[META] Synthetic data corpus generator

[gkamat]

Is your feature request related to a problem? Please describe

While OSB does come with a number of workloads and associated data corpora, they are limited in scope and size. Being able to generate documents synthetically based upon a specification provided by the user would be a useful capability. This will permit arbitrary-sized corpora to be created as well, and will work well with the data-stream model.