Update app.py

app.py

@@ -1,4 +1,3 @@
-
 import gradio as gr
 import pandas as pd
 import numpy as np
@@ -8,233 +7,382 @@ from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor, A
 from sklearn.linear_model import Ridge
 from sklearn.svm import SVR
 from sklearn.preprocessing import StandardScaler
+import matplotlib.pyplot as plt
+import matplotlib.patches as mpatches
+from matplotlib.gridspec import GridSpec
+import matplotlib.ticker as mticker
 import warnings
 warnings.filterwarnings('ignore')
 
-def get_forex_data(symbol="EURUSD=X", period="60d"):
-    """Forex verilerini indir"""
+# ─── Stil sabitleri ───────────────────────────────────────────────────────────
+BG       = "#0A0E17"
+PANEL    = "#111827"
+BORDER   = "#1E293B"
+ACCENT   = "#3B82F6"
+GREEN    = "#10B981"
+RED      = "#EF4444"
+AMBER    = "#F59E0B"
+MUTED    = "#64748B"
+TEXT     = "#F1F5F9"
+SUBTEXT  = "#94A3B8"
+
+MODEL_COLORS = ["#3B82F6", "#8B5CF6", "#10B981", "#F59E0B", "#EC4899"]
+
+# ─── Veri & feature mühendisliği ─────────────────────────────────────────────
+def get_forex_data(symbol="EURUSD=X", period="90d"):
     try:
         data = yf.download(symbol, period=period, interval="1d", progress=False)
+        if data is None or len(data) < 40:
+            return None
         return data
-    except Exception as e:
+    except Exception:
         return None
 
 def create_features(data):
-    """Teknik göstergeler ve özellikler oluştur"""
     df = data.copy()
-    
-    # Temel özellikler
-    df['Returns'] = df['Close'].pct_change()
-    df['High_Low'] = df['High'] - df['Low']
+    df['Returns']    = df['Close'].pct_change()
+    df['High_Low']   = df['High'] - df['Low']
     df['Close_Open'] = df['Close'] - df['Open']
-    
-    # Hareketli ortalamalar
-    df['SMA_5'] = df['Close'].rolling(window=5).mean()
-    df['SMA_10'] = df['Close'].rolling(window=10).mean()
-    df['SMA_20'] = df['Close'].rolling(window=20).mean()
-    
-    # RSI
+
+    for w in [5, 10, 20, 50]:
+        df[f'SMA_{w}'] = df['Close'].rolling(window=w).mean()
+
     delta = df['Close'].diff()
-    gain = (delta.where(delta > 0, 0)).rolling(window=14).mean()
-    loss = (-delta.where(delta < 0, 0)).rolling(window=14).mean()
-    rs = gain / loss
-    df['RSI'] = 100 - (100 / (1 + rs))
-    
-    # MACD
+    gain  = (delta.where(delta > 0, 0)).rolling(14).mean()
+    loss  = (-delta.where(delta < 0, 0)).rolling(14).mean()
+    df['RSI'] = 100 - (100 / (1 + gain / loss))
+
     exp1 = df['Close'].ewm(span=12, adjust=False).mean()
     exp2 = df['Close'].ewm(span=26, adjust=False).mean()
-    df['MACD'] = exp1 - exp2
-    
-    # Bollinger Bands
-    df['BB_middle'] = df['Close'].rolling(window=20).mean()
-    df['BB_std'] = df['Close'].rolling(window=20).std()
-    df['BB_upper'] = df['BB_middle'] + (df['BB_std'] * 2)
-    df['BB_lower'] = df['BB_middle'] - (df['BB_std'] * 2)
-    
-    # Lag features
-    for i in range(1, 6):
+    df['MACD']   = exp1 - exp2
+    df['Signal'] = df['MACD'].ewm(span=9, adjust=False).mean()
+
+    df['BB_middle'] = df['Close'].rolling(20).mean()
+    df['BB_std']    = df['Close'].rolling(20).std()
+    df['BB_upper']  = df['BB_middle'] + df['BB_std'] * 2
+    df['BB_lower']  = df['BB_middle'] - df['BB_std'] * 2
