文字处理篇——巧妙拆分中英文

文字处理—巧拆中英文

有时在进行问题时，需要把英文和中文拆开进行处理，这是就要用到拆解代码了！

#coding=utf-8

def split_ascii(s):
    for ii, ch in enumerate(s):
        if not ch.isspace() and ch.isascii():
            break
    return s[ii:],s[:ii].rstrip()

def first_deal():
    with open(rf"F:\物理练习图\九年级\字典组建试题\英语源DOC\ba1_1.txt") as f:
        quizFile = open(r'F:\物理练习图\九年级\字典组建试题\英语源DOC\ba1_1.txt','w',encoding='utf-8')
        #for num,line in enumerate(f):
        for line in f:
            a,b = split_ascii(line)
            a = ''.join(a)
            b = b.split('\n')
            b = ''.join(b)
            quizFile.write(b)
            quizFile.write('|')
            quizFile.write(a)
            quizFile.write('\n')
        quizFile.close()
            #print(b)

def second_deal():
    import re
    with open(rf"F:\物理练习图\九年级\字典组建试题\英语源DOC\ba1_1.txt") as f:
        quizFile = open(r'F:\物理练习图\九年级\字典组建试题\英语源DOC\ba1_last.txt', 'w', encoding='utf-8')
        for line in f:
            en_letter = '[\u0041-\u005a|\u0061-\u007a]+'  # 大小写英文字母
            zh_char = '[\u4e00-\u9fa5]+'  # 中文字符
            en = re.findall(en_letter,line)
            en = ' '.join(en)

            zh = re.findall(zh_char,line)
            zh = ''.join(zh)
            quizFile.write(zh)
            quizFile.write(en)
            quizFile.write('\n')

        quizFile.close()
second_deal()