汉语唤醒词的模糊判断（Python）

说明

这是一个简单的汉语模糊唤醒词的判断器，汉语发音中前后舌以及声母韵母的区别，如果进行精准判断，很容易误判。需要一个模糊判断的逻辑!

安装库

pip install pypinyin

代码Demo

# -*- coding: utf-8 -*-
# @Author : Dony YUAN
# @Time : 2024/12/06 20:02
from pypinyin import pinyin, Style


def compare_hanz(hanz_x:str, hanz_y:str, thresh:float=0.2)->bool:
    """
    
    :param hanz_x: 需要对比的汉字
    :param hanz_y: 需要对比的汉字
    :param thresh: 相似度阈值, default 0.2
    :return: bool 相似返回 True
    """
    py1 = pinyin(hanz_x, style=Style.NORMAL, errors='ignore')
    py2 = pinyin(hanz_y, style=Style.NORMAL, errors='ignore')
    py1 = [i[0] for i in py1]
    py2 = [i[0] for i in py2]
    print(py1, py2)
    diff = 0
    for m, n in zip(py1, py2):
        if m != n:
            l_m = len(m)
            l_n = len(n)
            if l_m > l_n:
                for x in n:
                    if x not in m:
                        diff += 1
            else:
                for x in m:
                    if x not in n:
                        diff += 1
            diff += abs(l_m - l_n)
    total_letters = max(len("".join(py1)), len("".join(py2)))
    error_rate = diff / total_letters
    print(f"error_rate: {error_rate}")
    # return "".join(py1) == "".join(py2)
    return error_rate < thresh


if __name__ == '__main__':
    wd1 = "您好小杜？"
    wd2 = "你好小布！"
    is_same = compare_hanz(wd1, wd2)
    print(f"是否相似:{is_same}")

[‘nin’, ‘hao’, ‘xiao’, ‘du’] [‘ni’, ‘hao’, ‘xiao’, ‘bu’]
error_rate: 0.16666666666666666
是否相似:True

其他

如大佬有更简便的方法，还请不吝赐教~