Inheritance from base model is erased in post

[vale9003]

Example code:

class BaseMessage(BaseModel):
id_trx: float
date: str
id_cost: float
n_id_issuer: float

@classmethod
def get_validators(cls):
    # yield dict_validator
    yield cls.validate

@classmethod
def validate(cls, value):
    if isinstance(value, cls):
        return value
    else:
        return cls(**dict_validator(value))

class SubMessageOne(BaseMessage):
a: str
b: str

class SubMessageTwo(BaseMessage):
c: str

@app.post("/test")
async def test(m: BaseMessage):
return isinstance(m, SubMessageOne)

POST JSON example:

{
"id_trx": 5,
"date": "20190101",
"id_cost": 250.7,
"n_id_issuer": 45,
"a" : "a uno",
"b" : "b uno"

}

It returns false from test method, I cannot access to SubMessageOne attributes a and b.
I would like to know if exits a workaround to solve this issue or it isn't supported.

Best Regards,
Valentina.

[nkhitrov]

This is the correct behavior, since the type specifies a base model in which a and b are not present. It is also logical that isinstance returns false, since in m is the parent class, not the child. May be you want something like that?

from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()


class BaseMessage(BaseModel):
    id_trx: float
    date: str
    id_cost: float
    n_id_issuer: float


class SubMessageOne(BaseMessage):
    a: str
    b: str


class SubMessageTwo(BaseMessage):
    c: str


@app.post("/test")
async def test(m: SubMessageOne):
    is_inst = isinstance(m, BaseMessage)
    return {"is_inst": is_inst, "model": m}

[tiangolo]

Thanks for the help here @SlyFoxy !

@vale9003 does that solve your use case?

If not, you can create a small self-contained app that shows your code and what you want to achieve. Check this to format it: https://help.github.com/en/github/writing-on-github/creating-and-highlighting-code-blocks . That way we can copy it and run it.

[github-actions]

Assuming the original issue was solved, it will be automatically closed now. But feel free to add more comments or create new issues.

[github-actions]

Assuming the original need was handled, this will be automatically closed now. But feel free to add more comments or create new issues or PRs.