Ajout analyse de données par réseau neuronal

src/app.py

@@ -8,6 +8,13 @@ import joblib
 from io import TextIOWrapper
 from ydata_profiling import ProfileReport
 
+# model
+from sklearn.model_selection import train_test_split
+from sklearn.preprocessing import StandardScaler
+from tensorflow.keras.models import Sequential
+from tensorflow.keras.layers import Dense
+from tensorflow.keras.utils import plot_model
+
 app = Flask(__name__)
 FREQTRADE_USERDATA_DIR = '/mnt/external'
 
@@ -121,6 +128,43 @@ def read_feather(filename):
         # dataframe['min200'] = talib.MIN(dataframe['close'], timeperiod=200)
         # dataframe['max200'] = talib.MAX(dataframe['close'], timeperiod=200)
 
+        # Choisir les colonnes techniques comme variables d'entrée (X)
+        feature_cols = ['rsi', 'sma20', 'sma5_1h', 'volume']
+        df = dataframe
+        # Variable cible
+        df['target'] = df['futur_price_1h']
+
+        # Supprimer les lignes avec des NaN
+        df.dropna(subset=feature_cols + ['target'], inplace=True)
+
+        X = df[feature_cols].values
+        y = df['target'].values
+
+        # Normalisation
+        scaler = StandardScaler()
+        X = scaler.fit_transform(X)
+
+        # Split
+        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
+
+        # Modèle
+        model = Sequential([
+            Dense(64, input_dim=X.shape[1], activation='relu'),
+            Dense(32, activation='relu'),
+            Dense(1)  # Prédiction continue
+        ])
+
+        model.compile(optimizer='adam', loss='mse', metrics=['mae'])
+
+        # Entraînement
+        model.fit(X_train, y_train, epochs=50, batch_size=32, validation_data=(X_test, y_test))
+
+        loss, mae = model.evaluate(X_test, y_test)
+        print(f"Erreur moyenne absolue : {mae:.4f}")
+
+        model.summary()
+
+        plot_model(model, show_shapes=True, show_layer_names=True, to_file=FREQTRADE_USERDATA_DIR + "/reports/model.png")
         return dataframe.to_json(orient="records")
     except Exception as e:
         print(e)
@@ -185,6 +229,24 @@ def get_chart_data():
 
     return df.to_json(orient="records") #jsonify(chart_data)
 
+@app.route('/model')
+def show_model():
+    # Créer un exemple de modèle si non encore généré
+    model_path = FREQTRADE_USERDATA_DIR + "/reports/model.png"
+    if not os.path.exists(model_path):
+        model = Sequential([
+            Dense(64, input_shape=(6,), activation='relu'),
+            Dense(32, activation='relu'),
+            Dense(1)
+        ])
+        plot_model(model, to_file=model_path, show_shapes=True, show_layer_names=True)
+    return render_template('model.html', model_image=model_path)
+
+# Route pour servir les fichiers statiques (optionnelle si bien configuré)
+@app.route('/static/<path:filename>')
+def static_files(filename):
+    return send_from_directory('static', filename)
+
 
 if __name__ == '__main__':
     app.run(debug=True, host='0.0.0.0', port=5000)