How to configure an EMR Spark Datasource

This article is for comments to:

Please comment +1 if this How to is important to you.

It seems using the S3GlobReaderBatchKwargsGenerator will translate and s3:// path into s3a:// which messes up with spark being able to open file within the EMRFS context.
I might be doing something wrong, having documentation will uncover if it’s a bug or not