对于小文件:
# -*- coding: utf-8 -*-
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('-m', '--mode', default='f2h', help='full2half or half2full')
parser.add_argument('-i', '--input_file', help='input filename')
parser.add_argument('-o', '--output_file', default='output.txt', help='output filename')
args = parser.parse_args()
# 全角对应半角字典
# ASCII字符
full_to_half = {chr(i): chr(i - 0xfee0) for i in range(0xFF01, 0xFF5E + 1)}
full_to_half[chr(0x3000)] = chr(0x20)
# 半角对应全角字典
half_to_full = dict((v, k) for k, v in full_to_half.items())
def full2half(text):
"""全角转半角
:param text: 输入文本
:return: 输出文本
"""
return ''.join([full_to_half.get(i, i) for i in text])
def half2full(text):
"""半角转全角
:param text: 输入文本
:return: 输出文本
"""
return ''.join([half_to_full.get(i, i) for i in text])
def main():
with open(args.input_file, 'r', encoding='utf-8') as f_in, \
open(args.output_file, 'w', encoding='utf-8') as f_out:
text = f_in.read()
if args.mode == 'f2h':
output_text = full2half(text)
else:
output_text = half2full(text)
f_out.write(output_text)
if __name__ == '__main__':
main()
对于大文件:
# -*- coding: utf-8 -*-
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('-m', '--mode', default='f2h', help='full2half or half2full')
parser.add_argument('-i', '--input_file', help='input filename')
parser.add_argument('-o', '--output_file', default='output.txt', help='output filename')
args = parser.parse_args()
# 全角对应半角字典
# ASCII字符
full_to_half = {chr(i): chr(i - 0xfee0) for i in range(0xFF01, 0xFF5E + 1)}
full_to_half[chr(0x3000)] = chr(0x20)
# 半角对应全角字典
half_to_full = dict((v, k) for k, v in full_to_half.items())
def full2half(text):
"""全角转半角
:param text: 输入文本
:return: 输出文本
"""
return ''.join([full_to_half.get(i, i) for i in text])
def half2full(text):
"""半角转全角
:param text: 输入文本
:return: 输出文本
"""
return ''.join([half_to_full.get(i, i) for i in text])
def main():
with open(args.input_file, 'r', encoding='utf-8') as f_in, \
open(args.output_file, 'w', encoding='utf-8') as f_out:
for text in f_in:
if args.mode == 'f2h':
output_text = full2half(text)
else:
output_text = half2full(text)
f_out.write(output_text)
if __name__ == '__main__':
main()