요청과 응답을 위한 Pydantic 스키마

스키마를 사용하는 이유

BaseModel

from pydantic import BaseModel

class PredictRequest(BaseModel):
    features: list[float]

요청 스키마

class PredictRequest(BaseModel):
    features: list[float]
    model_version: str = "latest"

응답 스키마

class PredictResponse(BaseModel):
    prediction: float
    probability: float

종단점에서 스키마 사용하기

from fastapi import FastAPI
app = FastAPI()

@app.post("/predict")
def predict(req: PredictRequest):
    pred = model.predict([req.features])[0]
    return {"prediction": pred, "probability": 0.92}

response_model 매개변수

@app.post("/predict", response_model=PredictResponse)
def predict(req: PredictRequest) -> PredictResponse:
    return PredictResponse(prediction=1.0, probability=0.92)

특성 벡터에 list[float] 사용하기

오류에 HTTPException 사용하기

from fastapi import HTTPException

@app.post("/predict")
def predict(req: PredictRequest):
    if len(req.features) != 4:
        raise HTTPException(status_code=422,
            detail="Expected exactly 4 features")
    return {"prediction": model.predict([req.features])[0]}

필드 검증기

from pydantic import BaseModel, field_validator

class PredictRequest(BaseModel):
    features: list[float]

    @field_validator("features")
    @classmethod
    def not_empty(cls, v):
        if not v:
            raise ValueError("features cannot be empty")
        return v

필드 제약 조건

from pydantic import BaseModel, Field

class PredictRequest(BaseModel):
    features: list[float] = Field(min_length=1)

자동으로 반환되는 422 응답

빠른 확인

복습