什么是 JSON Schema？
- JSON Schema is a vocabulary that allows you to annotate and validate JSON documents.
- http://json-schema.org/
- 示例，简而言之，json schema就是json数据格式的一个描述，可以用来作实例验证。
// JSON Example
{
"message": "操作成功",
"responseCode": "0",
"hasError": false,
"data": {
"id": 100123456
}
}
// 与之对应的JSON Schema
{
"$schema": "http://json-schema.org/schema#",
"type": "object",
"properties": {
"message": {
"type": "string"
},
"responseCode": {
"type": "string"
},
"hasError": {
"type": "boolean"
},
"data": {
"type": "object",
"properties": {
"id": {
"type": "integer"
}
},
"required": ["id"]
}
},
"required": ["data", "hasError", "message", "responseCode"]
}
接口自动化测试利用 JSON Schema 断言 Response？
接口自动化测试一些常用断言方法：
- status_code：response.status_code == 200
- headers.xxx：Content-Type == application/json
- content.xxx：content.responseCode == "0"、content.message == "操作成功"
今天来个不一样的，Json schema Validate：
#!/usr/bin/env python
# -*- encoding: utf-8 -*-
from jsonschema import validate
#! 正常的返回值
resp_pass = {
"message": "操作成功",
"responseCode": "0",
"hasError": False,
"data": {
"id": 100120384
}
}
#! 错误的返回值，没有返回Data ID
resp_fail = {
"message": "操作成功",
"responseCode": "0",
"hasError": False,
"data": {}
}
#! 预期Schema
resp_schema = {
"$schema": "http://json-schema.org/schema#",
"type": "object",
"properties": {
"message": {
"type": "string"
},
"responseCode": {
"type": "string"
},
"hasError": {
"type": "boolean"
},
"data": {
"type": "object",
"properties": {
"id": {
"type": "integer"
}
},
"required": ["id"]
}
},
"required": ["data", "hasError", "message", "responseCode"]
}
#! 验证成功
try:
validate(resp_pass, resp_schema)
except Exception as e:
print('An exception occurred: ', e)
#! 验证失败
try:
validate(resp_fail, resp_schema)
except Exception as e:
print('An exception occurred: ', e)
以上执行结果为：
# 看得懂哈，不解释了 :)
An exception occurred: 'id' is a required property
Failed validating 'required' in schema['properties']['data']:
{'properties': {'id': {'type': 'integer'}},
'required': ['id'],
'type': 'object'}
On instance['data']:
{}
json schema validate fail
问：这么好用，怎么写预期的 json schema呢？
https://www.liquid-technologies.com/online-json-to-schema-converter
可以”硬写“也可以在线转换，输入预期的返回值json，转换成对应的 json schema
online-json-to-schema-converter
进阶：初次自动生成 JSON Schema用于后续断言？
以上，我们在写接口Case的时候还需要打开额外的网站去转换 json to json schema，有没有更简单的方式？
#!/usr/bin/env python
# -*- encoding: utf-8 -*-
from genson import SchemaBuilder
#! 正确的返回值
resp_pass = {
"message": "操作成功",
"responseCode": "0",
"hasError": False,
"data": {
"id": 100120384
}
}
#! 通过正确的返回值生成预期Json schema
builder = SchemaBuilder()
builder.add_object(resp_pass)
resp_schema = builder.to_json()
print(resp_schema)
""" 输出结果
{
"$schema": "http://json-schema.org/schema#",
"type": "object",
"properties": {
"message": {
"type": "string"
},
"responseCode": {
"type": "string"
},
"hasError": {
"type": "boolean"
},
"data": {
"type": "object",
"properties": {
"id": {
"type": "integer"
}
},
"required": ["id"]
}
},
"required": ["data", "hasError", "message", "responseCode"]
}
"""
我们可以把预期的Json Schema 存在本地（比如CSV或者数据库中），validate时先查询有没有expect schema，有就取出来与response做validation，没有就自动用当前的response生成并保存expect schema（so调试时要确保用正确的返回值生成），后续都拿这个expect schema来validate。把这个过程写成一个方法一直复用，是不是又省时又省力？
from jsonschema import validate
from genson import SchemaBuilder
# 伪代码
def validate_jsonschema(resp_json, expect_flag):
"""
resp_json：response 返回值
expect_flag：去数据库查询对应期望schemat的一个flag，通常用接口名
创建schema逻辑
1. 数据库查询是否有对应flag（对应接口）的期望schema数据；
2. if 有： 直接拿来跟response的json数据做validate
3. if 无： 插入当前这条的响应json数据对应的schema进数据库，assert True！ 所以用例写好调试正常之后，第一次运行确保response正确 然后生成对应的schema并保存
"""
sql_expect = f"SELECT expect_schema from resp_schema WHERE flag ='{expect_flag}'"
with MySqlDB() as db:
expect_schema = db.fetch_one(sql_expect)
if isinstance(resp_json, dict):
# ? 判断response content为dict（json）格式，否则直接False
if expect_schema:
# ? 判断是否存在预期schema
try:
expect_schema = json.loads(expect_schema.get('expect_schema'))
validate(resp_json, expect_schema)
except Exception as e:
logger.log_error(e.text)
assert False
# ? ValidationFailure，Fail the case
else:
builder = SchemaBuilder()
builder.add_object(resp_json)
expect_schema = builder.to_json(indent=2, ensure_ascii=False)
resp_schema = json.dumps(resp_json, indent=2, ensure_ascii=False)
# ? 将respons json转换成expect schema，记录下来
sql_insert = "INSERT INTO resp_schema VALUES (%s, %s, %s)"
with MySqlDB() as db:
db.create_one(sql_insert,
(expect_flag, expect_schema, resp_schema))
assert True
else:
assert False
# ! resp 非dict格式Fail
