knapsack problem, solved in python

Knapsack problem can be described as follows: given n objects, whose weights are w[1], w[2], ..., w[n], respectively. There's a knapsack whose maximum capability is W. The question is: how to select objects from the n objects, so that the selected objects can be fitted in the knapsack?

Theoretically, dynamic progamming can solve knapsack problem in polynomial time. Inspired by this conclusion, I developed a proof-of-concept programme in Python, and here it is:

def knapsack(weiList, totalWei):
    elmSet = [set()] * (totalWei + 1)
    allSet = set(weiList)
    for s in range(1, totalWei + 1):
        for j in range(s):
            revSet = allSet - elmSet[j]
            for elm in revSet:
                if (sum(elmSet[j]) + elm == s):
                    elmSet[s] = elmSet[j].union([elm])
                    
    return elmSet[totalWei]

def knapsackTest():
    Weight = [1,3,5,7,9,11,13]
    for S in range(sum(Weight) + 1):
        kss = knapsack(Weight, S)
        if (len(kss)):
            assert(sum(kss) == S)
            print S, ":", list(kss)

if __name__ == '__main__':
    knapsackTest()