+    df['BB_width']  = (df['BB_upper'] - df['BB_lower']) / df['BB_middle']
+
+    df['ATR'] = (df['High'] - df['Low']).rolling(14).mean()
+
+    for i in range(1, 8):
         df[f'Lag_{i}'] = df['Close'].shift(i)
-    
-    df = df.dropna()
-    return df
-
-def train_ensemble_models(df):
-    """5 farklı model eğit"""
-    feature_cols = ['Returns', 'High_Low', 'Close_Open', 'SMA_5', 'SMA_10', 
-                    'SMA_20', 'RSI', 'MACD', 'BB_middle', 'BB_upper', 'BB_lower',
-                    'Lag_1', 'Lag_2', 'Lag_3', 'Lag_4', 'Lag_5']
-    
-    X = df[feature_cols].values
+
+    return df.dropna()
+
+FEATURE_COLS = [
+    'Returns','High_Low','Close_Open',
+    'SMA_5','SMA_10','SMA_20','SMA_50',
+    'RSI','MACD','Signal',
+    'BB_middle','BB_upper','BB_lower','BB_width',
+    'ATR',
+    'Lag_1','Lag_2','Lag_3','Lag_4','Lag_5','Lag_6','Lag_7'
+]
+
+# ─── Model katmanı ────────────────────────────────────────────────────────────
+MODELS_DEF = {
+    'Random Forest':     RandomForestRegressor(n_estimators=200, random_state=42, n_jobs=-1),
+    'Gradient Boosting': GradientBoostingRegressor(n_estimators=200, learning_rate=0.05, random_state=42),
+    'AdaBoost':          AdaBoostRegressor(n_estimators=150, random_state=42),
+    'Ridge':             Ridge(alpha=0.5),
+    'SVR':               SVR(kernel='rbf', C=200, gamma=0.05, epsilon=0.0001),
+}
+
+def train_and_evaluate(df):
+    X = df[FEATURE_COLS].values
     y = df['Close'].values
-    
-    # Train/test split (son 5 g��n test için)
-    split_idx = -5
-    X_train, X_test = X[:split_idx], X[split_idx:]
-    y_train, y_test = y[:split_idx], y[split_idx:]
-    
-    # Normalizasyon
+    split = -10
+
+    X_train, X_test = X[:split], X[split:]
+    y_train, y_test = y[:split], y[split:]
+
     scaler = StandardScaler()
-    X_train_scaled = scaler.fit_transform(X_train)
-    X_test_scaled = scaler.transform(X_test)
-    
-    # 5 Ensemble Model
-    models = {
-        'Random Forest': RandomForestRegressor(n_estimators=100, random_state=42),
-        'Gradient Boosting': GradientBoostingRegressor(n_estimators=100, random_state=42),
-        'AdaBoost': AdaBoostRegressor(n_estimators=100, random_state=42),
-        'Ridge': Ridge(alpha=1.0),
-        'SVR': SVR(kernel='rbf', C=100, gamma=0.1)
-    }
-    
-    predictions = {}
-    for name, model in models.items():
-        model.fit(X_train_scaled, y_train)
-        pred = model.predict(X_test_scaled)
-        predictions[name] = pred
-    
-    return models, scaler, predictions, y_test, feature_cols
-
-def predict_next_day(models, scaler, df, feature_cols):
-    """Ertesi gün tahmini yap"""
-    latest_features = df[feature_cols].iloc[-1].values.reshape(1, -1)
-    latest_features_scaled = scaler.transform(latest_features)
-    
-    predictions = {}
-    for name, model in models.items():
-        pred = model.predict(latest_features_scaled)[0]
-        predictions[name] = pred
-    
-    # Ensemble ortalaması
-    ensemble_pred = np.mean(list(predictions.values()))
-    
-    return predictions, ensemble_pred
+    Xtr = scaler.fit_transform(X_train)
+    Xte = scaler.transform(X_test)
+
+    trained, test_preds = {}, {}
+    for name, m in MODELS_DEF.items():
