Selenium应用系列2

关注

原创最新推荐文章于 2026-09-17 18:49:52 发布 2.3k 阅读 AI 写同款· GEO 优化 ›

学点Selenium玩点新鲜~，让分布式测试有更多玩法 standalone 模式能够在一台机器上面执行完整的分布式功能，是 Selenium Grid 的最简单的模式，默认情况下，服务会在 http://localhost:4444 地址监听，我们需要通过 RemoteWebDriver 指向这个地址。上述示例客户端脚本和 Grid 服务都是在本机，当然在实际应用中我们可以将 Grid 服务和脚本分别部署在两台主机上，此时脚本中指向的地址：http://localhost:4444/换成 Grid 服务所在主机的对外 IP 地址即可。阅读详情

一. Selenium背景与概念

Selenium是一个Web应用程序测试工具。Selenium测试直接运行在浏览器中，支持.Net、Java、Perl、Python等不同语言的测试脚本。

因为各大应用平台没有对应的服务发布接口，sdk也没有覆盖相应的功能，所以此处转换了一种思路，使用Selenium操作各应用平台的web页面间接实现app发布的功能。各大应用商像 “360开放平台”,“腾讯应用宝”,“VIVO应用商店”,“OPPO软件商店”,“百度开放平台”,“小米”,“华为”,“魅族”,"阿里应用分发开放平台"等，此处以腾讯应用平台为示例，实现app发布功能

二. Selenium组成

Selenium Hub ：也叫做 Selenium Server，用来接收服务请求，并将收到的请求转发给相应的node执行测试请求。
Selenium Node ：Selenium任务执行的主体，用来接收Server发过来的任务，并调用相应的浏览器执行测试请求。
Python test ：编写任务，并将任务发送给 Selenium Hub。

三. Selenium搭建

1. java环境准备,此处不做详解，Hub/node都需要。

2. Selenium Hub搭建

下载 selenium-server-standalone jar 包：/data/sel/selenium-server-standalone-2.42.2.jar

准备 selenium配置文件 /data/sel/config.json

{
"host": null,
"port": 4444,
"prioritizer": null,
"capabilityMatcher": "org.openqa.grid.internal.utils.DefaultCapabilityMatcher",
"throwOnCapabilityNotPresent": true,
"newSessionWaitTimeout": -1,
"jettyMaxThreads": -1,
"nodePolling": 5000,
"cleanUpCycle": 5000,
"timeout": 30000,
"browserTimeout": 0,
"maxSession": 5,
"unregisterIfStillDownAfter": 30000
}

启动 Selenium Hub服务

#] java -jar /data/sel/selenium-server-standalone-2.42.2.jar -role hub -hubConfig /data/sel/config.json
Sep 29, 2020 5:07:15 PM org.openqa.grid.selenium.GridLauncher main
INFO: Launching a selenium grid server
2020-09-29 17:07:16.404:INFO:osjs.Server:jetty-7.x.y-SNAPSHOT
2020-09-29 17:07:16.427:INFO:osjsh.ContextHandler:started o.s.j.s.ServletContextHandler{/,null}
2020-09-29 17:07:16.435:INFO:osjs.AbstractConnector:Started SocketConnector@0.0.0.0:4444

有监听端口提示后，代表 Selenium Hub服务端搭建启动完成

3. Selenium Node搭建

下载 selenium-server-standalone jar 包：D:\apps\selenium-server-standalone-2.42.2.jar
安装谷歌浏览器
安装 chromedriver.exe 谷歌浏览器驱动：D:\apps\chromedriver.exe

准备 selenium配置文件 D:\apps\config.json

{
  "capabilities": [
    {
      "browserName": "*googlechrome",
      "maxInstances": 1,
      "seleniumProtocol": "Selenium"
    },
    {
      "browserName": "chrome",
      "maxInstances": 1,
      "seleniumProtocol": "WebDriver"
    }
  ],
  "configuration": {
    "proxy": "org.openqa.grid.selenium.proxy.DefaultRemoteProxy",
    "maxSession": 1,
    "port": 5555,
    "register": true,
    "registerCycle": 5000
  }
}

