Hi all,
I have created a tool to replay S3 backups of raw data form realtime data pipeline: https://github.com/grzegorzewald/SnowplowRecovery
With no modifications it would consume gzipped backups and output them to stdout - just pipe output to local copy of realtime enrichment process.
If you need/want to fix data, look at lines 21-23. In theory you may do almost anything (i have used this to fix types).