Creation of the prepare_dataset_for_training function. Adding docstrings

The function is a bit weird but it seems to be what teachers asked us to do. Indeed, according to the description of the projetc, the function is supposed to : Split between training set and test set or Split the training set for cross-validation That is why according to the choice of using cross validation, output format is not the same. Also, I added docstring so that we will later on be able to automatically generate a documentation of the code by using libairies such as Sphinx (here is a tutorial if you are interested : https://towardsdatascience.com/documenting-python-code-with-sphinx-554e1d6c4f6d or the link to the library : https://www.sphinx-doc.org/en/master/)