app.py

import pandas as pd
import numpy as np
import gradio as gr
import os
from transformers import pipeline

# --- KONFIGURASI SISTEM ---
MODEL_PATH = "archon_v1" # Folder model BERT hasil training Anda
CATEGORIES = ["groceries", "utilities", "transport", "healthcare", "education", "restaurant", "entertainment"]

class ArchonBankEngine:
    def __init__(self):
        # AI Layer: Automasi Klasifikasi Narasi Transaksi (Tahap 2)
        self.classifier = pipeline("text-classification", model=MODEL_PATH, tokenizer=MODEL_PATH)
        self.load_data()

    def load_data(self):
        # Tahap 1: Data Foundation (Merge Dataset)
        self.df_txn = pd.read_csv('transactions.csv', parse_dates=['date'])
        self.df_cust = pd.read_csv('customers.csv')
        self.df_bal = pd.read_csv('balances_revised.csv', parse_dates=['month'])
        self.df_rep = pd.read_csv('repayments_revised.csv', parse_dates=['due_date'])

    def run_pipeline(self, customer_id):
        # 1. Filter Data Nasabah
        cust_txn = self.df_txn[self.df_txn['customer_id'] == customer_id].copy()
        
        # 2. Tahap 2: AI Intelligence (Automasi Labeling)
        # BERT mendeteksi merchant_category dari narasi yang ambigu
        cust_txn['category'] = cust_txn['raw_description'].apply(lambda x: CATEGORIES[int(self.classifier(x)[0]['label'].split('_')[-1])])
        
        # 3. Tahap 4: Risk Labeling (Early Warning)
        # Bobot: Expense 30%, Trend 20%, Overdraft 20%, Missed Payment 20%
        # (Logika kalkulasi skor sesuai rumus di dokumen Anda)
        risk_score = 0.75 # Simulasi hasil perhitungan
        risk_level = "HIGH" if risk_score >= 0.7 else "LOW"
        
        # 4. Tahap 5: NBO Engine (Action Layer)
        action = "restructuring_suggestion" if risk_level == "HIGH" else "promote_saving"
        
        # 5. Tahap 6: Explainable Summary (Untuk Manajemen)
        summary = f"Nasabah terdeteksi berisiko {risk_level} karena rasio pengeluaran tinggi."
        
        return {
            "ID Nasabah": customer_id,
            "Level Risiko": risk_level,
            "Rekomendasi Aksi": action,
            "Penjelasan": summary
        }

engine = ArchonBankEngine()

# --- DASHBOARD UNTUK MANAJEMEN ---
def manager_dashboard(cust_id):
    return engine.run_pipeline(cust_id)

demo = gr.Interface(
    fn=manager_dashboard,
    inputs=gr.Textbox(label="Masukkan Customer ID (Contoh: C0001)"),
    outputs="json",
    title="🛡️ Archon-AI: Industrial Banking Automation",
    description="Sistem automasi pengelolaan sumber daya informasi bank berdasarkan perilaku nasabah."
)

if __name__ == "__main__":
    demo.launch()