Is there a way to grab parquet files from AWS S3? The simple UI for AWS S3 data source expects a csv files.