Spaces:
Sleeping
Sleeping
File size: 4,758 Bytes
ee25cbd b79f79f ee25cbd b79f79f ee25cbd 82a3e55 ee25cbd 82a3e55 ee25cbd b79f79f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 | import pandas as pd
import joblib
import uvicorn
from fastapi import FastAPI, Body, Request
from fastapi.responses import JSONResponse, RedirectResponse
from fastapi.exceptions import RequestValidationError
from pydantic import BaseModel, Field
from typing import Literal
# Charger le modèle
model = joblib.load("get_around_v1.pkl")
# Classe Pydantic pour les entrées
class CarData(BaseModel):
mileage: float = Field(..., ge=0, description="Kilométrage doit être >= 0")
engine_power: float = Field(..., ge=1, le=423, description="Puissance moteur entre 1 et 423")
car_brand: Literal[
"Citroën", "Peugeot", "PGO", "Renault", "Audi", "BMW", "Ford", "Mercedes",
"Opel", "Porsche", "Volkswagen", "KIA Motors", "Alfa Romeo", "Ferrari",
"Fiat", "Lamborghini", "Maserati", "Lexus", "Honda", "Mazda", "Mini",
"Mitsubishi", "Nissan", "SEAT", "Subaru", "Suzuki", "Toyota", "Yamaha"
]
fuel: Literal["diesel", "petrol", "hybrid_petrol", "electro"]
paint_color: Literal[
"black", "grey", "white", "red", "silver", "blue",
"orange", "beige", "brown", "green"
]
car_type: Literal[
"convertible", "coupe", "estate", "hatchback", "sedan", "subcompact", "suv", "van"
]
private_parking_available: bool
has_gps: bool
has_air_conditioning: bool
automatic_car: bool
has_getaround_connect: bool
has_speed_regulator: bool
winter_tires: bool
# Initialisation FastAPI
app = FastAPI(
title="GetAround Prediction API",
description="""
Cette API prédit le prix de location journalier d’un véhicule.
👉 Pour les (car_brand, fuel, car_type, paint_color), vous devez choisir une valeur parmi les critères listés, à savoir :
- car_brand : "Citroën", "Peugeot", "PGO", "Renault", "Audi", "BMW", "Ford", "Mercedes",
"Opel", "Porsche", "Volkswagen", "KIA Motors", "Alfa Romeo", "Ferrari",
"Fiat", "Lamborghini", "Maserati", "Lexus", "Honda", "Mazda", "Mini",
"Mitsubishi", "Nissan", "SEAT", "Subaru", "Suzuki", "Toyota", "Yamaha".
- fuel : "diesel", "petrol", "hybrid_petrol", "electro".
- paint_color : "black", "grey", "white", "red", "silver", "blue",
"orange", "beige", "brown", "green"
👉 Pour les autres options (private_parking_available, has_gps, etc.), utiliser true pour Oui et false pour Non.
""",
version="1.0"
)
@app.get("/")
def root():
return RedirectResponse(url="/docs")
# Gestionnaire d'erreurs personnalisé
@app.exception_handler(RequestValidationError)
async def validation_exception_handler(request: Request, exc: RequestValidationError):
errors = []
for e in exc.errors():
field = e.get("loc")[-1]
expected = e.get("ctx", {}).get("expected")
msg = e.get("msg")
if expected:
errors.append({
"field": field,
"message": f"Valeur incorrecte pour '{field}'. Valeurs possibles : {expected}"
})
else:
errors.append({
"field": field,
"message": f"Erreur sur '{field}': {msg}"
})
return JSONResponse(
status_code=422,
content={
"error": "Certaines valeurs ne sont pas valides",
"details": errors
}
)
# Endpoint de prédiction
@app.post(
"/predict",
summary="Prédire le prix journalier d'une voiture",
description="Fournir toutes les caractéristiques de la voiture pour obtenir le prix prédictif."
)
def predict(
data: CarData = Body(
...,
examples={
"valid_example": {
"summary": "Exemple valide",
"value": {
"mileage": 15000,
"engine_power": 100,
"car_brand": "Citroën",
"fuel": "diesel",
"paint_color": "white",
"car_type": "sedan",
"private_parking_available": True,
"has_gps": True,
"has_air_conditioning": True,
"automatic_car": True,
"has_getaround_connect": True,
"has_speed_regulator": True,
"winter_tires": True
}
}
}
)
):
df = pd.DataFrame([data.model_dump()])
df.rename(columns={'car_brand': 'model_key'}, inplace=True)
prediction = model.predict(df)
return {"prediction": round(float(prediction[0]), 2)}
# Lancer localement (utile pour tests)
if __name__ == "__main__":
uvicorn.run(app, host="0.0.0.0", port=7860) |