| 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 |