Myers Diff paper -- 2

Declaration: Here is the original paper link . It is written by Eugene W. Myers at 1986. I (j_now. You may also call me cydooc) translated parts of the thesis in to Chinese with a few annotations. If it offends your copyright, please leave me a message. I will remove this post ASAP. Thanks.

Myers Diff Algorithm

Constant MAX ∈ [0,M+N]
Var V: Array [−MAX .. MAX] of Integer
   V[1] ← 0
   For D ← 0 to MAX Do
      For k ← −D to D in steps of 2 Do
         If k = −D or k ≠ D and V[k − 1] < V[k + 1] Then
            x ← V[k + 1]
         Else
            x ← V[k − 1]+1
         y ← x − k
         While x < N and y < M and a x + 1 = by + 1 Do (x,y) ← (x+1,y+1)
        V[k] ← x
        If x ≥ N and y ≥ M Then
           Length of an SES is D
           Stop
   Length of an SES is greater than MAX

Ruby Implementation

#!/usr/bin/env ruby

class V
  def initialize(n, m, undefined_value)
    @vn = []
    @vm = []
    @m  = m
    @n  = n
    @dv = undefined_value
  end

  def [](i)
    ensure_size i
    r = 0 > i ?  @vm[-i-1] : @vn[i]
    r ? r : @dv
  end

  def []=(i, value)
    ensure_size i

    if i < 0
      @vm[-i-1] = value
    else
      @vn[i] = value
    end
  end

  private
  def ensure_size(i)
    raise "#{i} is out of boundary, should be in [#{-@m}, #{@n}]" if -@m > i or @n < i
  end
end


A = 'abcabba'
B = 'cbabac'

N = A.length
M = B.length

# v is an array of cordinate x. For the v[i], it means
# the max x cordinate in i-diagonal. i-diagonal is a
# set of vertexes that belongs to the line y = x - i.
v = V.new(N, M, -1)
v[1] = 0
done = nil

# A = 'abcabba'
# B = 'cbabac'
0.upto(N + M) do |d|
  puts "Calculating #{d}-path: "
  (-d).step(d, 2) do |k|
    if k == -d or k != d and v[k-1] < v[k+1]
      # (k+1)-path: move down  from the k-path's end point
      x = v[k+1]
      direction = "move down"
    else
      # (k+1)-path: move right from the k-path's end point
      x = v[k-1] + 1
      direction = "move right"
    end
    puts "   k: #{k}, x: #{x}. #{direction}."

    y = x - k

    # String is indexed from 0. So A[x] == B[y] is judging
    # if the diagonal is a snake. i.e. A[x] == B[y] is the
    # same meaning as the paper a[x + 1] = b[y + 1]
    while x < N and y < M and A[x] == B[y]
      x, y = x+1, y+1
      puts "      move along the diagonal (#{x-1}, #{y-1})->(#{x}, #{y})"
    end

    v[k] = x
    if x >= N and y >= M
      done = 1
      break
    end
  end
  break if done
end

Outputs

Calculating 0-path: 
   k: 0, x: 0. move down.
Calculating 1-path: 
   k: -1, x: 0. move down.
   k: 1, x: 1. move right.
Calculating 2-path: 
   k: -2, x: 0. move down.
      move along the diagonal (0, 2)->(1, 3)
      move along the diagonal (1, 3)->(2, 4)
   k: 0, x: 1. move down.
      move along the diagonal (1, 1)->(2, 2)
   k: 2, x: 2. move right.
      move along the diagonal (2, 0)->(3, 1)
Calculating 3-path: 
   k: -3, x: 2. move down.
      move along the diagonal (2, 5)->(3, 6)
   k: -1, x: 3. move right.
      move along the diagonal (3, 4)->(4, 5)
   k: 1, x: 3. move down.
      move along the diagonal (3, 2)->(4, 3)
      move along the diagonal (4, 3)->(5, 4)
   k: 3, x: 4. move right.
      move along the diagonal (4, 1)->(5, 2)
Calculating 4-path: 
   k: -4, x: 3. move down.
   k: -2, x: 4. move down.
   k: 0, x: 5. move down.
   k: 2, x: 6. move right.
      move along the diagonal (6, 4)->(7, 5)
   k: 4, x: 6. move right.
      move along the diagonal (6, 2)->(7, 3)
Calculating 5-path: 
   k: -5, x: 3. move down.
   k: -3, x: 4. move down.
   k: -1, x: 5. move down.
   k: 1, x: 7. move down.

Graphical Representation

Reference:

• Eugene W. Myers paper: An O(ND) Difference Algorithm and Its Variations
  
• Investigating Myers' diff algorithm
• See another post: Myers Diff paper -- 0