Memoization using dictionary

Question

So I am trying to implement Lowest common subsequence in Python and was trying this alternative to my previous solution. I tried using a dictionary instead of a 2-D matrix to memoize the results.

def lcs(s1, s2):

    cache = {}
    if len(s1) == 0 or len(s2) == 0:
        return 0
    if (s1, s2) in cache:
        return cache[s1, s2]
    else:
        if s1[-1] == s2[-1]:
            cache[s1, s2] = 1 + lcs(s1[:-1], s2[:-1])
        else:
            cache[s1, s2] = max(lcs(s1[:-1], s2), lcs(s1, s2[:-1]))
    print cache

It's returning

TypeError: unsupported operand type(s) for +: 'int' and 'NoneType'

which I understand is because I am not returning anything so how can I do something like this.

return cache[s1, s2] = 1 + lcs(s1[:-1], s2[:-1])

And I am trying to implement it without using any decorators.

Answer 1

Try this

cache = {}
def lcs(s1, s2):
  global cache
  if len(s1) == 0 or len(s2) == 0:
      return 0
  if (s1, s2) in cache:
      return cache[(s1, s2)]
  else:
      if s1[-1] == s2[-1]:
          cache[(s1, s2)] = 1 + lcs(s1[:-1], s2[:-1])
      else:
          cache[(s1, s2)] = max(lcs(s1[:-1], s2), lcs(s1, s2[:-1]))
  return cache[(s1, s2)]