如何编写验证器方法以使用 python 验证 json 元素数据

Question

我是 python 的新手，正在尝试使用 json 架构编写 python 脚本来验证巨大的 json 输出文件的架构。想要确保我的 json 文件中没有任何空值。

写了一个方法来读取模式 json 文件和输出 json 文件，现在将它们都传递给验证函数。 json 文件中有很多重复对象。然后意识到我应该编写验证器 function/class 来传递每个对象并在 loop.But 中继续验证它们卡在这里不知道该怎么做

{
  "id": "test",
  "name": "name",
  "cake_name": "test",
  "metric": 0.5,
  "anticipations": [
    {
      "time": "2018-01-01 00:00:00",
      "points": 0.49128797804879504,
      "top_properties": {
        "LA:TB2341": 0.23,
        "LA:TB2342": 0.23,
        "LA:TB2343": 0.23
      },
      "status": 0,
      "alert": false
    },
    {
      "time": "2018-01-02 00:00:00",
      "points": 0.588751186433263,
      "top_properties": {
        "LA:TB2342": 0.23,
        "LA:TB2341": 0.23,
        "LA:TB2344": 0.23
      },
      "status": 0,
      "alert": true
    }
  ]
}

PS: 相应的架构文件是从“https://jsonschema.net/”生成的，这是我的moduleschema.json 和以上 json 是 modelout.json

我写的代码只是为了读取文件：

def test_json(self):
    with open('/Users/moduleschema.json', 'r') as json_file:

        schema = json_file.read()
        print(schema)

    with open('/Users/modelout.json', 'r') as output_json:
        outputfile = output_json.read()

        print(outputfile)

    strt = jsonschema.validate(outputfile, schema)
    jsonschema.Draft4Validator
    print(strt)

我想解析 json 文件以确保所有字段都显示正确的类型（整数表示整数，字符串值表示字符串）。我是 python 的新手，如果这是一个愚蠢的问题，请原谅我。谢谢！

Answer 1

所以我将给出一个依赖于我非常喜欢的第三方包的答案。我没有参与其中，但我已经使用过它，它非常有用，尤其是对于此处的验证类型。

是的，您可以创建一个自定义验证器，例如

import json
import typing

# here json_data is the data in your question
def custom_validator(json_data: typing.Dict):
    string_attributes = ["id", "name", "cake_name", "status", "time", "LA:TB2342", "LA:TB2341", "LA:TB2344"]
    int_attributes = [...]
    float_attributes = [...]
    validations_errors = []
    for attribute in string_attributes:
        if attribute in json_data:
            if attribute in string_attributes and not isinstance(json_data.get(attribute), str):
                validations_errors.append(f"key {attribute} is not a string, got {json_data.get(attribute)}")
                ...

这很快就会失控。也许你可以花更多的时间把它弄得漂亮等等

但是，我强烈建议您继续阅读 dataclasses and pydantic

这是我会使用的解决方案

import json
import typing
from pydantic import BaseModel

# if you look closely, this just represents those tiny dictionaries in your list
class Anticipation(BaseModel):
    time: str
    points: float
    top_properties: typing.Dict[str, float]
    status: int
    alert: bool

# this is the whole thing, note how we say that anticipations is a list of those objects we defined above
class Data(BaseModel):
    id: str
    name: str
    cake_name: "str"
    metric: float
    anticipations: typing.List[Anticipation]

json_data = """{
  "id": null,
  "name": "name",
  "cake_name": "test",
  "metric": 0.5,
  "anticipations": [
    {
      "time": "2018-01-01 00:00:00",
      "points": 0.49128797804879504,
      "top_properties": {
        "LA:TB2341": 0.23,
        "LA:TB2342": 0.23,
        "LA:TB2343": 0.23
      },
      "status": 0,
      "alert": false
    },
    {
      "time": "2018-01-02 00:00:00",
      "points": 0.588751186433263,
      "top_properties": {
        "LA:TB2342": 0.23,
        "LA:TB2341": 0.23,
        "LA:TB2344": 0.23
      },
      "status": null,
      "alert": true
    }
  ]
}
"""
data = json.loads(json_data)
data = Data(**data)

我把id改成了null，把status改成了null。如果你运行这个，它会失败并显示这条消息。这是相当有用的

pydantic.error_wrappers.ValidationError: 2 validation errors
id
  none is not an allowed value (type=type_error.none.not_allowed)
anticipations -> 1 -> status
  value is not a valid integer (type=type_error.integer)

显然这意味着您必须安装第 3 方软件包，对于新的 python 编码人员，人们会建议不要这样做。在这种情况下，下面的模板应该会为您指明正确的方向

def validate(my_dict: typing.Dict, string_attributes, int_attributes, float_attributes):
    validations_errors = []
    for attribute in string_attributes:
        if attribute in my_dict:
            if attribute in string_attributes and not isinstance(my_dict.get(attribute), str):
                validations_errors.append(f"key {attribute} is not a string, got {my_dict.get(attribute)}")
            if attribute in int_attributes and not isinstance(my_dict.get(attribute), int):
                # append to the list of errors
                pass
    return validations_errors


def custom_validator(json_data: typing.Dict):
    string_attributes = ["id", "name", "cake_name", "time", "LA:TB2342", "LA:TB2341", "LA:TB2344"]
    int_attributes = [...]
    float_attributes = [...]

    # now do it for anticipations
    validation_errors = validate(json_data, string_attributes, int_attributes, float_attributes)
    for i, anticipation in enumerate(json_data.get('anticipations')):
        validation_error = validate(anticipation, string_attributes, int_attributes, float_attributes)
        if validation_error:
            validation_errors.append(f"anticipation -> {i} error: {validation_error}")

    return validation_errors

data = json.loads(json_data)
custom_validator(data)

输出：['key id is not a string, got None']

您可以在该函数的基础上进行构建

如何编写验证器方法以使用 python 验证 json 元素数据

how to write validator method to validate the json element data with python

python

jsonschema