Hi, I am reading about Feature Engineering part of utilsforecast and I think it is a great and useful module. It will be even better if we can use it in MLForecast/NeuralForecast directly when creating features. Just making sure in case I missed the information, can't utilsforecast method (

pipeline

,

fourier

,

trend

) be used in MLForecast like how

lags

,

date_features

,

lag_transforms

, and

target_transforms

are defined?

You can use it to create features like mentioned here (simply adding stuff to your dataset): Exogenous features - Nixtla

Okay, so it must be created manually outside the MLForecast instance. Thank you!