#! /usr/bin/env ruby

# This is a game that helps people understand genome assembly. Given a string,
# it generates sequence reads giving perfect coverage of the string and with a
# fixed overlap. The idea is to print the generated reads, cut them out, and
# have learners assemble them by hand. Different difficulties can be
# demonstrated by using a string with repeats, low complexity, etc., to mimic
# real assembly problems, or by adjusting the parameters (overlap and number
# of fragments).

# generate n approximately equally-sized fragments with each contiguous pair
# of fragments overlapping by k such that the entire quote is covered by the
# resulting fragments
def sequence_quote_with_overlap(quote, n_fragments, k)

  quote = quote.downcase

  # the quote length may not be divisible by the number of fragments,
  # so we distribute the remainder across the fragments randomly.
  remainder = quote.length % n_fragments
  bump = ([1] * remainder + [0] * (n_fragments - remainder)).shuffle
  fraglen = quote.length / n_fragments + k

  # because the final fragment will be too short by the overlap size,
  # we recover k characters from the other fragments at random
  (0...bump.length).to_a.sample(k).each{ |i| bump[i] -= 1 }

  # for each fragment, adjust the fragment length by the bump
  # and sequence the fragment from the quote
  fragments = []
  firstchar = 0
  adj_fraglen = fraglen + bump[0]
  lastchar = firstchar + adj_fraglen - 1

  (1...n_fragments).each do |i|
    adj_fraglen = fraglen + bump[i]
    fragments << quote[firstchar..lastchar]
    firstchar = lastchar - k + 1
    lastchar = firstchar + adj_fraglen - 1
  end

  # store the final fragment
  fragments << quote[firstchar..lastchar]

  fragments
end


# demos

# simple assembly

quote = "Try a thing you haven’t done three times. Once, to get over the fear of doing it. Twice, to learn how to do it. And a third time, to figure out whether you like it or not."

n_fragments = 18
k = 6

frags = sequence_quote_with_overlap quote, n_fragments, k
frags.shuffle.each { |f| puts "\"#{f}\"" }

# with repeats

quote = "Happiness resides not in possessions, and not in gold, happiness dwells in the soul."
k = 3
frags = sequence_quote_with_overlap quote, n_fragments, k
frags.shuffle.each { |f| puts "\"#{f}\"" }