Update app.py

app.py

@@ -3,12 +3,10 @@ import os
 import logging
 import time
 
-# 🔧 Force log output to HF container
 logging.basicConfig(stream=sys.stdout, level=logging.INFO, format="%(asctime)s [%(levelname)s] %(message)s")
 logger = logging.getLogger("crypto_ml")
 logger.info("🚀 Starting HF Space container...")
 
-# 1. IMPORTS (Wrapped for safe logging)
 try:
     import gradio as gr
     import yfinance as yf
@@ -17,48 +15,71 @@ try:
     from sklearn.ensemble import RandomForestClassifier
     from sklearn.preprocessing import MinMaxScaler
     import joblib
-    logger.info("✅ All dependencies imported successfully")
+    logger.info("✅ Dependencies loaded")
 except Exception as e:
     logger.error(f"❌ Import failed: {e}")
     sys.exit(1)
 
-# 2. CONFIG
 SYMBOL = "ETH-USD"
 FEE = 0.0015
 WINDOW = 60
 MODEL_PATH = "/tmp/rf_model.pkl"
 SCALER_PATH = "/tmp/scaler.pkl"
 
-# 3. DATA & MODEL FUNCTIONS
 def fetch_and_prepare():
     logger.info("📥 Fetching ETH-USD from Yahoo Finance...")
     df = yf.download(SYMBOL, period="60d", interval="1h", progress=False)
-    df = df.resample("4h").agg({"Open":"first","High":"max","Low":"min","Close":"last","Volume":"sum"}).dropna()
-    df.columns = ["open","high","low","close","volume"]
+    if df.empty:
+        raise ValueError("yfinance returned empty data")
+
+    # 🔧 FIX: Handle yfinance MultiIndex columns (v0.2.30+)
+    if isinstance(df.columns, pd.MultiIndex):
+        df.columns = df.columns.get_level_values(1)
+    
+    # Map Adj Close -> Close if present
+    if 'Adj Close' in df.columns:
+        df.rename(columns={'Adj Close': 'Close'}, inplace=True)
+        
+    # Standardize to lowercase to avoid case-sensitivity issues
+    df.columns = [col.lower() for col in df.columns]
+    
+    # Verify required columns
+    required = ['open', 'high', 'low', 'close', 'volume']
+    missing = [c for c in required if c not in df.columns]
+    if missing:
+        raise ValueError(f"Missing columns: {missing}. Found: {list(df.columns)}")
+
+    # Resample to 4h candles
+    df = df.resample("4h").agg({
+        "open": "first", "high": "max", "low": "min",
+        "close": "last", "volume": "sum"
+    }).dropna()
+
     df.index.name = "timestamp"
     df = df.reset_index()
-    
-    # Indicators (Manual)
+
+    # 📊 Technical Indicators
     delta = df["close"].diff()
-    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))
-    
+    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))
+
     sma = df["close"].rolling(20).mean()
     std = df["close"].rolling(20).std()
-    df["bb_upper"] = sma + (std*2)
-    df["bb_lower"] = sma - (std*2)
-    
+    df["bb_upper"] = sma + (std * 2)
+    df["bb_lower"] = sma - (std * 2)
+
+    # Labels
     df["next_close"] = df["close"].shift(-1)
-    df["label"] = (df["next_close"] > df["close"]*(1+FEE)).astype(int)
+    df["label"] = (df["next_close"] > df["close"] * (1 + FEE)).astype(int)
     df = df.dropna()
-    
-    feat_cols = ["open","high","low","close","volume","rsi","bb_upper","bb_lower"]
+
+    feat_cols = ["open", "high", "low", "close", "volume", "rsi", "bb_upper", "bb_lower"]
     X_list, y_list = [], []
     for i in range(WINDOW, len(df)):
         X_list.append(df.iloc[i-WINDOW:i][feat_cols].values.flatten())
         y_list.append(df.iloc[i]["label"])
-        
+
     return np.array(X_list), np.array(y_list), df, feat_cols
 
 def run_train():
@@ -68,7 +89,7 @@ def run_train():
         scaler = MinMaxScaler()
         X_scaled = scaler.fit_transform(X)
         
