Applying Seasonality w/ Fourier Terms

[benfetch]

Hi there,
I am currently using statsforecast with statsmodels Calendar Fourier class to include fourier terms as exogenous regressors. Am I able to create a new class as a child class of AutoARIMA where these terms might be included natively. I have tried by implementing the following class but it fails as y is a numpy array and doesn't have an index.

class AutoARIMADHR(AutoARIMA):

def __init__(self, freq='D', monthly_k=0, annual_k=0, *args, **kwargs):
    super().__init__(*args, **kwargs)
    self.freq = freq
    self.monthly_k = monthly_k
    self.annual_k = annual_k
    self.fourier_terms = None

def fit(self, y, X=None):

    # Generate Fourier terms
    monthly_fourier = CalendarFourier(freq='MS', order=self.monthly_k)
    annual_fourier = CalendarFourier(freq='YS', order=self.annual_k)
      # Adjust frequency and order as needed
    dp = DeterministicProcess(index=y.index, constant=True, order=1, seasonal=False, additional_terms=[monthly_fourier, annual_fourier])
    self.fourier_terms = dp.in_sample()

    # Combine original features with Fourier terms
    if X is not None:
        X = pd.concat([X, self.fourier_terms], axis=1)
    else:
        X = self.fourier_terms
    
    return super().fit(y, X)

def predict(self, h, X=None):
    # Generate future Fourier terms
    future_fourier_terms = self.fourier_terms.iloc[-len(h):]
    
    # Combine original features with future Fourier terms
    if X is not None:
        X = pd.concat([X, future_fourier_terms], axis=1)
    else:
        X = future_fourier_terms
    
    return super().predict(h, X)

Is this possible? I want to be able to just call the StatsForecast.forecast method and compare multiples of these models simultaneously.