| import numpy as np | |
| from sklearn.preprocessing import StandardScaler | |
| # Simula o mesmo shape e pré-processamento do modelo | |
| scaler = StandardScaler() | |
| def preprocess_input_data(data): | |
| data = np.array(data).astype(np.float32) | |
| if data.ndim == 1: | |
| data = np.expand_dims(data, axis=0) | |
| data = scaler.fit_transform(data) | |
| return np.expand_dims(data, axis=2) # [batch, features, 1] | |