+        m.fit(Xtr, y_train)
+        test_preds[name] = m.predict(Xte)
+        trained[name] = m
+
+    return trained, scaler, test_preds, y_test
+
+def predict_next(trained, scaler, df):
+    x = df[FEATURE_COLS].iloc[-1].values.reshape(1, -1)
+    xs = scaler.transform(x)
+    preds = {n: float(m.predict(xs)[0]) for n, m in trained.items()}
+    ensemble = float(np.mean(list(preds.values())))
+    return preds, ensemble
+
+# ─── Grafik ───────────────────────────────────────────────────────────────────
+def build_figure(df, next_preds, ensemble_pred, test_preds, y_test):
+    plt.rcParams.update({
+        'font.family': 'monospace',
+        'text.color': TEXT,
+        'axes.labelcolor': SUBTEXT,
+        'xtick.color': SUBTEXT,
+        'ytick.color': SUBTEXT,
+        'figure.facecolor': BG,
+        'axes.facecolor': PANEL,
+        'axes.edgecolor': BORDER,
+        'grid.color': BORDER,
+        'grid.linewidth': 0.5,
+    })
+
+    fig = plt.figure(figsize=(14, 12))
+    gs = GridSpec(3, 2, figure=fig, hspace=0.45, wspace=0.35)
+
+    ax1 = fig.add_subplot(gs[0, :])   # Fiyat + BB
+    ax2 = fig.add_subplot(gs[1, 0])   # RSI
+    ax3 = fig.add_subplot(gs[1, 1])   # MACD
+    ax4 = fig.add_subplot(gs[2, :])   # Model tahminleri + hata
+
+    tail = df.tail(45)
+    idx  = np.arange(len(tail))
+    cp   = float(df['Close'].iloc[-1])
+
+    # ── Panel 1: Fiyat ──────────────────────────────────────────────
+    ax1.fill_between(idx, tail['BB_lower'].values, tail['BB_upper'].values,
+                     alpha=0.08, color=ACCENT, label='Bollinger Bandı')
+    ax1.plot(idx, tail['BB_upper'].values, lw=0.5, color=ACCENT, alpha=0.4)
+    ax1.plot(idx, tail['BB_lower'].values, lw=0.5, color=ACCENT, alpha=0.4)
+    ax1.plot(idx, tail['Close'].values,    lw=1.8, color=TEXT,   label='Kapanış')
+    ax1.plot(idx, tail['SMA_20'].values,   lw=1.1, color=AMBER,  alpha=0.8, label='SMA 20', ls='--')
+    ax1.plot(idx, tail['SMA_50'].values,   lw=1.1, color=MUTED,  alpha=0.8, label='SMA 50', ls=':')
+
+    # Ensemble tahmini noktası
+    arrow_x = len(idx) + 0.5
+    ax1.annotate(
+        f'▶ {ensemble_pred:.5f}',
+        xy=(len(idx) - 1, cp),
+        xytext=(len(idx) + 1, ensemble_pred),
+        arrowprops=dict(arrowstyle='->', color=GREEN, lw=1.5),
+        color=GREEN, fontsize=9, fontweight='bold',
+    )
+    ax1.axhline(y=cp, color=MUTED, lw=0.8, ls='--', alpha=0.5)
+
+    ax1.set_title('EUR/USD  ·  Son 45 Gün  +  Sonraki Gün Tahmini',
+                  fontsize=11, color=TEXT, pad=10, loc='left', fontweight='bold')
+    ax1.legend(fontsize=7, framealpha=0.15, labelcolor=SUBTEXT, ncol=4)
+    ax1.set_ylabel('Fiyat', fontsize=8)
+    ax1.yaxis.set_major_formatter(mticker.FormatStrFormatter('%.4f'))
+
+    # ── Panel 2: RSI ─────────────────────────────────────────────────
+    rsi_tail = tail['RSI'].values
+    ax2.plot(idx, rsi_tail, lw=1.4, color=ACCENT)
+    ax2.fill_between(idx, rsi_tail, 50, where=(rsi_tail >= 50), alpha=0.15, color=GREEN)
+    ax2.fill_between(idx, rsi_tail, 50, where=(rsi_tail < 50),  alpha=0.15, color=RED)
+    ax2.axhline(70, color=RED,   lw=0.8, ls='--', alpha=0.6)
+    ax2.axhline(30, color=GREEN, lw=0.8, ls='--', alpha=0.6)
+    ax2.set_ylim(0, 100)
