To not double-commit every library change, we simply export ALL of
the zotero library into a single file in the 'intermediate' data
directory. Technically this still works just as well since it
still reflects our 'intermediate' stage of tagging, screening,
keywording the library contents. It just contains the non-sampled
contents as well now.
Added a notebook which has the sole point of quickly allowing me
to grab and look at the data of the processed sample I am creating.
I.e. quickly list and uniq all interventions/outcomes/inequalities,
doing a tiny calculation or similar.
