如何用Pydantic模型处理带固定格式的动态字段名？

方案一：将动态键转换为子模型列表 #

完全可以实现你想要的结构，通过Pydantic的预处理验证器，把符合response_code_count/xxx格式的键值对转换成子模型列表。

Pydantic v2 实现代码 #

from pydantic import BaseModel, model_validator
from typing import List, Dict, Any

class ResponseCodeCount(BaseModel):
    response_code: str
    count: int

class Response(BaseModel):
    response_code_counts: List[ResponseCodeCount]
    some_other_field: str

    @model_validator(mode='before')
    def process_response_code_counts(cls, data: Dict[str, Any]) -> Dict[str, Any]:
        code_counts = []
        remaining_data = {}
        for key, value in data.items():
            if key.startswith('response_code_count/'):
                # 拆分出响应码部分
                response_code = key.split('/', 1)[1]
                code_counts.append({'response_code': response_code, 'count': value})
            else:
                remaining_data[key] = value
        # 将转换后的列表加入待解析数据
        remaining_data['response_code_counts'] = code_counts
        return remaining_data

# 测试示例
input_data = {
    "response_code_count/R1": 23,
    "response_code_count/PX1": 58,
    "some_other_field": "some_other_value"
}
response = Response(**input_data)
print(response.response_code_counts)
# 输出: [ResponseCodeCount(response_code='R1', count=23), ResponseCodeCount(response_code='PX1', count=58)]
print(response.some_other_field)
# 输出: some_other_value

Pydantic v1 实现代码 #

如果使用Pydantic v1，将model_validator替换为root_validator即可：

from pydantic import BaseModel, root_validator
from typing import List, Dict, Any

class ResponseCodeCount(BaseModel):
    response_code: str
    count: int

class Response(BaseModel):
    response_code_counts: List[ResponseCodeCount]
    some_other_field: str

    @root_validator(pre=True)
    def process_response_code_counts(cls, data: Dict[str, Any]) -> Dict[str, Any]:
        code_counts = []
        remaining_data = {}
        for key, value in data.items():
            if key.startswith('response_code_count/'):
                response_code = key.split('/', 1)[1]
                code_counts.append({'response_code': response_code, 'count': value})
            else:
                remaining_data[key] = value
        remaining_data['response_code_counts'] = code_counts
        return remaining_data

方案二：直接验证并保留动态键结构 #

如果不需要转换为列表，只是要验证动态键的格式并直接访问，可以使用带正则验证的字典字段：

Pydantic v2 实现代码 #

from pydantic import BaseModel
from typing import Dict
from annotated_types import Pattern

class Response(BaseModel):
    # 强制验证键必须符合 `response_code_count/[字母数字]` 的格式
    response_code_counts: Dict[Pattern[r'^response_code_count/[A-Z0-9]+$'], int]
    some_other_field: str

# 测试示例
input_data = {
    "response_code_count/R1": 23,
    "response_code_count/PX1": 58,
    "some_other_field": "some_other_value"
}
response = Response(**input_data)
# 直接通过原始键访问值
print(response.response_code_counts['response_code_count/R1'])
# 输出: 23

这种方式适合需要保留原始键结构的场景，同时确保所有response_code_count相关的键都符合指定格式。

内容的提问来源于stack exchange，提问作者Jossy