One of the projects on my secret TODO list is feeding Libgen into Elastic Search to get a cross referenced full text search. For now the hardware is prohibitively expensive, but time is on my side. I'm sure redundantly indexing a few 10s of TB will become trivial before the end of the decade.