打开cmd，启动 Selenium Node服务

java -Dwebdriver.chrome.driver=D:\apps\chromedriver.exe -jar D:\apps\selenium-server-tandalone-2.42.2.jar -nole node -host 本机地址 -hub http://HUB地址:4444/grid/register -nodeConfig D:\apps\config.json
9月 29, 2020 5:14:34 下午 org.openqa.grid.selenium.GridLauncher main
信息: Launching a selenium grid node
17:14:35.101 INFO - Java: Oracle Corporation 15+36-1562
17:14:35.101 INFO - OS: Windows 10 10.0 amd64
17:14:35.105 INFO - v2.42.2, with Core v2.42.2. Built from revision 6a6995d
17:14:35.144 INFO - RemoteWebDriver instances should connect to: http://127.0.0.1:5555/wd/hub
17:14:35.145 INFO - Version Jetty/5.1.x
17:14:35.145 INFO - Started HttpContext[/selenium-server,/selenium-server]
17:14:35.175 INFO - Started org.openqa.jetty.jetty.servlet.ServletHandler@71c3b41
17:14:35.175 INFO - Started HttpContext[/wd,/wd]
17:14:35.176 INFO - Started HttpContext[/selenium-server/driver,/selenium-server/driver]
17:14:35.176 INFO - Started HttpContext[/,/]
17:14:35.177 INFO - Started SocketListener on 0.0.0.0:5555
17:14:35.177 INFO - Started org.openqa.jetty.jetty.Server@3ce1e309
17:14:35.178 INFO - using the json request : {"capabilities":[{"seleniumProtocol":"Selenium","browserName":"*googlechrome","maxInstances":1,"platform":"XP"},{"seleniumProtocol":"WebDriver","browserName":"chrome","maxInstances":1,"platform":"XP"}],"configuration":{"role":"node","remoteHost":"http://192.168.89.1:5555","hubHost":"192.168.89.133","hubPort":4444,"nodeConfig":"D:\\apps\\config.json","url":"http://192.168.89.1:5555","proxy":"org.openqa.grid.selenium.proxy.DefaultRemoteProxy","hub":"http://192.168.89.133:4444/grid/register","port":5555,"host":"192.168.89.1","maxSession":1,"registerCycle":5000,"register":true},"class":"org.openqa.grid.common.RegistrationRequest"}
17:14:35.178 INFO - Starting auto register thread. Will try to register every 5000 ms.
17:14:35.178 INFO - Registering the node to hub :http://hub地址:4444/grid/register

有这样的提示，代表Selenium node启动成功。

注意：node与hub直接网络互通，如果无法访问，则不能连接

4. Python 任务编写运行

安装selenium
```
#] pip3 install selenium
```

编写python脚本 qqplatform.py

from selenium import webdriver
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
from selenium.webdriver.common.by import By
import time

class QQPlatform():
    def __init__(self):
        self.url = "https://open.qq.com/login"
        self.username = "username"
        self.password = "password"
        self.hub = 'hub地址'
        self.browser = webdriver.Remote(
            command_executor='http://%s:4444/wd/hub'%(self.hub),
            desired_capabilities=DesiredCapabilities.CHROME)
        self.browser.maximize_window()

    def login(self):
        print("开始登录")
        self.browser.get(self.url)
        time.sleep(0.2)
        self.browser.implicitly_wait(30)
        self.browser.switch_to.frame('login_frame')
        self.browser.find_element_by_id("switcher_plogin").click()
        time.sleep(0.2)
        self.browser.find_element_by_id("u").send_keys(self.username)
        self.browser.find_element_by_id("p").send_keys(self.password)
        time.sleep(0.2)
        self.browser.find_element_by_id("login_button").click()
        WebDriverWait(self.browser,15).until(lambda the_driver: the_driver.find_element_by_id('_leftSidebar').is_displayed())
        print("登录成功")