+    ax2.set_title('RSI (14)', fontsize=9, color=TEXT, pad=6, loc='left')
+    ax2.set_ylabel('RSI', fontsize=8)
+    for level, lbl in [(70, 'OB'), (30, 'OS')]:
+        ax2.text(len(idx), level, lbl, color=SUBTEXT, fontsize=7, va='center')
 
+    # ── Panel 3: MACD ────────────────────────────────────────────────
+    macd   = tail['MACD'].values
+    signal = tail['Signal'].values
+    hist   = macd - signal
+    colors_hist = [GREEN if v >= 0 else RED for v in hist]
+    ax3.bar(idx, hist, color=colors_hist, alpha=0.5, width=0.8, label='Histogram')
+    ax3.plot(idx, macd,   lw=1.3, color=ACCENT, label='MACD')
+    ax3.plot(idx, signal, lw=1.1, color=AMBER,  label='Signal', ls='--')
+    ax3.axhline(0, color=MUTED, lw=0.7)
+    ax3.set_title('MACD (12/26/9)', fontsize=9, color=TEXT, pad=6, loc='left')
+    ax3.legend(fontsize=7, framealpha=0.1, labelcolor=SUBTEXT)
+
+    # ── Panel 4: Model karşılaştırması ───────────────────────────────
+    names  = list(next_preds.keys())
+    values = [next_preds[n] for n in names]
+    maes   = [float(np.mean(np.abs(test_preds[n] - y_test))) for n in names]
+
+    x_pos = np.arange(len(names))
+    bar_colors = [GREEN if v > cp else RED for v in values]
+
+    bars = ax4.bar(x_pos, values, color=bar_colors, alpha=0.75, width=0.55, zorder=3)
+    ax4.axhline(y=cp, color=TEXT, lw=1.2, ls='--', alpha=0.7, label=f'Mevcut {cp:.5f}', zorder=4)
+    ax4.axhline(y=ensemble_pred, color=ACCENT, lw=1.5, ls='-', alpha=0.9,
+                label=f'Ensemble {ensemble_pred:.5f}', zorder=4)
+
+    for bar, val, mae in zip(bars, values, maes):
+        pct = (val - cp) / cp * 100
+        sign = '+' if pct >= 0 else ''
+        ax4.text(bar.get_x() + bar.get_width()/2, val + 0.00003,
+                 f'{sign}{pct:.3f}%\nMAE {mae:.5f}',
+                 ha='center', va='bottom', fontsize=7.5, color=TEXT, fontweight='bold')
+
+    ax4.set_xticks(x_pos)
+    ax4.set_xticklabels(names, fontsize=9)
+    ax4.set_title('5 Model Tahmini  vs  Mevcut Fiyat', fontsize=9, color=TEXT, pad=6, loc='left')
+    ax4.legend(fontsize=8, framealpha=0.15, labelcolor=SUBTEXT)
+    ax4.yaxis.set_major_formatter(mticker.FormatStrFormatter('%.4f'))
+
+    fig.suptitle('EUR/USD  ·  ENSEMBLE TAHMİN SİSTEMİ',
+                 fontsize=13, color=TEXT, fontweight='bold', y=0.98)
+
+    return fig
+
+# ─── Ana tahmin fonksiyonu ───────────────────────────────────────────────────
 def predict_eurusd():
-    """Ana tahmin fonksiyonu"""
+    import traceback
     try:
-        # Veri indir
         data = get_forex_data()
-        if data is None or len(data) == 0:
-            return "❌ Veri indirilemedi. Lütfen internet bağlantınızı kontrol edin.", None
-        
-        # Özellikler oluştur
-        df = create_features(data)
-        
-        # Modelleri eğit
-        models, scaler, test_predictions, y_test, feature_cols = train_ensemble_models(df)
-        
-        # Ertesi gün tahmini
-        next_day_preds, ensemble_pred = predict_next_day(models, scaler, df, feature_cols)
-        
-        # Mevcut fiyat
-        current_price = float(df['Close'].iloc[-1])
-        
-        # Sonuçları hazırla
-        result = f"""🔄 GÜNCELLEME ZAMANI: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}
-
