Rough code for comparing document similarity with MinHash

minhash.py

def minhash(text, window=25):  # assume len(text) > 50
    hashes = [murmurhash(text[i:i+window]) for i in range(len(text)-window+1)]
    return set(sorted(hashes)[0:20])

def similarity(text1, text2):
    hashes1 = minhash(text1)
    hashes2 = minhash(text2)
    return len(hashes1 & hashes2) / len(hashes1)

A = "one two three four five six seven eight nine ten eleven twelve thirteen fourteen fifteen sixteen seventeen eighteen"
B = "one two three four   5   6    7     8     9  ten eleven twelve thirteen fourteen fifteen sixteen seventeen eighteen"
C = " 1   2    3     4    5   6    7     8     9   10   11     12       13       14      15     16       17        18"

similarity(A, A)  # 1.0
similarity(A, B)  # 0.6
similarity(B, C)  # 0.1