    def upload(self,apk):
        # 新功能
        print("开始上传")
        apkName=apk.split('/')[-1]
        handle = self.browser.current_window_handle
        self.browser.find_element_by_class_name('cover').click()
        self.browser.find_element_by_partial_link_text('更新安装包').click()
        print(self.browser.current_url)
        handles=self.browser.window_handles
        for newhandle in handles:
            if newhandle != handle:
                self.browser.switch_to.window(newhandle)
        print(self.browser.current_url)
        try:
            self.browser.find_element_by_class_name("webuploader-element-invisible").send_keys(apk)
            self.browser.implicitly_wait(20)
        except:
            print("软件包上传失败")
            exit(2)
        # 下一步: 待优化
        time.sleep(30)
        # verifyTextPresent(pattern):校验当前页面是否出现该文字
        # WebDriverWait(self.browser, 30).until(
        #     lambda the_driver: the_driver.find_element(By.XPATH, '//p[text()="%s")]'%(apkName)).is_displayed())
        print("软件包上传成功")

    def submit(self,isTest):
        print("开始发布")
        self.browser.find_element_by_id('j-submit-btn').click()
        time.sleep(3)
        if isTest:
            # 测试打开
            self.browser.find_element_by_class_name('j-confirm-no').click()
        else:
            # 正式使用打开
            self.browser.find_element_by_id('j-confirm-yes').click()
        time.sleep(1)
        print("软件包已发布")

    def close(self):
        time.sleep(2)
        print("退出浏览器")
        time.sleep(2)
        self.browser.quit()
        print("应用发布结束")

if __name__ == '__main__':
    apk="/data/apks/_V_2.0.7_300_2020-09-27.apk"
    isTest=True
    qq = QQPlatform()
    qq.login()
    qq.upload(apk=apk)
    qq.submit(isTest=isTest)
    qq.close()

运行python脚本
```
#] python3 qqplatform.py
```

四. Selenium取消浏览器页面控制提示

selenium打开浏览器后，会有浏览器受selenium控制的提示，添加如下配置，则可隐藏这个提示

from selenium import webdriver
from selenium.webdriver import ChromeOptions
from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
    

option = ChromeOptions()
option.add_experimental_option('excludeSwitches', ['enable-automation'])
browser = webdriver.Remote(
            command_executor='http://hub:4444/wd/hub',
            desired_capabilities=DesiredCapabilities.CHROME,
            options=option
        )

此类配置并不能影响window.navigator.webdriver的值，只要使用selenium，此值仍然为true。

五. Selenium 控制浏览器标签页相关操作

方法	说明
set_window_size()	设置浏览器的大小
maximize_window()	最大化浏览器
refresh()	刷新当前页面
clear()	清除文本
send_keys(value)	模拟按键输入
click()	单击元素
is_displayed()	设置该元素是否用户可见
get_attribute(name)	获取元素属性值
back()	控制浏览器后退
forward()	控制浏览器前进
size	返回元素的尺寸
text	获取元素的文本
switch_to.frame(id)	切换到指定iframe框
current_url	获取当前浏览器url
current_window_handle	获取当前浏览器句柄
window_handles	获取当前浏览器所有句柄
switch_to.window(newhandle)	获取新的浏览器句柄
switch_to.default_content()	返回iframe上一层

六. Selenium 验证码识别技术

1. 介绍

很多网站都会做Selenium自动化识别，你手动登录的时候无需验证码直接登录，但是当你使用Selenium控制登录时就会有各种各样的验证码操作，此处只针对<字母数字干扰>验证码图做讲解

2. 百度图像识别API服务

文档地址： https://ai.baidu.com/ai-doc/REFERENCE/Ck3dwjhhu
vin码识别地址： https://cloud.baidu.com/doc/OCR/s/zk3h7y51e

3. 百度图片识别API服务操作

获取 client_id 和 client_secret
- 登录百度智能云
- 进入控制台 > 图像识别 > 创建应用(例:vin码识别) > 填写相应内容 > 立即创建 > 应用列表 > 根据名称可以找到API Key 和 Secret Key
- client_id 就是 API Key，client_secret 就是 Secret Key

通过API Key和Secret Key获取的access_token

import requests

client_id = "xxx"
client_secret = "jjj"
host = 'https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=%s&client_secret=%s' % (
            client_id, client_secret)
response = requests.get(host)
if response:
  res1 = response.json()
  access_token = res1['access_token']