-📊 MEVCUT EUR/USD FİYATI: {current_price:.5f}
-
-🎯 ERTESİ GÜN TAHMİNLERİ:
+        if data is None:
+            return "❌ Veri indirilemedi.", None
+
+        df      = create_features(data)
+        trained, scaler, test_preds, y_test = train_and_evaluate(df)
+        next_preds, ensemble = predict_next(trained, scaler, df)
+
+        cp = float(df['Close'].iloc[-1])
+        ens_chg = (ensemble - cp) / cp * 100
+        direction = "YUKARI ▲" if ens_chg > 0 else "AŞAĞI ▼"
+        dir_color = "📈" if ens_chg > 0 else "📉"
+
+        rsi_val  = float(df['RSI'].iloc[-1])
+        macd_val = float(df['MACD'].iloc[-1])
+        sig_val  = float(df['Signal'].iloc[-1])
+        rsi_status = "Aşırı Alım" if rsi_val > 70 else ("Aşırı Satım" if rsi_val < 30 else "Nötr")
+        macd_cross = "Boğa" if macd_val > sig_val else "Ayı"
+
+        lines = [
+            f"━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━",
+            f"  EUR/USD  ·  ENSEMBLE TAHMİN RAPORU",
+            f"  {datetime.now().strftime('%d %b %Y  %H:%M:%S')}",
+            f"━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━",
+            f"",
+            f"  MEVCUT FİYAT   :  {cp:.5f}",
+            f"  ENSEMBLE       :  {ensemble:.5f}  ({ens_chg:+.3f}%)  {dir_color} {direction}",
+            f"",
+            f"  ─── TEKNİK DURUM ───────────────────────────",
+            f"  RSI (14)       :  {rsi_val:.1f}  →  {rsi_status}",
+            f"  MACD Sinyali   :  {macd_cross} kesişimi",
+            f"  BB Pozisyonu   :  {float(df['BB_lower'].iloc[-1]):.5f}  ↔  {float(df['BB_upper'].iloc[-1]):.5f}",
+            f"",
+            f"  ─── MODEL TAHMİNLERİ ───────────────────────",
+        ]
+
+        for name, val in next_preds.items():
+            chg = (val - cp) / cp * 100
+            mae = float(np.mean(np.abs(test_preds[name] - y_test)))
+            arrow = "▲" if chg > 0 else "▼"
+            lines.append(f"  {name:<20s}:  {val:.5f}  ({chg:+.3f}%)  {arrow}  MAE={mae:.5f}")
+
+        lines += [
+            f"",
+            f"  ─── BACKTEST (Son 10 Gün) ──────────────────",
+        ]
+
+        ens_test = np.mean([test_preds[n] for n in test_preds], axis=0)
+        ens_mae  = float(np.mean(np.abs(ens_test - y_test)))
+        ens_rmse = float(np.sqrt(np.mean((ens_test - y_test) ** 2)))
+        hit_rate = float(np.mean(np.sign(np.diff(np.append([y_test[0]], ens_test))) ==
+                                  np.sign(np.diff(y_test)))) * 100
+
+        lines += [
+            f"  Ensemble MAE   :  {ens_mae:.6f}",
+            f"  Ensemble RMSE  :  {ens_rmse:.6f}",
+            f"  Yön Doğruluğu  :  {hit_rate:.1f}%",
+            f"",
+            f"━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━",
+            f"  ⚠  Yalnızca eğitim amaçlıdır. Finansal tavsiye değildir.",
+            f"━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━",
+        ]
+
+        output_text = "\n".join(lines)
+        fig = build_figure(df, next_preds, ensemble, test_preds, y_test)
+
+        return output_text, fig
 
-"""
-        
-        for model_name, pred in next_day_preds.items():
-            pred_value = float(pred)
-            change = ((pred_value - current_price) / current_price) * 100
-            arrow = "📈" if change > 0 else "📉"
