This framework uses just the BWT of the input string and a stack.
A programmable framework to find all the statistically over- and under-represented substrings of a string in small space and in parallel.