基于python+appium+yaml安卓UI自动化测试分享

结构介绍

之前分享过一篇安卓UI测试，但是没有实现数据与代码分离，后期维护成本较高，所以最近抽空优化了一下。
不想看文章得可以直接去Github，欢迎拍砖
大致结构如下：

 结构.png

• testyaml管理用例，实现数据与代码分离，一个模块一个文件夹

• public 存放公共文件，如读取配置文件、启动appium服务、读取Yaml文件、定义日志格式等

• page 存放最小测试用例集，一个模块一个文件夹

• results 存放测试报告及失败截图

   report.png

• logs 存放日志

   logs.png logdetail.png

• testcase 存放测试用例
• runtest.py 运行所有测试用例

yaml格式介绍

首先看下yaml文件的格式，之前也写过一点关于yaml语法学习的文章
testcase部分是重点，其中：

• element_info：定位元素信息

• find_type：属性，id、xpath、text、ids

• operate_type: click、sendkeys、back、swipe_up 为back就是返回，暂时就四种

  上面三个必填，operate_type必填!!!!!!

• send_content：send_keys 时用到

• index：ids时用到

• times: 返回次数或者上滑次数

testinfo:
    - id: cm001
      title: 新增终端门店
      execute: 1
testcase:
    -
      element_info: 客户
      find_type: text
      operate_type: click
    -
      element_info: com.fiberhome.waiqin365.client:id/cm_topbar_tv_right
      find_type: id
      operate_type: click
    -
      element_info: com.fiberhome.waiqin365.client:id/custview_id_singletv_inputtext
      find_type: ids
      operate_type: send_keys
      send_content: auto0205
      index: 0
    -
      element_info:
      find_type:
      operate_type: swipe_up
      times: 1
    -
      element_info: 提交
      find_type: text
      operate_type: click
    -
      element_info:
      find_type:
      operate_type: back
      times: 1

代码部分

公共部分

个人觉得核心的就是公共部分，相当于建房子，公共部分搞好了，后面仅仅是调用即可，建房子把架子搭好，后面就添砖加瓦吧。

读取配置文件readconfig.py
设置日志格式logs.py
获取设备GetDevices.py
这几个通用的就不做介绍了

• 读取yaml文件 GetYaml.py
  主要用来读取yaml文件

#coding=utf-8
#author='Shichao-Dong'import sys
reload(sys)
sys.setdefaultencoding('utf8')
import yaml
import codecsclass getyaml:
    def __init__(self,path):
        self.path = path    def getYaml(self):
        '''
        读取yaml文件
        :param path: 文件路径
        :return:
        '''
        try:
            f = open(self.path)
            data =yaml.load(f)
            f.close()
            return data
        except Exception:
            print(u"未找到yaml文件")    def alldata(self):
        data =self.getYaml()
        return data    def caselen(self):
        data = self.alldata()
        length = len(data['testcase'])
        return length    def get_elementinfo(self,i):
        data = self.alldata()
        # print data['testcase'][i]['element_info']
        return data['testcase'][i]['element_info']    def get_findtype(self,i):
        data = self.alldata()
        # print data['testcase'][i]['find_type']
        return data['testcase'][i]['find_type']    def get_operate_type(self,i):
        data = self.alldata()
        # print data['testcase'][i]['operate_type']
        return data['testcase'][i]['operate_type']    def get_index(self,i):
        data = self.alldata()
        if self.get_findtype(i)=='ids':
                    return data['testcase'][i]['index']
        else:
            pass    def get_send_content(self,i):
        data = self.alldata()
        # print data['testcase'][i]['send_content']
        if self.get_operate_type(i) == 'send_keys':
            return data['testcase'][i]['send_content']
        else:
            pass    def get_backtimes(self,i):
        data = self.alldata()
        if self.get_operate_type(i)=='back' or self.get_operate_type(i)=='swipe_up':
                    return data['testcase'][i]['times']
        else:
            pass    def get_title(self):
        data = self.alldata()
        # print data['testinfo'][0]['title']
        return  data['testinfo'][0]['title']

