IdentityPredictor failing when history is not long enough.

[melopeo]

I am using several models via GluonTS shell. When using IdentityPredictor I get the following error:
ValueError: operands could not be broadcast together with remapped shapes [original->remapped]: (1,107) and requested shape (100,672)

This error seems to be located here:

def predict_item(self, item: DataEntry) -> Forecast: 
    prediction = item["target"][-self.prediction_length :]
    samples = np.broadcast_to(
        array=np.expand_dims(prediction, 0),
        shape=(self.num_samples, self.prediction_length),
    ) 

I guess this means that when there are not enough observations in the past, then the IdentityPredictor just doesn't work.
Would it make sense to extend it to handle these cases and just generate a warning? Maybe copying its values until the desired prediction length is satisfied?