fitur baru, visualisasi, llm, summary

app.py

@@ -1,105 +1,117 @@
+import os
 import pandas as pd
 import numpy as np
 import gradio as gr
 import plotly.graph_objects as go
 from transformers import pipeline
 from huggingface_hub import InferenceClient
-import os
 
-# --- KONFIGURASI PILAR ---
-# Pastikan HF_TOKEN ada di Settings > Secrets di Hugging Face Space kamu
+# --- KONFIGURASI ENGINE ---
 HF_TOKEN = os.getenv("HF_TOKEN")
 client = InferenceClient(model="mistralai/Mistral-7B-Instruct-v0.3", token=HF_TOKEN)
 
-class ArchonFinalEngine:
+class ArchonIndustrialEngine:
     def __init__(self):
-        # Fase 2: AI Classifier (BERT)
-        self.classifier = pipeline("text-classification", model="archon_v1", tokenizer="archon_v1")
+        # Fase 2: AI Classifier (Automasi Kategori)
+        try:
+            self.classifier = pipeline("text-classification", model="archon_v1", tokenizer="archon_v1")
+        except:
+            self.classifier = None
         self.load_data()
 
     def load_data(self):
-        # Fase 1: Single Source of Truth
+        # Fase 1: Data Foundation (Single Source of Truth)
         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 get_risk_analysis(self, customer_id):
-        # --- FASE 3 & 4: DETERMINISTIC CALCULATION (Banking Standard) ---
-        user_txn = self.df_txn[self.df_txn['customer_id'] == customer_id].copy()
-        user_bal = self.df_bal[self.df_bal['customer_id'] == customer_id].sort_values('month')
+    def analyze(self, customer_id):
+        # 1. Validation ID
+        u_txn = self.df_txn[self.df_txn['customer_id'] == customer_id].copy()
+        u_bal = self.df_bal[self.df_bal['customer_id'] == customer_id].sort_values('month')
+        u_rep = self.df_rep[self.df_rep['customer_id'] == customer_id]
         
-        # 1. Expense Ratio Score (30%)
-        income = user_txn[user_txn['transaction_type'] == 'credit']['amount'].sum()
-        expense = user_txn[user_txn['transaction_type'] == 'debit']['amount'].sum()
-        er = expense / income if income > 0 else 1.0
-        er_score = 1 if er > 0.8 else (0.5 if er > 0.5 else 0)
+        if u_txn.empty or u_bal.empty:
+            return None, "ID Nasabah Tidak Ditemukan", None, None
 
-        # 2. Balance Trend Score (20%)
-        bt_score = 1 if len(user_bal) >= 2 and user_bal.iloc[-1]['avg_balance'] < user_bal.iloc[-2]['avg_balance'] else 0
+        # 2. Fase 4: Risk Scoring (Logika Bobot 30/20/20/20/10)
+        income = u_txn[u_txn['transaction_type'] == 'credit']['amount'].sum()
+        expense = u_txn[u_txn['transaction_type'] == 'debit']['amount'].sum()
+        er = expense / income if income > 0 else 1.1
         
-        # 3. Overdraft Score (20%)
-        od_score = 1 if (user_bal['min_balance'] <= 0).any() else 0
-
-        # Formula Final Risk Score
-        # risk_score = 0.3 * ER + 0.2 * BT + 0.2 * OD + 0.2 * MP + 0.1 * Vol
-        final_score = (0.3 * er_score) + (0.2 * bt_score) + (0.2 * od_score) + 0.1 
-        
-        risk_level = "HIGH" if final_score >= 0.7 else ("MEDIUM" if final_score >= 0.4 else "LOW")
+        er_score = 1.0 if er > 0.8 else (0.5 if er > 0.5 else 0.0)
+        bt_score = 1.0 if len(u_bal) >= 2 and u_bal.iloc[-1]['avg_balance'] < u_bal.iloc[-2]['avg_balance'] else 0.0
+        od_score = 1.0 if (u_bal['min_balance'] <= 0).any() else 0.0
+        mp_score = 1.0 if (u_rep['status'] == 'late').any() else 0.0
         
-        return risk_level, final_score, er, user_bal
+        final_score = (0.3 * er_score) + (0.2 * bt_score) + (0.2 * od_score) + (0.2 * mp_score) + 0.1
+        risk_lv = "HIGH" if final_score >= 0.7 else ("MEDIUM" if final_score >= 0.4 else "LOW")
 
-    def generate_llm_advice(self, risk_level, er_ratio, customer_id):
-        # --- FASE 5: GENERATIVE NBO (Action Layer) ---
-        # Menggunakan LLM agar saran tidak kaku seperti robot
-        prompt = f"""
-        Role: Senior Financial Advisor di Bank Indonesia.
-        Nasabah: {customer_id}
-        Kondisi: Risiko {risk_level}, Rasio Pengeluaran {er_ratio:.2%}.
-        Tugas: Berikan 1 paragraf saran yang natural, empati, dan sangat personal dalam Bahasa Indonesia. 
-        Gunakan gaya bahasa profesional tapi bersahabat. Jangan kaku.
+        # 3. Fase 6: Explainable Summary (Ringkasan Laporan)
+        flags = []
+        if er_score == 1: flags.append("⚠️ Pengeluaran Kritis (>80%)")
+        if bt_score == 1: flags.append("📉 Tren Saldo Menurun")
+        if od_score == 1: flags.append("🚫 Saldo Pernah Minus")
+        if mp_score == 1: flags.append("❌ Riwayat Telat Bayar")
+        
+        summary_report = f"""
+        ### 📋 LAPORAN RINGKAS NASABAH ({customer_id})
+        - **Status Risiko**: {risk_lv} (Skor: {final_score:.2f})
+        - **Cashflow**: Pemasukan Rp{income:,.0f} | Pengeluaran Rp{expense:,.0f}
+        - **Sinyal Perilaku**: {', '.join(flags) if flags else '✅ Keuangan Stabil'}
         """
+
+        # 4. Fase 5: Generative NBO (Dynamic Advice)
+        prompt = f"""[INST] Anda adalah penasihat keuangan bank. Nasabah {customer_id} memiliki risiko {risk_lv} dengan rasio pengeluaran {er:.2%}. 
+        Gunakan data ini untuk memberi 1 paragraf saran yang personal, tidak kaku, dan solutif dalam Bahasa Indonesia. [/INST]"""
         try:
-            response = client.chat_completion(messages=[{"role": "user", "content": prompt}], max_tokens=150)
-            return response.choices[0].message.content
+            advice = client.text_generation(prompt, max_new_tokens=150)
         except:
-            return "Tetap pantau pengeluaran Anda agar kesehatan finansial terjaga."
+            advice = "Disarankan melakukan efisiensi pada kategori non-esensial."
 