• 启动appium服务 StartAppiumServer.py
  主要是启动appium并返回端口port,这个port在下面的driver中需要

#coding=utf-8
#author='Shichao-Dong'from logs import log
import random,time
import platform
import os
from GetDevices import deviceslog = log()
dev = devices().get_deviceName()class Sp:
    def __init__(self, device):
        self.device = device    def __start_driver(self, aport, bpport):
        """
        :return:
        """
        if platform.system() == 'Windows':
            import subprocess
            subprocess.Popen("appium -p %s -bp %s -U %s" %
                             (aport, bpport, self.device), shell=True)    def start_appium(self):
        """
        启动appium
        p:appium port
        bp:bootstrap port
        :return: 返回appium端口参数
        """
        aport = random.randint(4700, 4900)
        bpport = random.randint(4700, 4900)
        self.__start_driver(aport, bpport)        log.info(
            'start appium :p %s bp %s device:%s' %
            (aport, bpport, self.device))
        time.sleep(10)
        return aport    def main(self):
        """
        :return: 启动appium
        """
        return self.start_appium()    def stop_appium(self):
        '''
        停止appium
        :return:
        '''
        if platform.system() == 'Windows':
            os.popen("taskkill /f /im node.exe")if __name__ == '__main__':
    s = Sp(dev)
    s.main()

• 获取driver GetDriver.py
  platformName、deviceName、appPackage、appActivity这些卸载配置文件config.ini文件中，可以直接通过readconfig.py文件读取获得。
  appium_port有StartAppiumServer.py文件返回

s = Sp(deviceName)
appium_port = s.main()def mydriver():
    desired_caps = {
                'platformName':platformName,'deviceName':deviceName, 'platformVersion':platformVersion,
                'appPackage':appPackage,'appActivity':appActivity,
                'unicodeKeyboard':True,'resetKeyboard':True,'noReset':True
                }
    try:
        driver = webdriver.Remote('http://127.0.0.1:%s/wd/hub'%appium_port,desired_caps)
        time.sleep(4)
        log.info('获取driver成功')
        return driver
    except WebDriverException:
        print 'No driver'if __name__ == "__main__":
    mydriver()

