def get_predictions()

in getting_started/utils/lookout_equipment_utils.py [0:0]

• 14 lines of code
• 4 McCabe index (conditional complexity)

    def get_predictions(self):
        """
        This method loops through all the inference executions and build a
        dataframe with all the predictions generated by the model
        
        RETURNS
        =======
            results_df: pandas.DataFrame
                A dataframe with one prediction by row (1 for an anomaly or 0
                otherwise). Each row is indexed by timestamp.
        """
        # Fetch the list of execution summaries if there were not queried yet
        if self.execution_summaries is None:
            _ = self.list_inference_executions()
            
        # Loops through the executions summaries:
        results_df = []
        for execution_summary in self.execution_summaries:
            bucket = execution_summary['CustomerResultObject']['Bucket']
            key = execution_summary['CustomerResultObject']['Key']
            fname = f's3://{bucket}/{key}'
            results_df.append(pd.read_csv(fname, header=None))

        # Assembles them into a DataFrame:
        results_df = pd.concat(results_df, axis='index')
        results_df.columns = ['Timestamp', 'Predictions']
        results_df['Timestamp'] = pd.to_datetime(results_df['Timestamp'])
        results_df = results_df.set_index('Timestamp')
        
        return results_df