Provide a better implementation of `Dataset.__init__`

[tizianocitro]

The implementation could change but the issue is the same: __init__ does too many things.

I want to refactor this, also to make dependency clearer. For example, process() requires self.hypergraph not None and self.__is_original == True, which is not very clear from the current implementation.

def __init__(
   self,
   hdata: Optional[HData] = None,
   sampling_strategy: SamplingStrategy = SamplingStrategy.HYPEREDGE,
    is_original: Optional[bool] = True,
) -> None:
    self.__is_original = is_original
    self.__sampler = create_sampler_from_strategy(sampling_strategy)

    self.sampling_strategy = sampling_strategy
    self.hypergraph = self.download() if hdata is None else HIFHypergraph.empty()
    self.hdata = self.process() if hdata is None else hdata

I want to refactor this to provide a better way to create datasets.