Storing raw data in a compressed format

[PadLex]

I was wondering if there would be interest in a pull request that stores the raw data in a compressed format. It would greatly reduce the storage requirements and you could fit the raw books in this repository's binary releases. This would make the initial download much faster and relieve pressure from official gutenberg mirrors.

Python provides packages to directly read and write to compressed file formats so it wouldn't complicate the scripts much.