Make preprocessing script work with any BIDS dataset

[hermancollin]

ATM, preprocessing.py downloads the SEM dataset from the github repo. However, the other datasets are not hosted publicly so we can't specify a URL for those. Downloading this data will have to be done separately using git-annex, and then the preprocessing script can be applied.

Instead of automatically downloading the SEM dataset, we should add an argument to specify the path to a BIDS dataset (that was downloaded prior).