如何基于Pydantic模型从单向结构实现无数据库查询的双向访问

需求背景 #

现有单向嵌套的Pydantic模型结构：Receipt包含多个Transactions实例，需要让每个Transactions实例无需查询数据库、无需直接引用Receipt模型，就能访问所属Receipt的receipt_id属性。

示例基础代码 #

from typing import List
from pydantic import BaseModel

data = {
    "receipt_id": 123,
    "transactions": [
        {"transaction_id": 456},
        {"transaction_id": 789}
    ]
}

class Transactions(BaseModel):
    transaction_id: int

class Receipt(BaseModel):
    receipt_id: int
    transactions: List[Transactions]

receipt = Receipt(**data)
# 期望实现：receipt.transactions[0].receipt_id → 返回123

推荐解决方案：利用Pydantic后初始化钩子自动注入属性 #

通过Pydantic的model_post_init生命周期钩子，在Receipt实例初始化完成后，自动为每个嵌套的Transactions实例赋值receipt_id，无需手动传递参数，完全满足需求：

from typing import List
from pydantic import BaseModel

data = {
    "receipt_id": 123,
    "transactions": [
        {"transaction_id": 456},
        {"transaction_id": 789}
    ]
}

class Transactions(BaseModel):
    transaction_id: int
    receipt_id: int  # 新增字段用于存储所属收据ID

class Receipt(BaseModel):
    receipt_id: int
    transactions: List[Transactions]

    def model_post_init(self, __context) -> None:
        # 初始化完成后，为每个交易实例注入receipt_id
        for transaction in self.transactions:
            transaction.receipt_id = self.receipt_id

# 初始化测试
receipt = Receipt(**data)
print(receipt.transactions[0].receipt_id)  # 输出：123

对您提出的两种方案的分析 #

• 方案1：传入基类模型并直接引用属性
  缺点：需要在所有使用Transactions的地方持有Receipt实例的引用，耦合度高，违背了“无需直接引用基类模型”的需求，代码灵活性差。

• 方案2：将基类属性赋值后传入函数
  缺点：需要手动传递receipt_id，容易遗漏，当Transactions实例被多处调用时维护成本高；若Receipt的receipt_id发生变更，Transactions实例的属性无法自动同步（静态数据场景下影响较小，但不够优雅）。

进阶方案：支持父模型属性同步更新 #

如果需要Transactions实例的receipt_id随Receipt实例的属性变更自动同步，可以使用Pydantic的私有属性+属性装饰器实现：

from typing import List
from pydantic import BaseModel, PrivateAttr

class Transactions(BaseModel):
    transaction_id: int
    _receipt: Receipt = PrivateAttr()  # 私有属性，不参与序列化

    @property
    def receipt_id(self):
        return self._receipt.receipt_id

class Receipt(BaseModel):
    receipt_id: int
    transactions: List[Transactions]

    def model_post_init(self, __context) -> None:
        for transaction in self.transactions:
            transaction._receipt = self

# 测试属性同步
receipt = Receipt(**data)
print(receipt.transactions[0].receipt_id)  # 输出：123

# 修改Receipt的receipt_id
receipt.receipt_id = 456
print(receipt.transactions[0].receipt_id)  # 输出：456

内容的提问来源于stack exchange，提问作者Eric