• 重新封装find等命令，BaseOperate.py
  里面主要是一些上滑、返回、find等一些基础操作

#coding=utf-8
#author='Shichao-Dong'from selenium.webdriver.support.ui import WebDriverWait
from logs import log
import os
import time'''
一些基础操作：滑动、截图、点击页面元素等
'''class BaseOperate:
    def __init__(self,driver):
        self.driver = driver    def back(self):
        '''
        返回键
        :return:
        '''
        os.popen("adb shell input keyevent 4")    def get_window_size(self):
        '''
        获取屏幕大小
        :return: windowsize
        '''
        global windowSize
        windowSize = self.driver.get_window_size()
        return windowSize    def swipe_up(self):
        '''
        向上滑动
        :return:
        '''
        windowsSize = self.get_window_size()
        width = windowsSize.get("width")
        height = windowsSize.get("height")
        self.driver.swipe(width/2, height*3/4, width/2, height/4, 1000)    def screenshot(self):
        now=time.strftime("%y%m%d-%H-%M-%S")
        PATH = lambda p: os.path.abspath(
            os.path.join(os.path.dirname(__file__), p)
        )
        screenshoot_path = PATH('../results/screenshoot/')
        self.driver.get_screenshot_as_file(screenshoot_path+now+'.png')    def find_id(self,id):
        '''
        寻找元素
        :return:
        '''
        exsit = self.driver.find_element_by_id(id)
        if exsit :
            return True
        else:
            return False    def find_name(self,name):
        '''
        判断页面是否存在某个元素
        :param name: text
        :return:
        '''
        findname = "//*[@text='%s']"%(name)
        exsit = self.driver.find_element_by_xpath(findname)
        if exsit :
            return True
        else:
            return False    def get_name(self,name):
        '''
        定位页面text元素
        :param name:
        :return:
        '''
        # element = driver.find_element_by_name(name)
        # return element        findname = "//*[@text='%s']"%(name)
        try:
            element = WebDriverWait(self.driver, 10).until(lambda x: x.find_element_by_xpath(findname))
            # element = self.driver.find_element_by_xpath(findname)
            self.driver.implicitly_wait(2)
            return element
        except:
            self.screenshot()
            log.error('未定位到元素：'+'%s')%(name)    def get_id(self,id):
        '''
        定位页面resouce-id元素
        :param id:
        :return:
        '''
        try:
            element = WebDriverWait(self.driver, 10).until(lambda x: x.find_element_by_id(id))
            # element = self.driver.find_element_by_id(id)
            self.driver.implicitly_wait(2)
            return element
        except:
            self.screenshot()
            log.error('未定位到元素：'+'%s')%(id)    def get_xpath(self,xpath):
        '''
        定位页面xpath元素
        :param id:
        :return:
        '''
        try:
            element = WebDriverWait(self.driver, 10).until(lambda x: x.find_element_by_xpath(xpath))
            # element = self.driver.find_element_by_xpath(xpath)
            self.driver.implicitly_wait(2)
            return element
        except:
            self.screenshot()
            log.error('未定位到元素：'+'%s')%(xpath)    def get_ids(self,id):
        '''
        定位页面resouce-id元素组
        :param id:
        :return:列表
        '''
        try:
            # elements = self.driver.find_elements_by_id(id)
            elements = WebDriverWait(self.driver, 10).until(lambda x: x.find_elements_by_id(id))
            self.driver.implicitly_wait(2)
            return elements
        except:
            self.screenshot()
            log.error('未定位到元素：'+'%s')%(id)    def page(self,name):
        '''
        返回至指定页面
        :return:
        '''
        i=0
        while i<10:
            i=i+1
            try:
                findname = "//*[@text='%s']"%(name)
                self.driver.find_element_by_xpath(findname)
                self.driver.implicitly_wait(2)
                break
            except :
                os.popen("adb shell input keyevent 4")
                try:
                    findname = "//*[@text='确定']"
                    self.driver.find_element_by_xpath(findname).click()
                    self.driver.implicitly_wait(2)
                except:
                    os.popen("adb shell input keyevent 4")
                try:
                    self.driver.find_element_by_xpath("//*[@text='工作台']")
                    self.driver.implicitly_wait(2)
                    break
                except:
                    os.popen("adb shell input keyevent 4")

• Operate.py
  我认为最关键的一步了，后面没有page都是调用这个文件进行测试，主要是根据读取的yaml文件，然后进行if…else…判断，根据对应的operate_type分别进行对应的click、sendkeys等操作

#coding=utf-8
#author='Shichao-Dong'from GetYaml import getyaml
from BaseOperate import BaseOperateclass Operate:
    def __init__(self,path,driver):
        self.path = path
        self.driver = driver
        self.yaml = getyaml(self.path)
        self.baseoperate=BaseOperate(driver)    def check_operate_type(self):
        '''
        读取yaml信息并执行
        element_info：定位元素信息
        find_type：属性，id、xpath、text、ids
        operate_type: click、sendkeys、back、swipe_up 为back就是返回，暂时就三种        上面三个必填，operate_type必填!!!!!!        send_content：send_keys 时用到
        index：ids时用到
        times:
        :return:
        '''        for i in range(self.yaml.caselen()):
            if self.yaml.get_operate_type(i) == 'click':
                if self.yaml.get_findtype(i) == 'text':
                    self.baseoperate.get_name(self.yaml.get_elementinfo(i)).click()
                elif self.yaml.get_findtype(i) == 'id':
                    self.baseoperate.get_id(self.yaml.get_elementinfo(i)).click()
                elif self.yaml.get_findtype(i) == 'xpath':
                    self.baseoperate.get_xpath(self.yaml.get_elementinfo(i)).click()
                elif self.yaml.get_findtype(i) == 'ids':
                    self.baseoperate.get_ids(self.yaml.get_elementinfo(i))[self.yaml.get_index(i)].click()            elif self.yaml.get_operate_type(i) == 'send_keys':
                if self.yaml.get_findtype(i) == 'text':
                    self.baseoperate.get_name(self.yaml.get_elementinfo(i)).send_keys(self.yaml.get_send_content(i))
                elif self.yaml.get_findtype(i) == 'id':
                    self.baseoperate.get_id(self.yaml.get_elementinfo(i)).send_keys(self.yaml.get_send_content(i))
                elif self.yaml.get_findtype(i) == 'xpath':
                    self.baseoperate.get_xpath(self.yaml.get_elementinfo(i)).send_keys(self.yaml.get_send_content(i))
                elif self.yaml.get_findtype(i) == 'ids':
                    self.baseoperate.get_ids(self.yaml.get_elementinfo(i))[self.yaml.get_index(i)].send_keys(self.yaml.get_send_content(i))            elif self.yaml.get_operate_type(i) == 'back':
                for n in range(self.yaml.get_backtimes(i)):
                    self.baseoperate.back()            elif self.yaml.get_operate_type(i) == 'swipe_up':
                for n in range(self.yaml.get_backtimes(i)):
                    self.baseoperate.swipe_up()    def back_home(self):
        '''
        返回至工作台
        :return:
        '''
        self.baseoperate.page('工作台')

