先说下本人是菜鸟，通过网上资料学习，终于调通了MonkeyRunner 实现自动点击截屏后与本地图库进行对比输出，以后做静态UI测试就不需要眼睛盯着看图了，这一切交给MonkeyRunner了。

首先通过monkeyrunner录制脚本 test.mr

test.mr

TOUCH|{'x':265,'y':1232,'type':'downAndUp',} WAIT|{'seconds':5.0,} TOUCH|{'x':355,'y':189,'type':'downAndUp',} WAIT|{'seconds':5.0,} TOUCH|{'x':85,'y':322,'type':'downAndUp',} WAIT|{'seconds':5.0,} TOUCH|{'x':371,'y':893,'type':'downAndUp',} WAIT|{'seconds':5.0,} TOUCH|{'x':371,'y':893,'type':'downAndUp',} WAIT|{'seconds':5.0,} TOUCH|{'x':47,'y':778,'type':'downAndUp',} WAIT|{'seconds':5.0,} TOUCH|{'x':657,'y':88,'type':'downAndUp',} WAIT|{'seconds':5.0,}

然后通过python写自动截屏对比功能 monkey_snapshot.py

#!/usr/bin/env monkeyrunner
# Copyright 2010, The Android Open Source Project
#coding=utf8
import sys from com.android.monkeyrunner import MonkeyDevice from com.android.monkeyrunner import MonkeyImage from com.android.monkeyrunner import MonkeyRunner CMD_MAP = { 'TOUCH': lambda dev, arg: dev.touch(**arg), 'DRAG': lambda dev, arg: dev.drag(**arg), 'PRESS': lambda dev, arg: dev.press(**arg), 'TYPE': lambda dev, arg: dev.type(**arg), 'WAIT': lambda dev, arg: MonkeyRunner.sleep(**arg) } def process_file(fp, device,photo_id): for line in fp: (cmd, rest) = line.split('|') try: rest = eval(rest) except: print 'unable to parse options' continue if cmd not in CMD_MAP: print 'unknown command: ' + cmd continue CMD_MAP[cmd](device, rest) print cmd pic_name='./shot'+str(photo_id)+'.png' result = device.takeSnapshot() #compare picture result2 = MonkeyRunner.loadImageFromFile("E:\MD\shot"+str(photo_id)+".png") print "E:\MD\shot"+str(photo_id)+".png" flag = result.sameAs(result2,0.95) if (flag == False): print 'prapre to save new png '+pic_name result.writeToFile(pic_name,'png') photo_id+=1 print "All PlayBack Action Captured,Total:"+str(photo_id)+' PNG Saved' print "===============================================================" def main(): print "Ok,Now StarRuning MyZXX MonkeyRuner SnopShoot Shell" print "==============================================================="