Fuzzy Jaccard similarity based search

fuzzy.py

def fuzzy_find(term, lst, threshold=0.75, shingle_len=3):
    # Create a set of n-grams of length shingle_len
    def shingle(s):
        shingles = []
        for t in re.split(' ', s):
            for i in range(max(1, len(t) - shingle_len)):
                shingles.append(t[i:i+shingle_len])
        return set(shingles)

    # Computes the Jaccard similarity between two sets 
    def jaccard_sim(A, B):
        return len(A & B) / len(A | B)

    # Shingle the search term
    term_shingles = shingle(term)

    # Search for the best-matching element in lst
    best_idx, best_score = None, None
    for i, elem in enumerate(lst):
        score = jaccard_sim(term_shingles, shingle(elem))
        if (i == 0) or (score > best_score):
            best_idx = i
            best_score = score

    # Return the best-matching element if the score is above the threshold
    if (not best_idx is None) and (best_score > threshold):
        print(term, lst[best_idx], best_score)
        return best_idx

    return None