MarkovSetN nth order markov set

superclass: LookupMarkovSet

constant order set. for variable order see ShannonFinger

uses an optimized identity lookup in which arrays of elements are stored internally as symbols.

instance creation:

*new(args, order,  updateSeeds)

args

a list of arrays that each represent 

one node in the set:

[[array_of_elements], [next_elements], [weigths]]

array_of_elements

should be of size 'order'.

next_elements

the next node is searched by simple equality.

if nil, this is a terminator node

weights

the array is normalized.

if weigths is nil, equal weight is used

if args is nil, the set is created and may be trained by the 

other methods.

order

the order of the set.

updateSeeds

if set to true, each element is always added to the seeds.

*fill( n, stream, order ) 

n number of items to read

stream a function or stream that returns items to read

order order of the new set

other methods see superclass MarkovSet

// direct node definition

(

m = MarkovSetN([

[[100, 120], [120, 130]],

[[120, 100], [100, 1900]]

], 2);

)

8.do { m.next([100, 120]).postln };

8.do { m.next([120, 100]).postln };

// training by a stream

m = MarkovSetN.fill(100, Pseq([1, 2, 3, 4, Prand([5, 55, 555])], inf).asStream, 2);

8.do { m.next([1,2]).postln };

8.do { m.next([3,4]).postln };

8.do { m.next([4,5]).postln };