import neuralforecast as nf

import neuralforecast.losses.pytorch as nflp

import pandas as pd

import numpy as np

from stdlib.qname import import_from, create_from

NF_LOSSES = {

None: nflp.MSE,

"mae": nflp.MAE,

"mse": nflp.MSE,

"rmse": nflp.RMSE,

"mape": nflp.MAPE,

"smape": nflp.SMAPE,

"mase": nflp.MASE,

"relmse": nflp.relMSE,

"quatileloss": nflp.QuantileLoss,

"mqloss": nflp.MQLoss,

"huberloss": nflp.HuberLoss,

"huberqloss": nflp.HuberQLoss,

"hubermqloss": nflp.HuberMQLoss,

"distributionloss": nflp.DistributionLoss

}

print(create_from("mse", aliases=NF_LOSSES))

print(create_from(dict(

clazz="distributionloss",

distribution="StudentT", level=[80, 90], return_params=False

), aliases=NF_LOSSES))