-            result += f"{arrow} {model_name}: {pred_value:.5f} ({change:+.2f}%)\n"
-        
-        ensemble_pred_value = float(ensemble_pred)
-        ensemble_change = ((ensemble_pred_value - current_price) / current_price) * 100
-        ensemble_arrow = "📈" if ensemble_change > 0 else "📉"
-        
-        result += f"\n{ensemble_arrow} 🏆 ENSEMBLE ORTALAMA: {ensemble_pred_value:.5f} ({ensemble_change:+.2f}%)\n"
-        
-        # Model performansları
-        result += f"\n\n📈 MODEL PERFORMANSLARI (Son 5 Gün):\n"
-        for i, (model_name, preds) in enumerate(test_predictions.items()):
-            mse = float(np.mean((preds - y_test) ** 2))
-            mae = float(np.mean(np.abs(preds - y_test)))
-            result += f"• {model_name}: MAE={mae:.5f}, MSE={mse:.8f}\n"
-        
-        # Grafik için veri hazırla
-        import matplotlib.pyplot as plt
-        
-        fig, (ax1, ax2) = plt.subplots(2, 1, figsize=(12, 10))
-        
-        # Üst grafik: Fiyat geçmişi
-        last_30_days = df.tail(30)
-        ax1.plot(range(len(last_30_days)), last_30_days['Close'].values, label='Gerçek Fiyat', linewidth=2)
-        ax1.axhline(y=current_price, color='gray', linestyle='--', alpha=0.5, label='Mevcut Fiyat')
-        ax1.axhline(y=ensemble_pred_value, color='red', linestyle='--', alpha=0.7, label='Tahmin')
-        ax1.set_title('EUR/USD Son 30 Gün Fiyat Hareketi', fontsize=14, fontweight='bold')
-        ax1.set_xlabel('Gün')
-        ax1.set_ylabel('Fiyat')
-        ax1.legend()
-        ax1.grid(True, alpha=0.3)
-        
-        # Alt grafik: Model tahminleri karşılaştırması
-        model_names = list(next_day_preds.keys())
-        model_values = [float(v) for v in next_day_preds.values()]
-        colors = ['#FF6B6B', '#4ECDC4', '#45B7D1', '#FFA07A', '#98D8C8']
-        
-        bars = ax2.barh(model_names, model_values, color=colors)
-        ax2.axvline(x=current_price, color='black', linestyle='--', linewidth=2, label='Mevcut Fiyat')
-        ax2.set_title('Model Tahminleri Karşılaştırması', fontsize=14, fontweight='bold')
-        ax2.set_xlabel('Tahmin Edilen Fiyat')
-        ax2.legend()
-        ax2.grid(True, alpha=0.3, axis='x')
-        
-        # Değerleri barların üzerine yaz
-        for i, (bar, val) in enumerate(zip(bars, model_values)):
-            change = ((val - current_price) / current_price) * 100
-            ax2.text(val, i, f' {val:.5f} ({change:+.2f}%)', 
-                    va='center', fontsize=9, fontweight='bold')
-        
-        plt.tight_layout()
-        
-        return result, fig
-        
     except Exception as e:
-        import traceback
-        error_msg = f"❌ Hata oluştu: {str(e)}\n\n{traceback.format_exc()}"
-        return error_msg, None
-
-# Gradio arayüzü
-with gr.Blocks(title="EUR/USD Fiyat Tahmini", theme=gr.themes.Soft()) as app:
-    gr.Markdown("""
-    # 💹 EUR/USD Fiyat Tahmin Sistemi
-    ## 5 Ensemble Model ile Otomatik Forex Tahmini
-    
-    Bu uygulama **Random Forest, Gradient Boosting, AdaBoost, Ridge Regression ve SVR** modellerini kullanarak EUR/USD paritesi için ertesi gün fiyat tahmini yapar.
-    
-    🔄 **Yenile** butonuna tıklayarak güncel verileri çekip yeni tahmin alabilirsiniz.
