initial preprocessing
stores results into stemmed-*.txt
% python preproc.py

create document vector for 28 files and computes tfidf weights
prints everything onto the screen
which I manually cut and paste into a file called cossim-matrix.txt
% python docvector.py

reads in content from cossim-matrix.txt into numeric matrix
repeatedly finds smallest distance then spits out the pair of documents
% python hiercluster.py