from fastapi import FastAPI
from fastapi.responses import RedirectResponse
import joblib
import pandas as pd
import numpy as np
from pydantic import BaseModel

# 1. Configuration de l'API
app = FastAPI(title="Fraud Detection API")

# 2. Chargement du modèle (assure-toi que le nom correspond)
model = joblib.load('fraud_model_hashing.pkl')

# 3. Redirection automatique vers /docs
@app.get("/", include_in_schema=False)
def root():
    return RedirectResponse(url="/docs")

# 4. Schéma des données d'entrée
class Transaction(BaseModel):
    amt: float
    trans_date_trans_time: str
    dob: str
    lat: float
    long: float
    merch_lat: float
    merch_long: float
    city_pop: float
    category: str
    gender: str
    state: str
    merchant: str
    job: str
    cc_num: int

# 5. Ta fonction de préparation (adaptée pour une seule ligne)
def prepare_input(data: dict):
    df = pd.DataFrame([data])
    # Calculs rapides (similaires à ton make_features)
    dt = pd.to_datetime(df["trans_date_trans_time"])
    df["hour"], df["day_of_week"] = dt.dt.hour, dt.dt.dayofweek
    df["day"], df["month"] = dt.dt.day, dt.dt.month
    
    dob = pd.to_datetime(df["dob"])
    df["age"] = ((dt - dob).dt.days / 365.25).astype("float32")
    
    # Distance Haversine
    lat1, lon1 = np.radians(df["lat"]), np.radians(df["long"])
    lat2, lon2 = np.radians(df["merch_lat"]), np.radians(df["merch_long"])
    d = np.sin((lat2-lat1)/2)**2 + np.cos(lat1)*np.cos(lat2)*np.sin((lon2-lon1)/2)**2
    df["distance"] = (6371 * 2 * np.arcsin(np.sqrt(d))).astype("float32")

    # Valeurs par défaut pour les agrégats (car une API reçoit souvent 1 seule transaction)
    df["avg_amt"] = df["amt"] 
    df["std_amt"] = 0.0
    df["nb_trans"] = 1.0

    cols = ['amt', 'hour', 'day_of_week', 'day', 'month', 'age', 'lat', 'long', 
            'city_pop', 'distance', 'avg_amt', 'std_amt', 'nb_trans', 
            'category', 'gender', 'state', 'merchant', 'job']
    return df[cols]

@app.post("/predict")
def predict(data: Transaction):
    X = prepare_input(data.dict())
    prediction = model.predict(X)
    return {"is_fraud": int(prediction[0])}