-    def create_plots(self, user_bal):
-        # --- FASE 6: INSIGHT VISUALIZATION
-        fig = go.Figure()
-        fig.add_trace(go.Scatter(x=user_bal['month'], y=user_bal['avg_balance'], name='Avg Balance', line=dict(color='blue', width=4)))
-        fig.add_trace(go.Bar(x=user_bal['month'], y=user_bal['min_balance'], name='Min Balance', marker_color='orange', opacity=0.6))
-        fig.update_layout(title='Tren Saldo Bulanan', xaxis_title='Bulan', yaxis_title='Jumlah (Rp)', template='plotly_white')
-        return fig
+        # 5. Fase 6: Visualizations (Income vs Expense & Balance Trend)
+        monthly_data = u_txn.groupby(u_txn['date'].dt.to_period('M')).agg(
+            Inflow=('amount', lambda x: x[u_txn.loc[x.index, 'transaction_type'] == 'credit'].sum()),
+            Outflow=('amount', lambda x: x[u_txn.loc[x.index, 'transaction_type'] == 'debit'].sum())
+        ).reset_index()
+        monthly_data['date'] = monthly_data['date'].dt.to_timestamp()
 
-# --- UI LOGIC ---
-engine = ArchonFinalEngine()
+        # Grafik 1: Inflow vs Outflow
+        fig1 = go.Figure()
+        fig1.add_trace(go.Bar(x=monthly_data['date'], y=monthly_data['Inflow'], name='Inflow', marker_color='#2ecc71'))
+        fig1.add_trace(go.Bar(x=monthly_data['date'], y=monthly_data['Outflow'], name='Outflow', marker_color='#e74c3c'))
+        fig1.update_layout(title="Inflow vs Outflow Bulanan", barmode='group', template='plotly_white')
 
-def run_archon(cust_id):
-    risk_lv, score, er, bal_data = engine.get_risk_analysis(cust_id)
-    advice = engine.generate_llm_advice(risk_lv, er, cust_id)
-    plot = engine.create_plots(bal_data)
-    
-    summary = {
-        "Risk Level": risk_lv,
-        "Risk Score": round(score, 2),
-        "Expense Ratio": f"{round(er*100, 2)}%"
-    }
-    return summary, advice, plot
+        # Grafik 2: Balance Trend
+        fig2 = go.Figure()
+        fig2.add_trace(go.Scatter(x=u_bal['month'], y=u_bal['avg_balance'], mode='lines+markers', name='Avg Balance', line=dict(color='#3498db', width=3)))
+        fig2.add_trace(go.Scatter(x=u_bal['month'], y=u_bal['min_balance'], mode='lines', name='Min Balance', line=dict(dash='dot', color='#f1c40f')))
+        fig2.update_layout(title="Tren Saldo (Avg vs Min)", template='plotly_white')
+
+        return summary_report, advice, fig1, fig2
 
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🪙 Archon-AI: Industrial Dashboard")
+# --- UI INTERFACE ---
+engine = ArchonIndustrialEngine()
+
+with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue")) as demo:
+    gr.Markdown("# 🪙 Archon-AI: Financial Resilience Dashboard")
+    
     with gr.Row():
         with gr.Column(scale=1):
-            input_id = gr.Textbox(label="Customer ID", placeholder="C0001")
-            btn = gr.Button("Analyze Customer", variant="primary")
-            out_json = gr.JSON(label="Deterministic Metrics")
+            input_id = gr.Textbox(label="Customer ID", placeholder="C0001 - C0120")
+            btn = gr.Button("Jalankan Analisis AI", variant="primary")
+            report_box = gr.Markdown(label="Summary Report")
+        
         with gr.Column(scale=2):
-            out_plot = gr.Plot(label="Financial Timeline")
-    
-    out_advice = gr.Textbox(label="Archon Generative Advice (LLM)", lines=5)
-    
-    btn.click(fn=run_archon, inputs=input_id, outputs=[out_json, out_advice, out_plot])
+            with gr.Tabs():
+                with gr.TabItem("Cashflow Analysis"):
+                    plot_cash = gr.Plot()
+                with gr.TabItem("Balance History"):
+                    plot_bal = gr.Plot()
+                    
+    advice_box = gr.Textbox(label="NBO (Generative AI)", lines=4)
+
+    btn.click(fn=engine.analyze, inputs=input_id, outputs=[report_box, advice_box, plot_cash, plot_bal])
 
 demo.launch()