Fastapi报错RuntimeError: no validator found for ＜class ‘typing.

报错RuntimeError: no validator found for ＜class ‘typing."/>

Fastapi报错RuntimeError: no validator found for ＜class ‘typing.

Fastapi报错RuntimeError: no validator found for <class ‘typing._UnionGenericAlias’>, see `arbitrary_ty

问题描述

使用FastAPI 请求体 - 嵌套模型的时候报错, 不嵌套的时候是可以的

不嵌套校验

class PredictGas(BaseModel):predict_level: strpredict_day_is_heating_season: boolpredict_type: str = Field(default="area", description="预测类型")heating_start_date: Union[datetime, date]heating_end_date: Union[datetime, date]# Optional可选类型  Optional[X]等价于Union[X, None]    int=1等价于Optional[int] = 2   Union[int, float]只能传int  floatmodel: Optional[int] = Field(default=1, title="算法模型id", description="可选不同模型调用", le=2, ge=1)data: List[dict] = []  # List[dict] = [] 等价于 list = []  # 前者可验证list内部第一层数据结构

其中 Optional可选类型 Optional[X]等价于Union[X, None] int=1等价于Optional[int] = 2 Union[int, float]只能传int float

使用嵌套模型 (其中data是第一层列表字典数据)

其中验证流程是 从外到内层 PredictGas —>>> DataBody ----->>> (PredictParams, PastParams)

class PredictParams(BaseModel):# is_heating = Optional[bool]# ts: Union[datetime, date]# low_temp = Union[int, float]# low_temp = float# high_temp = float# is_heating = boolts = datelow_temp: float = 1.0high_temp: float = 10.0is_heating: bool = False# class Config:#     arbitrary_types_allowed = Trueclass PastParams(BaseModel):# ts: Union[datetime, date]# gas_value: float# low_temp = float# high_temp = float# is_heating = boolts: dategas_value: floatlow_temp: float = 1.0high_temp: float = 10.0is_heating: bool = Falseclass DataBody(BaseModel):code: str# predict_params: Union[List[PredictParams], None] =None# predict_params: List  # 不进行内层验证predict_params: List[PredictParams]past_params: List[PastParams]class PredictGas(BaseModel):predict_level: strpredict_day_is_heating_season: boolpredict_type: str = Field(default="area", description="预测类型")heating_start_date: Union[datetime, date]heating_end_date: Union[datetime, date]# Optional可选类型  Optional[X]等价于Union[X, None]    int=1等价于Optional[int] = 2   Union[int, float]只能传int  floatmodel: Optional[int] = Field(default=1, title="算法模型id", description="可选不同模型调用", le=2, ge=1)# data: List[dict] = []  # List[dict] = [] 等价于 list = []  # 前者可验证list内部第一层数据结构# 嵌套模型校验# data: Union[List[Image], None] = Nonedata: List[DataBody]

我的检验数据格式

{"predict_level": "day","predict_day_is_heating_season": False,"predict_type": "area","heating_start_date": "2023-10-10","heating_end_date": "2024-3-10","model": 1,"data": [{"code": "1","predict_params": [{"ts": "2023-10-12","low_temp": -1,"high_temp": 2,"is_heating": True},{"ts": "2023-10-13","low_temp": -1,"high_temp": 2,"is_heating": True}],"past_params": [{"ts": "2023-10-05","gas_value": 1555,"low_temp": -1,"high_temp": 2,"is_heating": True},{"ts": "2023-10-06","gas_value": 2000,"low_temp": -1,"high_temp": 2,"is_heating": True},{"ts": "2023-10-07","gas_value": 2000,"low_temp": -1,"high_temp": 2,"is_heating": True},{"ts": "2023-10-08","gas_value": 1555,"low_temp": -1,"high_temp": 2,"is_heating": True},{"ts": "2023-10-09","gas_value": 1555,"low_temp": -1,"high_temp": 2,"is_heating": True},{"ts": "2023-10-10","gas_value": 1555,"low_temp": -1,"high_temp": 2,"is_heating": True},{"ts": "2023-10-11","gas_value": 1555,"low_temp": -1,"high_temp": 2,"is_heating": True}]}]}

报错情况

当我最内层嵌套的class使用 Union的时候会报错RuntimeError: no validator found for <class ‘typing._UnionGenericAlias’>, see `arbitrary_ty

github提供的解决方案

用了上面的解决方式后fastapi启动不报错,但是接口测试报错

最终定位发现,使用python3.10版本 将Union[int, float]的使用方式换位low_temp = int, | float这种就好了
如果python3.10一下版本的话 使用单种类型验证才可以例如:

class PredictParams(BaseModel):ts = datelow_temp: float = 1.0high_temp: float = 10.0is_heating: bool = False

github错误参考地址