[FEATURE] Add a step to prepare datasets for training like we had with `prepare_dataset`

[plaguss]

Is your feature request related to a problem? Please describe. In distilabel<1.0.0 we have prepare_datasets, we need a step offering the same behaviour.

Describe the solution you'd like A new PrepareDatasetForTraining or similar that offers the same functionality.

Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

Additional context Add any other context or screenshots about the feature request here.