移动性能测试 Android 获取详细功耗

 目录 

Android系统 >= 5.0

清空耗电数据

adb shell dumpsys batterystats --rest

获取电量数据

adb shell dumpsys batterystats >battery.txt

battery.txt 各个数值的意义参考 https://developer.android.com/studio/profile/battery-historian-charts.html#FilteringBatterystats

其中关键词Estimated power use （估算电量）下有具体uid的耗电，uid根据文本头部通过进程名可以筛选到。

另外，根据uid，还可以找到流量数据，alarm和唤醒次数，等等。

比如：
第一行70.0为微信的耗电

Estimated power use (mAh):    Capacity: 3900, Computed drain: 439, actual drain: 117-156    Uid u0a166: 70.0 ( cpu=1.66 wake=1.69 radio=66.1 sensor=0.540 )    Idle: 45.5    Uid u0a84: 43.4 ( cpu=0.732 wake=0.000292 radio=1.02 gps=0.304 sensor=41.3 )    Uid u0a15: 42.5 ( cpu=0.118 wake=0.120 radio=42.2 )    Uid 0: 42.2 ( cpu=8.27 wake=7.58 radio=26.4 )    Uid 1000: 26.8 ( cpu=8.09 wake=1.24 radio=16.8 gps=0.534 sensor=0.118 )    Uid u0a298: 10.9 ( cpu=3.11 wake=4.49 radio=2.74 sensor=0.560 )

具体统计：

u0a166:  Mobile network: 181.22KB received, 79.26KB sent (packets 488 received, 617 sent)  Mobile radio active: 22m 2s 198ms (29.1%) 76x @ 1197 mspp  Wake lock *alarm*: 2s 75ms partial (48 times) realtime  Wake lock WakerLock:overflow: 2m 51s 280ms partial (220 times) realtime  TOTAL wake: 2m 53s 355ms partial realtime  Sensor 17: 2h 20m 58s 990ms realtime (0 times)  Background for: 7h 33m 26s 672ms   Total running: 7h 33m 26s 672ms 

一段提取耗电量的脚本：

pkgName = "com.xx.xx"ps = subprocess.Popen("adb shell dumpsys batterystats", shell=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)uid = ""uid_tag = Falsepower_tag = Falseuid_re = '=(.+?):"{}"'.format(pkgName)while True:    line = ps.stdout.readline()    if line:        if "Battery History" in line:            uid_tag = True        if "Estimated power use (mAh):" in line:            power_tag = True    else:        break    if uid_tag:        temp = re.findall('=(.+?):"{}"'.format(pkgName), line)        if len(temp) > 0:            uid = temp[0]            print uid            uid_tag = False    if power_tag:        temp = re.findall("Uid {}: (.+?) ".format(uid), line)        if len(temp) > 0:            print temp[0]            power_tag = False

battery-historian

更详细的图表需要battery-historian下的/script/historian.py生成图表。

python historian.py -a battery.txt > battery.html

这里生产的网页查看需要翻墙，因为请求了一些google的js和css，这里把需要的资料放网盘了。（这样本地化也方便集成到其他系统）

百度网盘

一共需要替换两个地方

1. 资源中jsapi最后两行writeLoadTag方法中load资源的路径，改成自己的路径

oogle.loader.writeLoadTag("css", google.loader.ServiceBase + "/api/visualization/1.0/40ff64b1d9d6b3213524485974f36cc0/ui+zh_CN.css", false);google.loader.writeLoadTag("script", google.loader.ServiceBase + "/api/visualization/1.0/40ff64b1d9d6b3213524485974f36cc0/dygraph,format+zh_CN,default+zh_CN,ui+zh_CN,timeline+zh_CN.I.js", false);

1. battery-historian生成的html文件，需要导入所有的资源（两个css，两个js，一个jsapi）

比如：

<script src="https://ajax.googleapis.com/ajax/libs/jquery/1.11.1/jquery.min.js"></script>      <script type="text/javascript" src="https://www.google.com/jsapi?autoload={'modules':[{'name':'visualization','version':'1','packages':['timeline']}]}"></script>

替换成：

<link href="/static/dataProcess/js/battery/tooltip.css" rel="stylesheet" type="text/css">      <script src="/static/dataProcess/js/battery/jquery.min.js"></script>      <script type="text/javascript" src="/static/dataProcess/js/battery/jsapi"></script>      <link href="/static/dataProcess/js/battery/ui+zh_CN.css" type="text/css" rel="stylesheet">      <script src="/static/dataProcess/js/battery/dygraph,format+zh_CN,default+zh_CN,ui+zh_CN,timeline+zh_CN.I.js" type="text/javascript"></script>

一切都可以写成python脚本~美滋滋（historian.py中有很多可以拿来现成用的代码，建议可以看看）

目前刚开始做，准备集成到性能测试的自动化脚本中，求喷求建议

坑：
记录一个坑，在连接usb数据线的情况下，数据不会更新，数据不会更新，数据不会更新。必须拔掉数据线，才会开始记录。

https://testerhome.com/topics/10098