playwright也是可以做接口测试的,但个人觉得还是没有requests库强大,但和selenium相比的话,略胜一筹,毕竟支持API登录,也就是说可以不用交互直接调用接口操作了。
怎么用
既然是API的测试了,那肯定就别搞UI自动化那套,搞什么浏览器交互,那叫啥API测试,纯属扯淡。
也不像有些博主更懒,直接贴的官方例子,难道我用你再帮我复制一次?
来下面,说明下使用playwright如何做API测试?
实例化request对象
示例代码如下:
playwright.request.new_context()
没错,实例化后,就是调API,看吧,其实也不是很难是不是?
实战举栗
这里用我自己写的学生管理系统的部分接口来做演示,并对部分常用api做以说明,代码示例都是用同步的写法。
1、GET请求
示例如下:
def testQueryStudent(playwright: Playwright): """ 查询学生 """ url = 'http://localhost:8090/studentFindById' param = { 'id': 105 } request_context = playwright.request.new_context() response = request_context.get(url=url, params=param) assert response.ok assert response.json() print('n', response.json())
效果:

2、POST请求
示例代码:
def testAddStudent(playwright: Playwright): """ 新增学生 :return: """ url = 'http://localhost:8090/studentAdd' request_body = { "className": "banji", "courseName": "wuli", "email": "ales@qq.com", "name": "ales", "score": 70, "sex": "boy", "studentId": "92908290" } header = {"Content-Type": "application/json"} request_context = playwright.request.new_context() response = request_context.post(url=url, headers=header, data=request_body) assert response.ok assert response.json() print('n', response.json())
效果:

3、PUT请求
示例代码:
def testUpdateStudents(playwright: Playwright): """ 修改学生 """ url = 'http://localhost:8090/studentUpdate/100' param = { 'studentId': "id" + str(100), 'name': "name" + str(100), 'score': 100, "sex": "girl", "className": "class" + str(100), "courseName": "course" + str(100), "email": str(100) + "email@qq.com" } request_context = playwright.request.new_context() response = request_context.put(url=url, form=param) assert response.ok assert response.json() print('n', response.json())
效果:

4、DELETE请求
示例代码:
def testDeleteStudents(playwright: Playwright): """ 删除学生 """ url = 'http://localhost:8090/studentDelete/' + str(105) request_context = playwright.request.new_context() response = request_context.delete(url=url) assert response.ok assert response.json() print('n', response.json())
效果:

5、上传文件
这个是特例吧,按照官方给的方法,我真的是死活也不能成功,一直都是提示上上传文件不能为空,也不到为啥,结果我用了一个替代方案,就是抓包模拟的构造入参,才成功,也是曲折呀。
示例代码:
def test_upload_file(playwright: Playwright): ''' 上传文件 :param playwright: :return: ''' # 创建请求上下文 request_context = playwright.request.new_context() # 定义上传文件的URL upload_url = "http://localhost:8090/fileUpload" # 文件路径 file_path = "d:/demo.txt" # 获取文件名和MIME类型 filename = file_path.split('/')[-1] mime_type, _ = mimetypes.guess_type(file_path) if not mime_type: mime_type = 'application/octet-stream' # 读取文件内容 with open(file_path, 'rb') as file: file_content = file.read() # 构造multipart/form-data的边界字符串 boundary = '---------------------' + str(random.randint(1e28, 1e29 - 1)) # 构造请求体 body = ( f'--{boundary}rn' f'Content-Disposition: form-data; name="file"; filename="{filename}"rn' f'Content-Type: {mime_type}rnrn' f'{file_content.decode("utf-8") if mime_type.startswith("text/") else file_content.hex()}' f'rn--{boundary}--rn' ).encode('utf-8') # 设置请求头 headers = { 'Content-Type': f'multipart/form-data; boundary={boundary}', } # 发起POST请求 response = request_context.post(upload_url, data=body, headers=headers) # 检查响应 assert response.status == 200, f"Upload failed with status: {response.status}" assert response.ok assert response.json() print('n', response.json())
效果:

官方写法:
# 读取文件内容 with open(file_path, 'rb') as file: file_content = file.read() response = request_context.post(upload_url, multipart={ "fileField": { "name": "demo.txt", "mimeType": "text/plain", "buffer": file_content, } }) print('n', response.json())
效果:

官方写法,我不知道为啥,有大侠知道,还请帮忙给个例子,小弟不胜感激呀!
写在最后
我还是觉得微软很强呀,这套框架确实比selenium略胜一筹,综合来看。
终于有时间了,来更新一篇,感觉文章对你有用,转发留言都可,谢谢!
对了,那个上传文件的为啥不行,还请前辈们帮看一下呀!