feature request: only apply random noise in `RandomAdder` to training data

Currently, RandomAdder adds noise to data both at training and at prediction time. This causes predictions to become non-deterministic and it offers no clear benefit in most cases I can think of.

I suggest changing the default behaviour of the transformer to only add random noise to the train data and optionally through a constructor flag also to the prediction data.

[MBrouns]

So I got asked a similar type of thing in todays training where someone wanted to drop rows with too many missing values from train but not from test so I was toying around to see if I could find something that would work.

I might have figured out a way but I’m not sure I like it all that much:

import pandas as pd
import hashlib

class TrainOnlyMixin(BaseEstimator, TransformerMixin):
    
    def fit(self, X, y):
        self.df_hash_ = self.hash_df(X)
        return self
    
    
    @staticmethod
    def hash_df(df):
        return hashlib.sha256(pd.util.hash_pandas_object(df, index=True).values).hexdigest()
    
    def transform(self, X, y=None):
    
        if self.hash_df(X) == self.df_hash_:
            return self.transform_train(X)
        
        else:
            return self.transform_test(X)

I basically store a hash of the train dataframe and compare X with it in transform and then call transform_train or transform_test. I think this can be made quite generic and I can’t think of a case where it wouldn’t work. What do you think?

[MBrouns]

#81