BUG: ML Tooling can't handle models with no y values

[andersbogsnes]

If we have a model that has no y values when training, ML-tooling will fail, as it always does a train_test_split which fails when y is None.

To Reproduce
Steps to reproduce the behavior:

1. Define a model where y=None

class RandomRankEstimator(BaseEstimator, ClassifierMixin):
    def fit(x, y=None):
          ... do modelstuff here...
    def predict(x):
        ...predict...

class CovRanker(ModelData):
       def get_training_data():
             return x, None
       def get_prediction_data(x):
            return x

model = CovRanker(RandomRankEstimator)
model.train_estimator()

Expected behavior
.train_estimator should not fail if no y value is supplied

Additional context
Add any other context about the problem here.

• I have updated ML Tooling to the latest version
• I have checked if this issue has been reported previously

[thomasfrederikhoeck]

Could be solved in #275

[andersbogsnes]

solved by #348