爬虫逆向滑块验证

selenium滑动验证

import time
from selenium import webdriver
from selenium.webdriver.common.by import By  # 按照什么方式查找，By.ID,By.CSS_SELECTOR
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.support.wait import WebDriverWait  # 等待页面加载某些元素
import cv2 #pip install opencv-python

from urllib import request
from selenium.webdriver.common.action_chains import ActionChains


#获取要滑动的距离
def get_distance():
    #滑动验证码的整体背景图片
    background = cv2.imread("background.png", 0)
    #缺口图片
    gap = cv2.imread("gap.png", 0)

    res = cv2.matchTemplate(background, gap, cv2.TM_CCOEFF_NORMED)
    value = cv2.minMaxLoc(res)[2][0]
    print(value)
    #单位换算
    return value * 278 / 360


def main():
    chrome = webdriver.Chrome(executable_path='./chromedriver')
    chrome.implicitly_wait(5)

    chrome.get('https://passport.jd.com/new/login.aspx?')

    login = chrome.find_element(By.ID, 'pwd-login')
    login.click()

    loginname = chrome.find_element(By.ID, 'loginname')
    loginname.send_keys("xxxx")

    nloginpwd = chrome.find_element(By.ID, 'nloginpwd')
    nloginpwd.send_keys("xxxxx")

    loginBtn = chrome.find_element(By.CLASS_NAME, 'login-btn')
    loginBtn.click()
    #带缺口的大图
    img_src = chrome.find_element(By.XPATH, '//*[@class="JDJRV-bigimg"]/img').get_attribute("src")
    #缺口图片
    temp_src = chrome.find_element(By.XPATH, '//*[@class="JDJRV-smallimg"]/img').get_attribute("src")
    #两张图片保存起来
    request.urlretrieve(img_src, "background.png")
    request.urlretrieve(temp_src, "gap.png")

    distance = int(get_distance())
    print("distance:", distance)

    print('第一步,点击滑动按钮')
    element = chrome.find_element(By.CLASS_NAME, 'JDJRV-slide-btn')
    ActionChains(chrome).click_and_hold(on_element=element).perform()  # 点击鼠标左键，按住不放

    ActionChains(chrome).move_by_offset(xoffset=distance, yoffset=0).perform()
    ActionChains(chrome).release(on_element=element).perform()

    time.sleep(2)
if __name__ == '__main__':
    main()

pyppeteer滑动验证

import random
from pyppeteer import launch
import asyncio
import cv2
from urllib import request


async def get_track():
    background = cv2.imread("background.png", 0)
    gap = cv2.imread("gap.png", 0)

    res = cv2.matchTemplate(background, gap, cv2.TM_CCOEFF_NORMED)
    value = cv2.minMaxLoc(res)[2][0]
    return value * 278 / 360 - 13

async def main():
    browser = await launch({
        # headless指定浏览器是否以无头模式运行，默认是True。
        "headless": False,
        #设置窗口大小
        "args": ['--window-size=1366,768'],
        "executablePath" : '/Applications/Google Chrome.app/Contents/MacOS/Google Chrome'
    })
    # 打开新的标签页
    page = await browser.newPage()
    # 设置页面大小一致
    await page.setViewport({"width": 1366, "height": 768})
    # 访问主页
    await page.goto("https://passport.jd.com/new/login.aspx?")


    # 模拟输入用户名和密码,输入每个字符的间隔时间delay ms
    await page.type("#loginname", '', {
        "delay": random.randint(30, 60)
    })
    await page.type("#nloginpwd", '', {
        "delay": random.randint(30, 60)
    })

    # page.waitFor 通用等待方式，如果是数字，则表示等待具体时间（毫秒）: 等待2秒
    await page.waitFor(2000)
    await page.click("div.login-btn")
    await page.waitFor(2000)
    # page.Jeval（selector，pageFunction）#定位元素，并调用js函数去执行
    #=>表示js的箭头函数：el = function(el){return el.src}
    img_src = await page.Jeval(".JDJRV-bigimg > img", "el=>el.src")
    temp_src = await page.Jeval(".JDJRV-smallimg > img", "el=>el.src")

    request.urlretrieve(img_src, "background.png")
    request.urlretrieve(temp_src, "gap.png")

    # 获取gap的距离
    distance = await get_track()
    """
        # Pyppeteer 三种解析方式
        Page.querySelector()  # 选择器
        Page.querySelectorAll()
        Page.xpath()  # xpath  表达式
        # 简写方式为：
        Page.J(), Page.JJ(), and Page.Jx()
        """
    #定位到滑动按钮标签
    el = await page.J("div.JDJRV-slide-btn")
    # 获取元素的边界框，包含x,y坐标
    box = await el.boundingBox()
    #box={'x': 86, 'y': 34, 'width': 55.0, 'height': 55.0}
    #将鼠标悬停/一定到指定标签位置
    await page.hover("div.JDJRV-slide-btn")
    #按下鼠标
    await page.mouse.down()
    #模拟人的行为进行滑动
    # steps 是指分成几步来完成，steps越大，滑动速度越慢
    #move(x,y)表示将鼠标移动到xy坐标位置
    #random.uniform生成指定范围的随机浮点数
    await page.mouse.move(box["x"] + distance + random.uniform(20, 40),
                          box["y"],
                          {"steps": 100})
    await page.waitFor(1000)

    await page.mouse.up()
    await page.waitFor(2000)


loop = asyncio.get_event_loop()
loop.run_until_complete(main())