现在,大多公司都采用 soa 架构。那么我们怎么知道 soa 提供哪些服务方法呢?是通过浏览器不断的点击尝试?还是通过 wcf?还是通过阅读 soa 接口文档,在软件开发行业,总会有文档落后于代码的情况?这些手法都太拙劣了。其实我们可以通过一些更快捷高效的手段达到目的。
萌生这个想法的背景是,我们公司部署了几个 soa 服务。然后我们测试自动化非常分散,形式各种各样,五花八门,简直就是八仙过海各显神通,非常不易于统计自动化覆盖率统计。于是乎,就想到了使用脚本扫描统计的办法。不多说了,直接上代码,脚本如下:
#coding=utf-8
import requests
from lxml import etree
from suds.client import Client
import csv
preSoa = r"http://192.168.1.20:8085/"
soaurl =requests.get(preSoa)
detail_html = etree.HTML(soaurl.content)
links =detail_html.xpath(ur"//a")
hrefs = [preSoa +x.attrib["href"] for x in links if ".config" not in x.text and "bin" not in x.text and "Demo" not in x.text and ".asax" not in x.text ]
svclinks =[]
def getsvcurl(templink):
tempsoaurlcontent = requests.get(templink).content
temphtml = etree.HTML(tempsoaurlcontent)
temphtmllink =temphtml.xpath(ur"//a/text()")
svclink = [templink + x + "?singleWsdl" for x in temphtmllink if ".svc" in x]
[svclinks.append(x) for x in svclink]
[getsvcurl(x) for x in hrefs]
getsvcurl(preSoa)
print "svc count:{0}".format(len(svclinks))
client = Client(svclinks[0])
svcfuns =[]
def dealsvcmethods(svclink):
methods=[x[0] for x in Client(svclink).sd[0].ports[0][1]]
[svcfuns.append({"funcname":x,"svc":svclink}) for x in methods]
[dealsvcmethods(x) for x in svclinks]
with open(ur"E:\主站soa2016-10-25.csv", 'ab+') as csvfile :
writer = csv.DictWriter(csvfile,fieldnames = ["funcname","svc"])
writer.writeheader()
writer.writerows(svcfuns)
当然,使用这种方式是有前提条件的。首先,soa 站点必须要公开服务。有了这个,我们能够很容易找出前后版本新增的服务方法。