Sicket learn Multi label encoder with automatic most frequent imputation of unseen labels

encoding.py

class EncodeCategorical(TransformerMixin):
    """
    Encodes a specified list of columns or all columns if None. Impute unseen labels with most frequent value per column
    """

    def __init__(self, columns=None):
        self.columns = columns
        self.encoders = None
        self.fillNewLabels = None

    def fit(self, X, y, **kwargs):
        """
        Expects a data frame with named columns to encode.
        """
        # Encode all columns if columns is None
        if self.columns is None:
            self.columns = X.columns

        # Fit a label encoder for each column in the data frame
        self.encoders = {}
        for column in self.columns:
            # to fix weird unorderable types: str() > float() issue
            # when loading data via loadNeverDataMultipleFiles or loadNeverDataMultipleFilesDirect
            self.encoders[column] = LabelEncoder().fit(X[column].astype(str))

        # collect most frequent label per column as replacement for unseen labels
        self.fillNewLabels = pd.Series([X[c].value_counts().index[0] for c in self.columns],
                                       index=self.columns)
        return self

    def transform(self, data):
        """
        Uses the encoders to transform a data frame.
        """
        for column, encoder in self.encoders.items():
            fittedLabels = encoder.classes_
            replacementForUnseen = self.fillNewLabels[column]
            data[column] = data[column].astype(str)
            data.loc[~data[column].isin(fittedLabels), column] = str(replacementForUnseen)
            data.loc[:, column] = encoder.transform(data[column])  # .astype(str))
        return data