es: Elasticsearch
            The Elasticsearch client
        """
        # Convert the new_predictions DataFrame to a list of dictionaries
        # The columns of the new_predicitons DataFrame are 'PMID', 'date', 'date_completed', 'label', 'text'
        # Only the 'PMID', 'date', 'date_completed', and 'label' columns are needed
        new_predictions = new_predictions[['PMID', 'date', 'date_completed', 'label']].to_dict(orient='records')

        logger.debug(f'Adding {len(new_predictions)} new predictions to the predictions index {index}: First 5: {new_predictions[:5]}')

        # Add the new predictions to the predictions index
        helpers.bulk(es, [{'_index': index, '_id': int(doc['PMID']), '_source': doc} for doc in new_predictions])

    def _compute_cluster_names(self, X: Union[List[str], pd.Series]) -> None:
        """
        Compute the cluster names

        Parameters
        ----------
        X : Union[List[str], Series]
            The documents
        y : np.ndarray
            The labels

        Returns
        -------
        Dict[int, str]
            The cluster names
        """
        raise NotImplementedError

    @abstractmethod
    def _preprocess(self, X: Union[List[str], pd.Series]) -> Union[List[str], pd.Series]:
        """
        Preprocess a batch of documents

        Parameters
        ----------
        X : Union[List[str], Series]
            The documents

        Returns
        -------
        Union[List[str], Series]
            The preprocessed documents
        """
        raise NotImplementedError

    @abstractmethod
    def _preprocess_single(self, x: str) -> str:
        """
        Preprocess a single document

        Parameters
        ----------
        x : str
            The document

        Returns
        -------
        st