通过 access_token 获取验证码

import requests
import base64

f = open('/data/03.png', 'rb')
img = base64.b64encode(f.read())
params = {"image": img}

request_url = "https://aip.baidubce.com/rest/2.0/ocr/v1/vin_code"
access_token = "xxxaaafff"
request_url = request_url + "?access_token=" + access_token
headers = {'content-type': 'application/x-www-form-urlencoded'}
response = requests.post(request_url, data=params, headers=headers)
res = response.json()
if "words_result" in res.keys():
  words_result = res['words_result']
  words = words_result[0]['words']

4. Selenium 使用百度图像功能集成验证码识别

# coding = utf-8
from selenium import webdriver
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.common.desired_capabilities import DesiredCapabilities
import time
from PIL import Image, ImageEnhance
import requests
import base64
from selenium.webdriver import ChromeOptions

class TreePlatform():
    def __init__(self):
        ""
        self.url = "http://dev.360.cn/"
        self.username = "xxx"
        self.password = "yyy"
        self.hub = '192.168.89.133'
        self.option = ChromeOptions()
        self.option.add_experimental_option('excludeSwitches', ['enable-automation'])
        self.browser = webdriver.Remote(
            command_executor='http://%s:4444/wd/hub' % (self.hub),
            desired_capabilities=DesiredCapabilities.CHROME,
            options=self.option
        )
        self.browser.maximize_window()
        self.num = 0

    def checkYZM(self):
        self.num += 1
        self.browser.save_screenshot("/data/01.png")
        imgElement = self.browser.find_element_by_class_name('quc-captcha-img')
        ran = Image.open("/data/01.png")
        left = imgElement.location['x']  # 区块截图左上角在网页中的x坐标
        top = imgElement.location['y']  # 区块截图左上角在网页中的y坐标
        right = left + imgElement.size['width']  # 区块截图右下角在网页中的x坐标
        bottom = top + imgElement.size['height']  # 区块截图右下角在网页中的y坐标
        box = (left, top, right, bottom)
        ran.crop(box).save("/data/02.png")
        imageCode = Image.open("/data/02.png")
        sharp_img = ImageEnhance.Contrast(imageCode).enhance(2.0)
        sharp_img.save("/data/03.png")
        sharp_img.load()  # 对比度增强
        time.sleep(2)
        request_url = "https://aip.baidubce.com/rest/2.0/ocr/v1/vin_code"
        # request_url = "https://aip.baidubce.com/rest/2.0/ocr/v1/webimage_loc"
        # 二进制方式打开图片文件
        f = open('/data/03.png', 'rb')
        img = base64.b64encode(f.read())
        params = {"image": img}
        # 获取百度图片验证的access_token
        client_id = "xxxxxx"
        client_secret = "yyyyyy"
        host = 'https://aip.baidubce.com/oauth/2.0/token?grant_type=client_credentials&client_id=%s&client_secret=%s' % (
            client_id, client_secret)
        response = requests.get(host)
        access_token = "afaefefadf"
        if response:
            res1 = response.json()
            access_token = res1['access_token']
        # access_token = '24.6446d03681246f106b5778014157565c.2592000.1588924992.282335-19323883'
        request_url = request_url + "?access_token=" + access_token
        headers = {'content-type': 'application/x-www-form-urlencoded'}
        response = requests.post(request_url, data=params, headers=headers)
        res = response.json()
        print(res)
        if "words_result" in res.keys():
            words_result = res['words_result']
            if len(words_result) >= 1:
                words = words_result[0]['words']
                return words
            else:
                return -1
        else:
            return -1

    def whileCheckYZM(self):
        if self.num >= 20:
            print("尝试次数过多,已退出,请手动发布")
            exit(2)
        result = self.checkYZM()
        print("result", result)
        if result == -1:
            self.browser.find_element_by_class_name("quc-captcha-change-link").click()
            return self.whileCheckYZM()
        else:
            print(result)
            if len(result) <= 3:
                self.browser.find_element_by_class_name("quc-captcha-change-link").click()
                return self.whileCheckYZM()
            self.browser.find_element_by_name('phrase').clear()
            self.browser.find_element_by_name('phrase').send_keys(result)
            time.sleep(3)
            self.browser.find_element_by_class_name("quc-button-sign-in").click()
            time.sleep(3)
            try:
                if self.browser.find_element_by_class_name('quc-tip-error').is_displayed():
                    return self.whileCheckYZM()
            except:
                print("登录成功")

