v2.0: 33 AI coins - KITE, TAO, VIRTUAL, FET, LINK y 25 mas

download_data.py

@@ -51,7 +51,17 @@ def _get_working_endpoint() -> str:
     logger.warning("\u26a0\ufe0f Ning\u00fan endpoint directo disponible, usando data-api")
     return "https://data-api.binance.vision"
 
-DEFAULT_SYMBOLS = ["BTCUSDT", "ETHUSDT", "SOLUSDT"]
+DEFAULT_SYMBOLS = [
+    # Base (mercado general)
+    "BTCUSDT", "ETHUSDT", "SOLUSDT",
+    # AI Coins (baja correlacion con BTC)
+    "LINKUSDT", "TAOUSDT", "WLDUSDT", "VIRTUALUSDT", "FETUSDT",
+    "INJUSDT", "GRTUSDT", "KITEUSDT", "THETAUSDT",
+    "KAITOUSDT", "SENTUSDT", "LPTUSDT", "AWEUSDT", "TURBOUSDT",
+    "SAHARAUSDT", "VANAUSDT", "NMRUSDT", "OPENUSDT", "ROBOUSDT",
+    "HOLOUSDT", "RLCUSDT", "IOUSDT", "PHAUSDT", "IQUSDT",
+    "AIXBTUSDT", "SAPIENUSDT", "FLUXUSDT", "ALLOUSDT", "MIRAUSDT",
+]
 DEFAULT_TIMEFRAME = "4h"
 DEFAULT_DAYS = 1825  # ~5 años
 MAX_CANDLES_PER_REQUEST = 1000
@@ -420,11 +430,14 @@ def main():
 
     # ── 1. Klines spot ──
     for symbol in args.symbols:
+        path = os.path.join(DATA_DIR, f"klines_{symbol}_{args.timeframe}.parquet")
+        if os.path.exists(path):
+            logger.info("Klines %s ya existe, salteando", symbol)
+            continue
         df = download_klines(symbol, args.timeframe, args.days)
         if not df.empty:
-            path = os.path.join(DATA_DIR, f"klines_{symbol}_{args.timeframe}.parquet")
             df.to_parquet(path)
-            logger.info("💾 Guardado: %s (%d filas)", path, len(df))
+            logger.info("Guardado: %s (%d filas)", path, len(df))
 
     # ── 2. Datos de derivados (futuros) ──
     if not args.no_derivatives:

feature_engine.py

@@ -365,7 +365,15 @@ def main():
     parser.add_argument("--all", action="store_true")
     args = parser.parse_args()
 
-    symbols = ["BTCUSDT", "ETHUSDT", "SOLUSDT"] if args.all else [args.symbol]
+    symbols = [
+        "BTCUSDT", "ETHUSDT", "SOLUSDT",
+        "LINKUSDT", "TAOUSDT", "WLDUSDT", "VIRTUALUSDT", "FETUSDT",
+        "INJUSDT", "GRTUSDT", "KITEUSDT", "THETAUSDT",
+        "KAITOUSDT", "SENTUSDT", "LPTUSDT", "AWEUSDT", "TURBOUSDT",
+        "SAHARAUSDT", "VANAUSDT", "NMRUSDT", "OPENUSDT", "ROBOUSDT",
+        "HOLOUSDT", "RLCUSDT", "IOUSDT", "PHAUSDT", "IQUSDT",
+        "AIXBTUSDT", "SAPIENUSDT", "FLUXUSDT", "ALLOUSDT", "MIRAUSDT",
+    ] if args.all else [args.symbol]
 
     for symbol in symbols:
         df = generate_features(symbol, args.timeframe)

model_signals.py

@@ -302,17 +302,20 @@ def main():
         return
 
     if args.multi:
+        import glob
+        pattern = os.path.join(DATA_DIR, f"labeled_*_{args.timeframe}.parquet")
+        files = sorted(glob.glob(pattern))
         dfs = []
-        for sym in ["BTCUSDT", "ETHUSDT", "SOLUSDT"]:
-            path = os.path.join(DATA_DIR, f"labeled_{sym}_{args.timeframe}.parquet")
-            if os.path.exists(path):
-                df_sym = pd.read_parquet(path)
-                logger.info("Loaded %s: %d rows", sym, len(df_sym))
-                dfs.append(df_sym)
+        for path in files:
+            df_sym = pd.read_parquet(path)
+            sym = os.path.basename(path).replace(f"labeled_", "").replace(f"_{args.timeframe}.parquet", "")
+            logger.info("Loaded %s: %d rows", sym, len(df_sym))
+            dfs.append(df_sym)
         if not dfs:
+            logger.error("No labeled files found")
             return
         df = pd.concat(dfs, axis=0).sort_index()
-        logger.info("Combined: %d rows", len(df))
+        logger.info("Combined: %d rows (%d symbols)", len(df), len(dfs))
     else:
         path = os.path.join(DATA_DIR, f"labeled_{args.symbol}_{args.timeframe}.parquet")
         if not os.path.exists(path):

regime_detector.py

@@ -275,19 +275,20 @@ def main():
 
