Añadir modelo BBVA y endpoint de predicción

app.py

@@ -1,7 +1,129 @@
+# ============================================================
+# 📈 app.py — Backend FastAPI para predicciones BBVA
+# ============================================================
+
 from fastapi import FastAPI
+from pydantic import BaseModel
+from fastapi.middleware.cors import CORSMiddleware
+import torch
+from torch import nn
+import joblib
+import numpy as np
+import pandas as pd
+
+# ============================================================
+# 1️⃣ Configuración base de FastAPI
+# ============================================================
+app = FastAPI(title="Predicción BBVA API")
+
+# Permitir acceso desde cualquier dominio (para GitHub Pages luego)
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+# ============================================================
+# 2️⃣ Definición del modelo LSTM (idéntico al entrenamiento)
+# ============================================================
+class LSTMModel(nn.Module):
+    def __init__(self, n_features, hidden_size=64):
+        super().__init__()
+        self.lstm = nn.LSTM(input_size=n_features, hidden_size=hidden_size, batch_first=True)
+        self.fc = nn.Linear(hidden_size, 1)
+    def forward(self, x):
+        out, _ = self.lstm(x)
+        out = out[:, -1, :]
+        out = self.fc(out)
+        return out.squeeze()
+
+# ============================================================
+# 3️⃣ Cargar el modelo y los scalers
+# ============================================================
+MODEL_PATH = "models_bbva/bbva_lstm_model.pth"
+SCALER_FULL_PATH  = "models_bbva/scaler_features_full.pkl"
+SCALER_TARGET_PATH = "models_bbva/scaler_target.pkl"
 
-app = FastAPI()
+n_features = 5
+hidden_size = 64
+device = torch.device("cpu")
 
+# Cargar modelo y scalers
+model = LSTMModel(n_features, hidden_size)
+model.load_state_dict(torch.load(MODEL_PATH, map_location=device))
+model.to(device)
+model.eval()
+
+scaler_features = joblib.load(SCALER_FULL_PATH)
+scaler_target = joblib.load(SCALER_TARGET_PATH)
+
+# ============================================================
+# 4️⃣ Modelo de entrada (JSON)
+# ============================================================
+class PredictRequest(BaseModel):
+    n_future: int = 3
+
+# ============================================================
+# 5️⃣ Endpoint raíz de prueba
+# ============================================================
 @app.get("/")
 def home():
-    return {"message": "✅ Hola Martina, el backend está funcionando correctamente!"}
+    return {"message": "✅ Backend BBVA activo y listo para predecir"}
+
+# ============================================================
+# 6️⃣ Endpoint de predicción
+# ============================================================
+@app.post("/predict/bbva")
+def predict_bbva(req: PredictRequest):
+    # === Cargar dataset base ===
+    PATH_BBVA = "data/bbva_pred.csv"  # opcional, puedes cambiarlo
+    try:
+        bbva = pd.read_csv(PATH_BBVA)
+    except Exception:
+        return {"error": "No se encontró el archivo de datos base (data/bbva_pred.csv)"}
+
+    bbva["Date"] = pd.to_datetime(bbva["Date"])
+    bbva = bbva.set_index("Date")
+    start_date = "2021-01-01"
+    end_date = "2025-10-31"
+    bbva = bbva.loc[(bbva.index >= start_date) & (bbva.index <= end_date)].copy()
+
+    # === Columnas del modelo ===
+    cols_to_scale = [
+        'Open', 'High', 'Low', 'Close', 'Adj Close',
+        'ma_5', 'close_lag1', 'close_lag2', 'close_lag3',
+        'Volume', 'ibex_momentum_5d'
+    ]
+    features = ['Open', 'High', 'Low', 'return_1d', 'return_3d']
+    window_size = 3
+    n_future = req.n_future
+
+    # === Escalar ===
+    bbva_scaled = bbva.copy()
+    bbva_scaled[cols_to_scale] = scaler_features.transform(bbva[cols_to_scale])
+
+    # === Última ventana ===
+    last_window = bbva_scaled[features].iloc[-window_size:].values
+
+    # === Predicción iterativa ===
+    preds_scaled = []
+    window = last_window.copy()
+    for _ in range(n_future):
+        X_input = torch.tensor(window, dtype=torch.float32).unsqueeze(0).to(device)
+        with torch.no_grad():
+            pred_scaled = model(X_input).cpu().numpy().flatten()[0]
+        preds_scaled.append(pred_scaled)
+        next_day = window[-1].copy()
+        window = np.vstack([window[1:], next_day])
+
+    # === Desescalar ===
+    preds_real = scaler_target.inverse_transform(np.array(preds_scaled).reshape(-1, 1)).flatten().tolist()
+
+    # === Fechas futuras ===
+    last_date = bbva.index[-1]
+    future_dates = pd.date_range(last_date + pd.Timedelta(days=1), periods=n_future, freq="B")
+    fechas = [str(d.date()) for d in future_dates]
+
+    return {"fechas": fechas, "predicciones": preds_real}

models_bbva/bbva_lstm_model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c83ee225e37688bc90d9c19d3a43d9fcf6dc6da245fe43c56f4cb9960792075c
+size 75497

models_bbva/scaler_features_full.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:362d0ed626b992c46b7abfb343876c908065350f9645d8b6b3c07176b23f5f5e
+size 1487

models_bbva/scaler_features_train.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3738f565788faf7031bb48aae5e9569bbb1c13b9c04ced86265b08d20395e73
+size 1487

models_bbva/scaler_target.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:88a0fa16db10ebc2c1959eb96de8397e2c091febad0006f83aa10abcaaf81b2c
+size 975