如何设置关联主类别列表的带值限制嵌套Pydantic模型？

首先修正你给出的枚举类定义错误，正确的枚举需明确每个成员的取值，建议按主类别拆分不同的子类别枚举：

from enum import Enum
from pydantic import BaseModel, field_validator, ValidationError, Discriminator, Union, Annotated, Literal


# 主类别枚举
class Category(str, Enum):
    MEAT = "Meat"
    VEGETABLE = "Vegetable"
    GRAIN = "Grain"


# 按主类别拆分的子类别枚举
class MeatSubCategory(str, Enum):
    BEEF = "Beef"
    POULTRY = "Poultry"
    SEAFOOD = "Seafood"


class VegetableSubCategory(str, Enum):
    GREEN_VEG = "GreenVegetable"
    ROOT_VEG = "RootVegetable"


class GrainSubCategory(str, Enum):
    WHEAT = "Wheat"
    RICE = "Rice"

方法一：使用字段验证器（field_validator） #

定义包含主类别和子类别的模型，通过验证器关联两者的取值限制：

class ProductCategory(BaseModel):
    category: Category
    sub_category: str

    @field_validator("sub_category")
    def validate_sub_category(cls, v, values):
        # 获取当前传入的主类别
        category = values.data.get("category")
        if not category:
            return v
        
        # 映射主类别到允许的子类别列表
        category_sub_map = {
            Category.MEAT: [item.value for item in MeatSubCategory],
            Category.VEGETABLE: [item.value for item in VegetableSubCategory],
            Category.GRAIN: [item.value for item in GrainSubCategory]
        }
        
        if v not in category_sub_map[category]:
            allowed_subs = ", ".join(category_sub_map[category])
            raise ValueError(f"当主类别为{category}时，子类别只能是：{allowed_subs}")
        return v

测试示例：

# 合法数据
valid_data = ProductCategory(category=Category.MEAT, sub_category="Beef")
print(valid_data)

# 非法数据（主类别是Meat，子类别用了RootVegetable）
try:
    invalid_data = ProductCategory(category=Category.MEAT, sub_category="RootVegetable")
except ValidationError as e:
    print(e)

方法二：使用鉴别联合（Discriminated Union） #

这种方式类型提示更清晰，Pydantic会自动根据主类别字段匹配对应的子模型：

# 为每个主类别创建对应子模型
class MeatCategory(BaseModel):
    category: Literal[Category.MEAT] = Category.MEAT
    sub_category: MeatSubCategory


class VegetableCategory(BaseModel):
    category: Literal[Category.VEGETABLE] = Category.VEGETABLE
    sub_category: VegetableSubCategory


class GrainCategory(BaseModel):
    category: Literal[Category.GRAIN] = Category.GRAIN
    sub_category: GrainSubCategory


# 定义鉴别联合模型
ProductCategory = Annotated[
    Union[MeatCategory, VegetableCategory, GrainCategory],
    Discriminator("category")
]

测试示例：

# 合法数据
valid_meat = ProductCategory(category=Category.MEAT, sub_category=MeatSubCategory.BEEF)
print(valid_meat)

# 非法数据（主类别是Vegetable，子类别用了Beef）
try:
    invalid_data = ProductCategory(category=Category.VEGETABLE, sub_category=MeatSubCategory.BEEF)
except ValidationError as e:
    print(e)

两种方法对比：

• 方法一适合简单场景，修改灵活，无需创建多个子模型
• 方法二更符合类型安全，IDE能提供准确的类型提示，适合复杂的类别体系

内容的提问来源于stack exchange，提问作者Oday Salim