.window_handles)

# #打开新的选项卡

# browser.execute_script('window.open()')

# print(browser.window_handles)

#

# #切换选项卡,switch_to切换功能

# browser.switch_to.window(browser.window_handles[1])

# browser.get('https://www.baidu.com')

#点击右边登录按钮

browser.find_element(

By.CSS_SELECTOR,

'body > div.screen-outer.clearfix > '

'div.col-right > div.tbh-member.J_Module > div > div.member-ft > div.member-logout.J_MemberLogout > a.btn-login.ml1.tb-bg.weight').click()

browser.switch_to.window(browser.window_handles[1])

#防止检测

# 防止selenium被监测

# 先修改js，再加载js

browser.execute_cdp_cmd(

"Page.addScriptToEvaluateOnNewDocument",

{

"source": "Object.defineProperty(navigator, 'webdriver', {get: () =>undefined})"

}

)

#显示等待，在指定时间内出现等待的元素，等待结束，如果未出现，超时异常。

WebDriverWait(browser,5).until(

EC.text_to_be_present_in_element((By.CSS_SELECTOR,'#login-form > div.fm-btn > button'),'登录')

)

#寻找账号输入框

number = '19--------3'

password = 'l-------56'

#添加账号

browser.find_element(By.CSS_SELECTOR,'#fm-login-id').send_keys(number)

#添加密码

browser.find_element(By.CSS_SELECTOR,'#fm-login-password').send_keys(password)

#登录

browser.find_element(By.CSS_SELECTOR,'#login-form > div.fm-btn > button').click()

#跳过滑块检测

browser.execute_cdp_cmd(

"Page.addScriptToEvaluateOnNewDocument",

{

"source": "Object.defineProperty(navigator, 'webdriver', {get: () =>undefined})"

}

)

time.sleep(5)

goods = input("请输入商品名")

#在搜索框输入商品

#send_keys,向浏览器定位的某个元素发送信息。（向输入框中输入内容、按键事件等）

browser.find_element(By.CSS_SELECTOR,'#q').send_keys(goods)

#方法一：搜索框直接回车

#直接回车

#browser.find_element(By.CSS_SELECTOR,'#q').send_keys(Keys.ENTER)

#方法二：点击按钮

browser.find_element(By.CSS_SELECTOR,'#J_TSearchForm > div.search-button > button').click()

WebDriverWait(browser, 5).until(

EC.text_to_be_present_in_element((By.CSS_SELECTOR, '#J_SearchForm > button'), '搜索')

)

#向下刷新界面，拉取界面。

y = 0

y_max = 7000

while y < y_max:

y += 500

#调用js实现页面滚动刷新

browser.execute_script(f'window.scrollTo(0,{y})')

WebDriverWait(browser, 5).until(

EC.text_to_be_present_in_element((By.CSS_SELECTOR, '#J_SearchForm > button'), '搜索')

)

#保存所有手机的列表

list_div = browser.find_elements(By.CSS_SELECTOR,'#mainsrp-itemlist > div > div > div:nth-child(1) > div')

for i in list_div:

title = i.find_element(By.CSS_SELECTOR,'div > div.ctx-box.J_MouseEneterLeave.J_IconMoreNew > div.row.row-2 > a').text

price = i.find_element(By.CSS_SELECTOR,'div > div.ctx-box.J_MouseEneterLeave.J_IconMoreNew > div.row.row-1.g-clearfix > div.price.g_price.g_price-highlight').text

print(title,price)

mywrite.writerow([title,price])

#print(list_div,len(list_div))

f.close()

# 版权归个人所有！ 请正确使用