-    """)
-    
-    with gr.Row():
-        predict_btn = gr.Button("🚀 Tahmin Yap / Yenile", variant="primary", scale=2)
-    
+        return f"❌ Hata:\n{traceback.format_exc()}", None
+
+# ─── CSS ─────────────────────────────────────────────────────────────────────
+CSS = """
+body, .gradio-container {
+    background: #0A0E17 !important;
+    font-family: 'JetBrains Mono', 'Fira Code', monospace !important;
+    color: #F1F5F9 !important;
+}
+.gr-button-primary {
+    background: linear-gradient(135deg, #1D4ED8, #3B82F6) !important;
+    border: none !important;
+    color: #fff !important;
+    font-weight: 700 !important;
+    letter-spacing: 0.05em !important;
+    padding: 12px 28px !important;
+    border-radius: 6px !important;
+    transition: opacity .2s !important;
+}
+.gr-button-primary:hover { opacity: 0.85 !important; }
+.gr-textbox textarea {
+    background: #111827 !important;
+    color: #94A3B8 !important;
+    border: 1px solid #1E293B !important;
+    font-family: 'JetBrains Mono', monospace !important;
+    font-size: 12.5px !important;
+    line-height: 1.7 !important;
+}
+.gr-box, .gr-panel {
+    background: #111827 !important;
+    border: 1px solid #1E293B !important;
+}
+footer { display: none !important; }
+"""
+
+HEADER_HTML = """
+<div style="
+    background: linear-gradient(135deg, #0A0E17 0%, #111827 100%);
+    border: 1px solid #1E293B;
+    border-radius: 10px;
+    padding: 28px 32px 20px;
+    margin-bottom: 8px;
+">
+  <div style="display:flex; align-items:center; gap:14px; margin-bottom:10px;">
+    <span style="font-size:2rem;">💹</span>
+    <div>
+      <h1 style="margin:0; font-size:1.5rem; font-weight:800; color:#F1F5F9; letter-spacing:.04em;">
+        EUR/USD  ·  ENSEMBLE FORECAST ENGINE
+      </h1>
+      <p style="margin:4px 0 0; color:#64748B; font-size:.82rem; font-family:monospace;">
+        Random Forest  ·  Gradient Boosting  ·  AdaBoost  ·  Ridge  ·  SVR
+      </p>
+    </div>
+  </div>
+  <div style="display:flex; gap:20px; flex-wrap:wrap; margin-top:14px;">
+    <span style="background:#1E293B; border-radius:4px; padding:4px 12px; font-size:.78rem; color:#94A3B8;">📡 Yahoo Finance — 90g</span>
+    <span style="background:#1E293B; border-radius:4px; padding:4px 12px; font-size:.78rem; color:#94A3B8;">📐 22 Teknik Gösterge</span>
+    <span style="background:#1E293B; border-radius:4px; padding:4px 12px; font-size:.78rem; color:#94A3B8;">🧪 10-Gün Backtest</span>
+    <span style="background:#1E293B; border-radius:4px; padding:4px 12px; font-size:.78rem; color:#94A3B8;">⚡ Tek Tıkla Güncelle</span>
+  </div>
+</div>
+"""
+
+# ─── Gradio arayüzü ──────────────────────────────────────────────────────────
+with gr.Blocks(css=CSS, title="EUR/USD Ensemble Forecast") as app:
+    gr.HTML(HEADER_HTML)
+
     with gr.Row():
-        output_text = gr.Textbox(label="📊 Tahmin Sonuçları", lines=25)
-    
+        predict_btn = gr.Button("⟳  TAHMİN YAP / YENİLE", variant="primary", scale=1)
+
     with gr.Row():
-        output_plot = gr.Plot(label="📈 Grafikler")
-    
+        with gr.Column(scale=1):
+            output_text = gr.Textbox(
+                label="RAPOR",
+                lines=30,
+                interactive=False,
+                show_copy_button=True,
+            )
+        with gr.Column(scale=2):
+            output_plot = gr.Plot(label="GRAFİKLER")
+
     predict_btn.click(
         fn=predict_eurusd,
         inputs=[],
-        outputs=[output_text, output_plot]
+        outputs=[output_text, output_plot],
     )
-    
-    gr.Markdown("""
-    ---
-    ### 📝 Notlar:
-    - Veriler Yahoo Finance'ten otomatik olarak çekilir
-    - Tahminler son 60 günlük veriye dayanır
-    - 5 farklı makine öğrenmesi modeli kullanılır
-    - Ensemble tahmini, tüm modellerin ortalamasıdır
-    - **Finansal tavsiye değildir, sadece eğitim amaçlıdır**
-    """)
 
 if __name__ == "__main__":
     app.launch()