utils.py 文件源码

def get_best_similar(data):
    import difflib
    key, use_similar, similar_pool = data

    # try to find some close key in existing messages...
    # Optimized code inspired by difflib.get_close_matches (as we only need the best match).
    # We also consider to never make a match when len differs more than -len_key / 2, +len_key * 2 (which is valid
    # as long as use_similar is not below ~0.7).
    # Gives an overall ~20% of improvement!
    #tmp = difflib.get_close_matches(key[1], similar_pool, n=1, cutoff=use_similar)
    #if tmp:
        #tmp = tmp[0]
    tmp = None
    s = difflib.SequenceMatcher()
    s.set_seq2(key[1])
    len_key = len(key[1])
    min_len = len_key // 2
    max_len = len_key * 2
    for x in similar_pool:
        if min_len < len(x) < max_len:
            s.set_seq1(x)
            if s.real_quick_ratio() >= use_similar and s.quick_ratio() >= use_similar:
                sratio = s.ratio()
                if sratio >= use_similar:
                    tmp = x
                    use_similar = sratio
    return key, tmp