Unfortunately we're not just searching for things but extracting word frequencies of every user for stylometric analysis, so we need to do custom crunching. Sp

Unfortunately we're not just searching for things but extracting word frequencies of every user for stylometric analysis, so we need to do custom crunching. Spreading this task into many sub-slices of the files is annoying because the frequencies per user add up quite a lot, which results in quite a massive amount of data.