How to send pyspark results to results.json in Athena Spark start_calculation_execution?

If I try to return the dataframe or use json.dumps(df) in Athena Spark start_calculation_execution codeBlock, they results appear in stdout file in S3 folder for the session and query execution. How do I make it appear in the results.json instead? It appears that the results.json is always empty ({}). And the stdout is always escaping the contents which is creating blockers downstream and I am not sure if that is the right place to store execution results.