PatchEvaluation.py 文件源码

def best_string_mapping(threshold, left_list, right_list):
    """
    This function tries to find the closest mapping with the best weight of two lists of strings.
    Example:

      List A        List B

    0:  'abc'         'abc'
    1:  'cde'         'cde'
    2:  'fgh'         'fgh
    3:                'fgj

    map_lists will try to map each element of List A to an element of List B, in respect to the given threshold.

    As a[{0,1,2}] == b[{0,1,2}], those values will automatically be mapped. Additionally, a[2] will also be mapped to
    b[3], if the threshold is low enough (cf. 0.5).
    """
    def injective_map(ll, rl, inverse_result=False):
        ret = dict()
        for l_entry in ll:
            for r_entry in rl:
                if l_entry == r_entry:
                    sim = 1
                else:
                    sim = fuzz.token_sort_ratio(l_entry, r_entry) / 100

                if sim < threshold:
                    continue

                if l_entry in ret:
                    _, old_sim = ret[l_entry]
                    if sim < old_sim:
                        continue

                ret[l_entry] = r_entry, sim
        return {(r, l) if inverse_result else (l, r) for l, (r, _) in ret.items()}

    return injective_map(left_list, right_list) | injective_map(right_list, left_list, True)