Pydantic V2迁移问题：如何检测复杂字段？

在Pydantic V2中，model_fields返回的FieldInfo对象已移除is_complex()方法，需通过字段的核心Schema或注解类型判断是否为复杂字段（通常指嵌套的BaseModel或包含BaseModel的容器类型）。以下是两种可行方案：

方案一：基于Pydantic核心Schema判断（推荐） #

通过获取字段的核心Schema，能准确覆盖更多边界场景：

from pydantic import BaseModel
from pydantic_core import core_schema
import typing as t


def is_field_complex(field) -> bool:
    # 获取字段核心Schema
    if hasattr(field, 'default') and hasattr(field.default, 'get_core_schema'):
        schema = field.default.get_core_schema()
    else:
        schema = field.get_core_schema()
    
    # 判断是否为嵌套BaseModel类型
    if isinstance(schema, core_schema.ModelSchema):
        return True
    
    # 判断是否为包含BaseModel的容器类型（List/Dict等）
    if isinstance(schema, core_schema.ListSchema):
        return isinstance(schema.items_schema, core_schema.ModelSchema)
    if isinstance(schema, core_schema.DictSchema):
        return isinstance(schema.value_schema, core_schema.ModelSchema)
    
    # 可按需扩展Set、Tuple等其他复杂类型判断
    return False


class ClassItems(BaseModel):
    item_id: str


class A(BaseModel):
    id: str
    items: t.List[ClassItems]


for field_name, field in A.model_fields.items():
    if is_field_complex(field):
        print(f"{field_name} is complex")

运行后输出：items is complex

方案二：基于字段注解类型判断（简单直观） #

通过解析字段注解，直接判断是否为BaseModel子类或包含BaseModel的泛型容器：

from pydantic import BaseModel
import typing as t


def is_field_complex(field) -> bool:
    annotation = field.annotation
    
    # 处理泛型容器类型（List[BaseModel]、Dict[str, BaseModel]等）
    if hasattr(annotation, '__origin__'):
        origin = annotation.__origin__
        if origin in (list, t.List, dict, t.Dict, set, t.Set):
            type_args = t.get_args(annotation)
            if type_args and isinstance(type_args[0], type) and issubclass(type_args[0], BaseModel):
                return True
    
    # 处理直接嵌套的BaseModel类型
    return isinstance(annotation, type) and issubclass(annotation, BaseModel)


class ClassItems(BaseModel):
    item_id: str


class A(BaseModel):
    id: str
    items: t.List[ClassItems]


for field_name, field in A.model_fields.items():
    if is_field_complex(field):
        print(f"{field_name} is complex")

运行后输出：items is complex

内容的提问来源于stack exchange，提问作者Lazloo Xp