公共部分的代码就介绍这么多，在编写这个框架的时候，大部分精力都花在这部分，所以个人觉得还是值得好好研究的

Page部分

page部分是最小用例集，一个模块一个文件夹，以客户为例，
目前写了两个用例，一个新增，一个排序，文件如下：

 file.png

代码如下，非常的简洁,

import sys
reload(sys)
sys.setdefaultencoding('utf8')
import codecs,os
from public.Operate import Operate
from public.GetYaml import getyamlPATH = lambda p: os.path.abspath(
    os.path.join(os.path.dirname(__file__), p)
)
yamlpath = PATH("../../testyaml/cm/cm-001addcm.yaml")class AddcmPage:    def __init__(self,driver):
        self.path = yamlpath
        self.driver = driver
        self.operate = Operate(self.path,self.driver)    def operateap(self):
        self.operate.check_operate_type()    def home(self):
        self.operate.back_home()

运行用例

这部分用了unittest,运行所有测试用例和生成报告。
一个模块一个用例，以客户为例：CmTest.py

from page.cm.CmAddcmPage import AddcmPage
from page.cm.CmSortcmPage import SortcmPagefrom public.GetDriver import mydriver
driver = mydriver()import unittest,time
class Cm(unittest.TestCase):    def test_001addcm(self):
        '''
        新增客户
        :return:
        '''
        add = AddcmPage(driver)
        add.operateap()
        add.home()
    def test_002sortcm(self):
        '''
        客户排序
        :return:
        '''
        sort = SortcmPage(driver)
        sort.sortlist()
        sort.home()    def test_999close(self):
        driver.quit()
        time.sleep(10)if __name__ == "__main__":
    unittest.main()

首先从page层将需要运行的用例都import进来，然后用unittest运行即可。
如果想要运行所有的测试用例，需要用到runtest.py

import time,os
import unittest
import HTMLTestRunner
from testcase.CmTest import Cmdef testsuit():
    suite = unittest.TestSuite()
    suite.addTests([unittest.defaultTestLoader.loadTestsFromTestCase(Cm),])    # runner = unittest.TextTestRunner(verbosity=2)
    # runner.run(suite)    now=time.strftime("%y-%m-%d-%H-%M-%S")
    PATH = lambda p: os.path.abspath(
        os.path.join(os.path.dirname(__file__), p)
    )
    dirpath = PATH("./results/waiqin365-")    filename=dirpath + now +'result.html'
    fp=open(filename,'wb')
    runner=HTMLTestRunner.HTMLTestRunner(stream=fp,title='waiqin365 6.0.6beta test result',description=u'result:')    runner.run(suite)
    fp.close()if __name__ =="__main__":
    testsuit()

这边的思路差不多，也是先导入再装入suite即可

总结

就目前而言，暂时算是实现了数据与用例的分离，但是yaml的编写要求较高，不能格式上出错。
同时也有一些其他可以优化的地方，如：

• 对弹窗的判断
• 断开后重连机制
• 失败后重跑机制
  等等，后续可以根据需求进行优化
  最后再贴一下开源地址Github，有兴趣的小伙伴可以去看一下，欢迎拍砖
  备注：完成过程中参考了Louis-me 和 auto 这两个开源项目，感谢！！！

作者：迈阿密小白
链接：https://www.jianshu.com/p/00aff8435a92
來源：简书
著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处