    def login(self):
        print("开始登录")
        self.browser.get(self.url)
        time.sleep(0.2)
        self.browser.implicitly_wait(30)
        self.browser.find_element_by_class_name("js-signIn").click()
        time.sleep(0.2)
        self.browser.find_element_by_name("account").send_keys(self.username)
        self.browser.find_element_by_name("password").send_keys(self.password)
        time.sleep(2)
        # 难点: 输入验证码 请输入验证码 可见,则需要输入验证码
        self.browser.find_element_by_class_name("quc-button-sign-in").click()
        try:
            if self.browser.find_element_by_class_name('quc-tip-error').is_displayed():
                self.whileCheckYZM()
        except:
            print("登录成功")

    def close(self):
        time.sleep(2)
        print("退出浏览器")
        self.browser.quit()

七. Selenium跳过登录验证

1.使用 cookies 屏蔽 selenium 登录验证问题

操作流程
- 打开要测试的网页，获取登录前的cookie（可以抓包获取，可以代码实现，下面会附上代码）。
- 手动登录，再获取登录后的cookie。
- 对比两次获取的cookie，找出登录后多出来的cookie，只要多出来的name和value就行（一般name就是token）。
- 在代码里加上写入cookie，把找出来的name和value写入。然后再写一遍打开网页的代码。
可以删除某个cookie，刷新看看影响登录状态不，如果影响，则需要找出，精简，越少越好
通过浏览器获取当前网站的cookies
- control控制台： document.cookie，将结果复制到字符串中

脚本配置

from selenium import webdriver
from selenium.webdriver import ChromeOptions
from selenium.webdriver.common.desired_capabilities import DesiredCapabilities


option = ChromeOptions()
option.add_experimental_option('excludeSwitches', ['enable-automation'])
bro=webdriver.Remote(
           command_executor='http://192.168.89.144:4444/wd/hub',
           desired_capabilities=DesiredCapabilities.CHROME,
           options=option
       )
bro.get('http://www.xxx.com')
bro.delete_all_cookies()
cookie_str="Hm_lvt_15a5a39cc30c333ba9fae9270351ef30=1603075219,1603075249,1603075436,1603076535; Hm_lpvt_15a5a39cc30c333ba9fae9270351ef30=1603076844; HW_refts_developer_huawei_com/consumer_developer_huawei_com=1602816770937; HW_id_developer_huawei_com/consumer_developer_huawei_com=83c25345ea61434d85d371a96e1f914d; HW_idts_developer_huawei_com/consumer_developer_huawei_com=1602816770938; HW_idn_developer_huawei_com/consumer_developer_huawei_com=0eaa0805460b40e2a03dfccf964e80c7; HW_viewts__developer_huawei_com=1603075212712; _ga=GA1.2.1762325546.1603075273; _gid=GA1.2.69869945.1603075273; urlBeforeLogin=https%3A%2F%2Fdeveloper.huawei.com%2Fconsumer%2Fcn%2F; state=2127401; hwsso_login=""; Hm_lvt_48e5a2ca327922f1ee2bb5ea69bdd0a6=1603075273,1603076561; Hm_lpvt_48e5a2ca327922f1ee2bb5ea69bdd0a6=1603076825; X-HD-SESSION=737a07c5-0b9d-b4c0-22f8-a688197a5093; x-uid=890086000102320858; x-siteId=1; x-hd-grey=0; developer_userinfo=%7B%22siteid%22%3A%221%22%2C%22expiretime%22%3A%2220201019T040721Z%22%2C%22csrftoken%22%3A%225FA3FBE57EE0CA455F2E782A7FD528A84C26D0C31FF2C0F0FE%22%7D; HW_viewts_developer_huawei_com/consumer_developer_huawei_com=1603076842837; x-userType=2; HW_idvc_developer_huawei_com/consumer_developer_huawei_com=43"
cookis_list=cookie_str.split('; ')
for item in cookis_list:
    cookie={u'domain': u'.developer.huawei.com',u'httpOnly': False,u'name': item.split('=')[0],u'path': u'/',u'secure': False,u'value': item.split('=')[1]}
    bro.add_cookie(cookie)
bro.refresh()