from opencc import OpenCC
cc = OpenCC('t2s')  # 't2s'表示繁体转简体
from h_corpus import Fileset
import unicodedata
import re
import json
def fullwidth_to_halfwidth(input_str):
    return ''.join([unicodedata.normalize('NFKC', char) for char in input_str])

def clearT(s):
    s = cc.convert(fullwidth_to_halfwidth(s))
    return s.strip()

def safe_index(s, _s, _i):
    try:
        retn = s.index(_s, _i)
    except:
        return 0, ''
    return retn+1, s[_i:retn]

def get_p(s, sep=' ', _max=4):
    retn = []
    idx = 0
    for i in range(_max):
        idx, tmp = safe_index(s, sep, idx)
        if idx == 0:
            break
        retn.append(tmp)
    retn.append(s[idx:])
    return retn

a = Fileset(r'D:\datasets\tmp', ext='.ks')
b = r'D:\datasets\b-corpus\KAMIKIMIEX'
in_talk = False
_n = {
  "？？？": "?",
  "心の声": "旁白",
  "块斗": "块斗",
  "Ｃ‐ＡＩ": "C-AI",
  "月咏": "月咏",
  "雾香": "雾香",
  "喧哗的女生": "喧哗的女生",
  "爱凑热闹的女生": "爱凑热闹的女生",
  "帅气的男生": "帅气的男生",
  "拉娜": "拉娜",
  "值日生的山本": "值日生的山本",
  "目出岛": "目出岛",
  "塊斗": "块斗",
  "安保系统": "安保系统",
  "叶月": "叶月",
  "老师": "老师",
  "OldMan": "OldMan",
  "索菲娅": "索菲娅",
  "玲音": "玲音",
  "爱彩梨": "爱彩梨",
  "高个子的男生": "高个子的男生",
  "松冈": "松冈",
  "霧香": "雾香",
  "伊澄博士": "伊澄博士",
  "记者": "记者",
  "喜欢八卦的女生": "喜欢八卦的女生",
  "不受欢迎的男生": "不受欢迎的男生",
  "新闻部的男生": "新闻部的男生",
  "新闻部部长": "新闻部部长",
  "月咏1": "月咏1",
  "月咏2": "月咏2",
  "月咏3": "月咏3",
  "月咏们": "月咏们",
  "盖着帽子的男人": "盖着帽子的男人",
  "女性播报员": "女性播报员",
  "神秘的男子": "神秘的男子",
  "特工": "特工",
  "泽库斯": "泽库斯",
  "艾恩斯": "艾恩斯",
  "采访人员": "采访人员",
  "一年级男生": "一年级男生",
  "孤独者": "孤独者",
  "完美超人": "完美超人",
  "愛彩梨": "爱彩梨",
  "部员": "部员",
  "街头大屏幕": "街头大屏幕",
  "部員": "部员",
  "护士": "护士",
  "店员": "店员",
  "雾斗": "块斗",
  "樱木明里": "樱木明里",
  "黑江": "黑江",
  "小女孩": "小女孩",
  "医生": "医生",
  "啊斗": "块斗",
  "书记": "书记",
  "不认识的男人": "不认识的男人",
  "帅哥": "帅哥",
  "新闻主持人": "新闻主持人",
  "大妈": "大妈",
  "大叔": "大叔",
  "块斗＆雾香": "块斗＆雾香",
  "爱斗": "块斗",
  "女子篮球部员": "女子篮球部员",
  "顾問": "顾问",
  "对手队长": "对手队长",
  "ＡＩ": "AI",
  "流浪仓鼠": "流浪仓鼠",
  "爱爱凑热闹的女生": "爱爱凑热闹的女生",
  "ツクヨミ": "ツクヨミ",  # 月咏 先不改，便于定位后进行全文翻译
  "黄毛小混混": "黄毛小混混",
  "纹身小混混": "纹身小混混",
  "金发小混混": "金发小混混",
  "冬守": "冬守",
  "梦山": "梦山",
  "强壮的保安": "强壮的保安",
  "高个子保安": "高个子保安",
  "强悍的保安": "强悍的保安",
  "海枝所长": "海枝所长",
  "女新闻主持人": "女新闻主持人",
  "男新闻主持人": "男新闻主持人"
}
reg_name = re.compile(r' name=(.+?)($| )')
reg_cs = re.compile(r' text=(.+?)($| )')
d_s = []
idx = 0
sub_idx = 0
for i in range(len(a)):
    sc = []
    with open(a[i], encoding='UTF-16LE') as f:
        for line in f:
            if len(line) < idx+1:
                continue
            line = line[idx:]
            if line.startswith('else'):
                continue
            if line.startswith('@if') and not line.startswith('@if exp="ChkSelect'):
                idx += 1
                continue
            if line.startswith('endif') and idx > 0:
                idx -= 1
                continue
            if line.startswith('@AddSelect') and 'invalid' not in line:
                d_s.append(reg_cs.findall(line)[0][0])
            if (line.startswith('@if exp="ChkSelect') or line.startswith('elsif exp="ChkSelect')) and d_s:
                if line.startswith('@'):
                    idx += 1
                n = '选择'
                d = clearT(d_s[0])
                d_s = d_s[1:]
                if d:
                    sc.append(n + '：' + d)
            if line.startswith('@Hitret'):
                in_talk = False
            if in_talk:
                in_talk = False
                d = clearT(line)
                if d:
                    sc.append(n + '：' + d)
            if line.startswith(r'@Talk'):
                n = reg_name.findall(line)[0][0]
                n = clearT(_n.get(n, n))
                in_talk = True
    if sc:
        with open(b + f'\\{i}.txt', 'w', encoding='utf-8') as f:
            f.write('\n'.join(sc))

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