     # Cargar datos - multi-symbol si habilitado
     if args.multi:
-        symbols = ["BTCUSDT", "ETHUSDT", "SOLUSDT"]
+        import glob
+        pattern = os.path.join(DATA_DIR, f"labeled_*_{args.timeframe}.parquet")
+        files = sorted(glob.glob(pattern))
         dfs = []
-        for sym in symbols:
-            path = os.path.join(DATA_DIR, f"labeled_{sym}_{args.timeframe}.parquet")
-            if os.path.exists(path):
-                df_sym = pd.read_parquet(path)
-                logger.info("Cargado %s: %d velas", sym, len(df_sym))
-                dfs.append(df_sym)
+        for path in files:
+            df_sym = pd.read_parquet(path)
+            sym = os.path.basename(path).replace("labeled_", "").replace(f"_{args.timeframe}.parquet", "")
+            logger.info("Cargado %s: %d velas", sym, len(df_sym))
+            dfs.append(df_sym)
         if not dfs:
             logger.error("No se encontraron datos labeled")
             return
         df = pd.concat(dfs, axis=0).sort_index()
-        logger.info("Dataset combinado: %d velas (3 pares)", len(df))
+        logger.info("Dataset combinado: %d velas (%d pares)", len(df), len(dfs))
     else:
         labeled_path = os.path.join(DATA_DIR, f"labeled_{args.symbol}_{args.timeframe}.parquet")
         if not os.path.exists(labeled_path):

regime_labeler.py

@@ -164,7 +164,15 @@ def main():
     parser.add_argument("--all", action="store_true")
     args = parser.parse_args()
 
-    symbols = ["BTCUSDT", "ETHUSDT", "SOLUSDT"] if args.all else [args.symbol]
+    symbols = [
+        "BTCUSDT", "ETHUSDT", "SOLUSDT",
+        "LINKUSDT", "TAOUSDT", "WLDUSDT", "VIRTUALUSDT", "FETUSDT",
+        "INJUSDT", "GRTUSDT", "KITEUSDT", "THETAUSDT",
+        "KAITOUSDT", "SENTUSDT", "LPTUSDT", "AWEUSDT", "TURBOUSDT",
+        "SAHARAUSDT", "VANAUSDT", "NMRUSDT", "OPENUSDT", "ROBOUSDT",
+        "HOLOUSDT", "RLCUSDT", "IOUSDT", "PHAUSDT", "IQUSDT",
+        "AIXBTUSDT", "SAPIENUSDT", "FLUXUSDT", "ALLOUSDT", "MIRAUSDT",
+    ] if args.all else [args.symbol]
 
     for symbol in symbols:
         features_path = os.path.join(DATA_DIR, f"features_{symbol}_{args.timeframe}.parquet")

startup.sh

@@ -1,23 +1,19 @@
 #!/bin/bash
 echo "========================================="
-echo "  AURORA BRAIN v1.3 — Starting Pipeline"
+echo "  AURORA BRAIN v2.0 — 33 AI Coins Pipeline"
 echo "========================================="
 
-# Solo limpiar features y modelos (NO los parquets de datos crudos)
+# Limpiar features, labels y modelos (conservar klines y funding)
 rm -f data/features_*.parquet data/labeled_*.parquet
 rm -f models/*.pkl models/*.json
 echo "Features y modelos limpiados"
 
-# Paso 1: Descargar datos (solo si no existen las klines)
-if [ ! -f "data/klines_BTCUSDT_4h.parquet" ]; then
-    echo "Descargando datos historicos..."
-    python download_data.py --symbols BTCUSDT ETHUSDT SOLUSDT --days 1825
-else
-    echo "Klines ya existen"
-fi
+# Paso 1: Descargar datos (descarga SOLO las que faltan, las existentes las saltea)
+echo "Descargando datos historicos (33 pares)..."
+python download_data.py --days 1825
 
-# Paso 2: Generar features
-echo "Generando features..."
+# Paso 2: Generar features para todos
+echo "Generando features (33 pares)..."
 python feature_engine.py --all
 
 # Paso 3: Etiquetar regimenes
@@ -28,7 +24,7 @@ python regime_labeler.py --all
 echo "Entrenando detector de regimen..."
 python regime_detector.py --multi
 
-# Paso 5: Entrenar modelo de senales (multi-symbol, threshold 2%)
+# Paso 5: Entrenar modelo de senales (multi-symbol)
 echo "Entrenando modelo de senales..."
 python model_signals.py --multi --threshold 2.0 --horizon 12