python自动化测试简易封装

背景

工作中遇到一些类似责任链试的全链路测试场景，总是得临时写一些测试脚本，刚好觉得这种场景和「junit」的代码编写类似，于是考虑用「python」开发支持下

细节说明

主体代码

代码量很小，直接贴出来：

import sys
import traceback
import time
import inspect
from functools import wraps

AUTO_TEST_PREFIX = "autotest"
SPLIT = "_"


def test(message=""):
    def decorate(func):
        @wraps(func)
        def wrapper(*args, **kwargs):
            print("-" * 80)
            start = time.time()
            print("start testing, api: %s" % func.__name__)
            print("detail message: ", message)
            print("=" * 20 + "body start" + "=" * 20)
            res = func(*args, **kwargs)
            print("=" * 20 + "body   end" + "=" * 20)
            print("run cost: %fs" % (time.time() - start))
            return res

        return wrapper

    return decorate


def start(name):
    mod = sys.modules[name]
    if mod:
        def func_filter(func):
            return SPLIT in func.__name__ and func.__name__.startswith(AUTO_TEST_PREFIX)

        def func_sort(func):
            pairs = func.__name__.split(SPLIT)
            if len(pairs) == 2:
                return int(pairs[1])
            return 0

        funcs = inspect.getmembers(mod, inspect.isfunction)
        funcs = [func[1] for func in funcs if func_filter(func[1])]
        for func in sorted(funcs, key=func_sort):
            try:
                res = func()
                if type(res) is bool:
                    if not res:
                        print("error: api %s run failed!" % func.__name__)
                        break
                    else:
                        print("info: api %s run success!" % func.__name__)
                elif not res:
                    print("warn: api %s with no return, you may need to check." % func.__name__)
                else:
                    print("info: api %s run with result:%s" % (func.__name__, str(res)))
            except Exception:
                traceback.print_exc()
                break
            finally:
                print("\n\n")

其实就是利用装饰器，黑盒处理了自动化测试的流程

用法

引入两个主题代码的两个「func」，「test」类似「junit」的注释，「start」为执行入口
此外可以简单对「func」添加下接口说明，便于理解接口功能，在自动化测试时方便快速定位

from auto_test import test
from auto_test import start


@test(message="login")
def autotest_0():
    print("process about login")

@test(message="create instance")
def autotest_1():
    print("process about create")
    return True

@test(message="config")
def autotest_2():
    print("process about config")
    return False

@test(message="destroy instance")
def autotest_3():
    print("process about destroy")


start(__name__)

查看执行结果:

--------------------------------------------------------------------------------
start testing, api: autotest_0
detail message:  login
====================body start====================
process about login
====================body   end====================
run cost: 0.000025s
warn: api autotest_0 with no return, you may need to check.

--------------------------------------------------------------------------------
start testing, api: autotest_1
detail message:  create instance
====================body start====================
process about create
====================body   end====================
run cost: 0.000022s
info: api autotest_1 run success!

--------------------------------------------------------------------------------
start testing, api: autotest_2
detail message:  config
====================body start====================
process about config
====================body   end====================
run cost: 0.000020s
error: api autotest_2 run failed!

按照接口顺序额打印出了各个接口的信息，执行耗时以及接口的输出
整个测试链路仅执行prefix为「autotest」的「func」,执行顺序按照subfix的编号，当且仅当该「func」执行不抛异常且执行结果不为「false」时才会继续流转
再针对具体的工作场景，沉淀一些「tools」,「utils」包，后面就交给他人填充「case」就好了，干净整洁