-        split = int(len(X)*0.95)
+        split = int(len(X) * 0.95)
         model = RandomForestClassifier(n_estimators=100, max_depth=10, random_state=42, n_jobs=-1)
         model.fit(X_scaled[:split], y[:split])
         
@@ -77,10 +98,10 @@ def run_train():
         joblib.dump(scaler, SCALER_PATH)
         
         logger.info(f"✅ Training complete. Accuracy: {acc:.4f}")
-        return {"status": "✅ Success", "accuracy": round(acc,4), "message": "Model saved to /tmp/"}
+        return {"status": "✅ Success", "accuracy": round(acc, 4), "message": "Model saved."}
     except Exception as e:
         logger.error(f"❌ Train failed: {e}")
-        return {"status": "❌ Error", "message": str(e)}
+        return {"status": "❌ Error", "message": str(e)[:150]}
 
 def run_predict():
     try:
@@ -89,20 +110,24 @@ def run_predict():
             
         X, y, df_full, feat_cols = fetch_and_prepare()
         scaler = joblib.load(SCALER_PATH)
-        latest = df_full.iloc[-WINDOW:][feat_cols].values.flatten().reshape(1,-1)
+        latest = df_full.iloc[-WINDOW:][feat_cols].values.flatten().reshape(1, -1)
         latest_scaled = scaler.transform(latest)
         
         model = joblib.load(MODEL_PATH)
         pred = int(model.predict(latest_scaled)[0])
         prob = float(model.predict_proba(latest_scaled)[0][1])
         
-        return {"status": "✅ Prediction", "signal": "BUY" if pred==1 else "HOLD", "confidence": f"{prob*100:.1f}%"}
+        return {
+            "status": "✅ Prediction",
+            "signal": "BUY (Hold 4h)" if pred == 1 else "HOLD/SELL",
+            "confidence_pct": f"{prob*100:.1f}%",
+            "logic": "Predicts >0.15% move next 4h"
+        }
     except Exception as e:
         logger.error(f"❌ Predict failed: {e}")
-        return {"status": "❌ Error", "message": str(e)}
+        return {"status": "❌ Error", "message": str(e)[:150]}
 
-# 4. GRADIO UI
-logger.info("🎨 Building Gradio UI...")
+# 🎨 UI
 tradingview_html = """
 <div style="width:100%;height:400px;margin:10px 0;">
   <iframe src="https://s.tradingview.com/widgetembed/?symbol=BINANCE%3AETHUSDT&interval=240&hidesidetoolbar=1&symboledit=1&saveimage=1&toolbarbg=transparent&studies=%5B%5D&theme=dark&style=1&timezone=UTC&withdateranges=1&showpopupbutton=1&studies_overrides=%7B%7D&overrides=%7B%7D&enabled_features=%5B%5D&disabled_features=%5B%5D&showpopupbutton=1&locale=en" 
@@ -124,12 +149,7 @@ with gr.Blocks(title="Crypto ML Strategy") as demo:
         
     train_btn.click(run_train, outputs=train_out)
     pred_btn.click(run_predict, outputs=pred_out)
-    gr.Markdown("💡 *Model disimpan di `/tmp/`. Retrain otomatis jika Space restart.*")
+    gr.Markdown("💡 *Model saved in `/tmp/`. Retrain if Space restarts.*")
 
-# 5. QUEUE & LAUNCH
 demo.queue(default_concurrency_limit=1)
-logger.info("✅ Gradio app built & queued. Ready for HF Spaces runner.")
-
-# Fallback untuk local testing, diabaikan di HF
-if __name__ == "__main__":
-    demo.launch(server_name="0.0.0.0", server_port=7860)
+logger.info("✅ App ready. HF Spaces runner will take over.")