"""
Churn Prediction API -- Baseline from P2
Basic serving endpoint without infrastructure.
"""

import json
import joblib
import numpy as np
from fastapi import FastAPI

app = FastAPI()

# Load model at startup
model = joblib.load("model.pkl")

# Load data profile for reference
with open("data_profile.json", "r") as f:
    data_profile = json.load(f)

FEATURE_ORDER = [
    "tenure_months",
    "monthly_charges",
    "total_charges",
    "num_complaints",
    "data_usage_gb",
    "contract_type",
    "payment_method",
    "segment",
]


@app.post("/predict")
def predict(data: dict):
    """Accept a JSON body and return churn prediction."""
    try:
        features = []
        for col in FEATURE_ORDER:
            features.append(data[col])

        # Convert categorical features to numeric (simple encoding)
        contract_map = {"month-to-month": 0, "one-year": 1, "two-year": 2}
        payment_map = {"bank_transfer": 0, "credit_card": 1, "mobile_money": 2, "cash": 3}
        segment_map = {"prepaid": 0, "postpaid": 1}

        processed = [
            features[0],  # tenure_months
            features[1],  # monthly_charges
            features[2],  # total_charges
            features[3],  # num_complaints
            features[4],  # data_usage_gb
            contract_map.get(features[5], 0),
            payment_map.get(features[6], 0),
            segment_map.get(features[7], 0),
        ]

        input_array = np.array([processed])
        prediction = int(model.predict(input_array)[0])
        probability = float(model.predict_proba(input_array)[0][1])

        return {
            "prediction": prediction,
            "probability": round(probability, 4),
        }
    except Exception as e:
        return {"error": str(e)}