require '/home/neil/programming/ruby/programming-collective/intelligence/k-means-clustering'

def read_genome_file(filename)
  genomes = []
  IO.foreach(filename) { |line| genomes << line.chomp.split('').collect {|i| i.to_i} }
  genomes
end

def find_centroid_sets(rows, k = 10, n = 10)
  centroid_set = []
  1.upto(n) do
    centroid_set << (k_means_cluster(rows, k, :domain => :discrete) {|x, y| hamming_distance x, y})
  end
  centroid_set
end

def order_centroids_by_cluster_size(centroids, rows)
  clusters = clusters_of_centroids(centroids, rows) {|x, y| hamming_distance x, y}
  cluster_sizes = clusters.collect {|c| c.length}
  sized_centroids = centroid.zip(cluster_sizes)
  sorted_sized_centroids = (sized_centroids.sort_by {|t| t[1]}).reverse
end

def best_centroids(centroid_set)
  centroid_set.collect {|tc| tc[0][0]}
end