shuffle produces “randomized” or “quasi-randomized” sequences.
It allows to select permutations that avoid repetitions of similar items: The user can set limits on the number of consecutive occurrences of items associated with some labels, or, can specify the minimal distance separating two items with the same labels. This is useful to extract random samples from a dictionary, or to generate quasi-randomized lists of stimuli that avoid runs of trials belonging to the same categories (see shuffle.pdf for more details).
In practice, shuffle reads text lines and then outputs them (or a subset of them) in a random order, possibly with constraints on successive lines. If no contraint is specified, all permutations are equiprobable. Otherwise, constraints can be specified as maximum numbers of repetitions of labels in one or several columns of the output, or on the minimum distance between successive repetitions.
There are three implementations, in AWK, Perl and Python.
Here are some examples of command lines with the AWK implementation,
shuffle -? # to display a short help
shuffle sample.txt
shuffle n=10 sample.txt # limits output to 10 lines
shuffle n=10 sample.txt # new permutation of 10 lines
shuffle n=5 seed=134 sample.txt # sets the seed of the random generator
shuffle n=5 seed=134 sample.txt # you get the same permutation as before...
shuffle n=8 constr='2' sample.txt # no more than 2 succesive lines with
# the same value in column 1
shuffle n=8 constr='0 3' sample.txt # no more than 3 succesive lines with
# the same value in column 2
shuffle n=8 constr='2 2' sample.txt # no more than 2 succesive lines with
# the same values in column 1 or column 2
shuffle n=8 constr='1 1' sample.txt # not repetition of colum 1 and 2 on successive lines
Christophe@Pallier.org