def get_all_files_in_directory(directory, ext=''):
    import os
    import re
    custom_sort_key_re = re.compile('([0-9]+)')

    def custom_sort_key(s):
        # 将字符串中的数字部分转换为整数，然后进行排序
        return [int(x) if x.isdigit() else x for x in custom_sort_key_re.split(s)]

    all_files = []
    for root, dirs, files in os.walk(directory):
        for file in files:
            if file.endswith(ext):
                file_path = os.path.join(root, file)
                all_files.append(file_path)
    return sorted(all_files, key=custom_sort_key)


def clearT():
    import unicodedata
    from opencc import OpenCC

    def full2half(input_str):
        return ''.join([unicodedata.normalize('NFKC', char) for char in input_str])

    cc = OpenCC('t2s')  # 't2s'表示繁体转简体

    def _clearT(s):
        s = cc.convert(full2half(s))
        return s.lstrip('n').strip().strip(r'\n').replace('\n', '\\n')

    return _clearT


clearT = clearT()


def startsWithAny(s: str, keys):
    for x in keys:
        if s.startswith(x):
            return x
    else:
        return False


def startsWithAlnum(s: str, _chrs):
    retn = ''
    for char in s:
        if (char.isalnum()) or (char in _chrs):
            retn += char
        else:
            break
    return retn


def listRfind_idx(_i, _data, _condition):
    for j in range(_i - 1, -1, -1):
        if _condition(_data[j]):
            return j
    return -1


def listRfind(_i, _data, _condition, _filter, delta=1):
    return list(filter(_filter, _data[max(listRfind_idx(_i, _data, _condition) + delta, 0):_i + 1]))


def startsWithCmd(s: str, _chrs=None):
    if _chrs is None:
        _chrs = {'_', '@'}
    cmd = startsWithAlnum(s, _chrs)
    if cmd:
        return s.startswith(cmd + ' ')
    else:
        return False


# =================

a = get_all_files_in_directory(r'E:\tmp\幼性反应', ext='.txt')
b = r'D:\datasets\tmp'

# =================

sc = {}

_n = {}

# =================
for path in a:
    name = path[path.rindex('\\'):]
    if name not in sc:
        sc[name] = []
        print(name)
    # =================

    with open(path, 'r', encoding='utf-8') as f:
        data = list(filter(lambda x: not x.startswith(';'), (x.rstrip() for x in f.readlines())))


    csel = {}
    w_i = -1
    while w_i < len(data) - 1:
        w_i += 1
        line = data[w_i]
        if line.startswith('select'):
            while w_i < len(data) - 1:
                w_i += 1
                line = data[w_i]
                if not line:
                    break
                line = line.split(',*')
                assert len(line) == 2
                csel[line[1].rstrip(',')] =  '选择:' + clearT(line[0].strip('"')[1:])
                # print(csel)
            continue
        if line.startswith('*') and line.isascii():
            line = line[1:]
            if line in csel:
                sc[name].append('旁白' + '：' + csel[line])
            continue
        if not line.endswith('\\'):
            continue
        tmp = listRfind(w_i, data,
                        lambda x: x.startswith('dwave ') or x.endswith('\\') or not x,
                        lambda x: (x.startswith('dwave ') or not startsWithCmd(x)) and not(x.startswith('bgm')),
                        delta=0)
        if (len(tmp) > 1 and tmp[0].endswith('\\')) or not tmp[0]:
            tmp = tmp[1:]
        if not tmp:
            continue
        if tmp[0].startswith('dwave '):
            n = '由佳'
            tmp = tmp[1:]
            assert tmp
        else:
            n = '旁白'
        d = '\\n'.join(tmp)
        d = clearT(d)
        if d:
            sc[name].append(n + '：' + d)

# =================

for k, v in sc.items():
    with open(b + f'\\{k}.txt', 'w', encoding='utf-8') as f:
        f.write('\n'.join(v))

# =================
import json

tmp = json.dumps(_n, ensure_ascii=False, indent=2)