Remove Vietnamese Accent

remove_vietnamese_tones.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

"""Remove Vietnamese tones function"""
"""Author: Khiem Doan"""

uni_chars_l = 'áàảãạâấầẩẫậăắằẳẵặđèéẻẽẹêếềểễệíìỉĩịóòỏõọôốồổỗộơớờởỡợúùủũụưứừửữựýỳỷỹỵ'
uni_chars_u = 'ÁÀẢÃẠÂẤẦẨẪẬĂẮẰẲẴẶĐÈÉẺẼẸÊẾỀỂỄỆÍÌỈĨỊÓÒỎÕỌÔỐỒỔỖỘƠỚỜỞỠỢÚÙỦŨỤƯỨỪỬỮỰÝỲỶỸỴ'

no_tone_chars_l = 'a'*17 + 'd' + 'e'*11 + 'i'*5 + 'o'*17 + 'u'*11 + 'y'*5
no_tone_chars_u = 'A'*17 + 'D' + 'E'*11 + 'I'*5 + 'U'*11 + 'O'*17 + 'Y'*5

no_tone_dict = dict(zip(uni_chars_l + uni_chars_u, no_tone_chars_l + no_tone_chars_u))


def remove_vietnamese_tones(text):
    text = [no_tone_dict.get(c, c) for c in text]
    return ''.join(text)


if __name__ == '__main__':
    print(remove_vietnamese_tones('Khiêm Đoàn'))
    print(remove_vietnamese_tones('Đoàn Hoà Khiêm'))