tools.py

import pandas as pd

KIFU_TO_SQUARE_NAMES = [
                    '１一', '１二', '１三', '１四', '１五', '１六', '１七', '１八', '１九',
                    '２一', '２二', '２三', '２四', '２五', '２六', '２七', '２八', '２九',
                    '３一', '３二', '３三', '３四', '３五', '３六', '３七', '３八', '３九',
                    '４一', '４二', '４三', '４四', '４五', '４六', '４七', '４八', '４九',
                    '５一', '５二', '５三', '５四', '５五', '５六', '５七', '５八', '５九',
                    '６一', '６二', '６三', '６四', '６五', '６六', '６七', '６八', '６九',
                    '７一', '７二', '７三', '７四', '７五', '７六', '７七', '７八', '７九',
                    '８一', '８二', '８三', '８四', '８五', '８六', '８七', '８八', '８九',
                    '９一', '９二', '９三', '９四', '９五', '９六', '９七', '９八', '９九',
                ]
KIFU_FROM_SQUARE_NAMES = [
                    '11', '12', '13', '14', '15', '16', '17', '18', '19',
                    '21', '22', '23', '24', '25', '26', '27', '28', '29',
                    '31', '32', '33', '34', '35', '36', '37', '38', '39',
                    '41', '42', '43', '44', '45', '46', '47', '48', '49',
                    '51', '52', '53', '54', '55', '56', '57', '58', '59',
                    '61', '62', '63', '64', '65', '66', '67', '68', '69',
                    '71', '72', '73', '74', '75', '76', '77', '78', '79',
                    '81', '82', '83', '84', '85', '86', '87', '88', '89',
                    '91', '92', '93', '94', '95', '96', '97', '98', '99',
                ]

def nomalize_precedence_name(df):
    #先手の対局者の名前から段位、タイトル名を削除する
    for x in range(len(df)):
        df["precedence_name"].iloc[x] = df["precedence_name"].iloc[x].replace(" ","").replace("　","").replace("\u3000","")
    if df["precedence_name"].iloc[x].endswith("段"):
        df["precedence_name"].iloc[x] = df["precedence_name"].iloc[x][:-2]
    df["precedence_name"].iloc[x] = df["precedence_name"].iloc[x].replace("十七世名人","").replace("十八世名人","").replace("十九世名人","")
    df["precedence_name"].iloc[x] = df["precedence_name"].iloc[x].replace("王将","").replace("王座","").replace("名人","").replace("竜王","").replace("棋聖","").replace("叡王","").replace("王位","").replace("棋王","")
    df["precedence_name"].iloc[x] = df["precedence_name"].iloc[x].replace("・","").replace("二冠","").replace("三冠","")
    return df

def nomalize_kif(df):
    for x in range(len(df)):
        kif = eval(df.iloc[x]["kif"])
        #kifの正規化処理　手数、消費時間を削除する
        cnt = -1
        for y in kif:
            cnt += 1
            while(1):
                    if "0" <= y[0] <= "9":
                        y = y[1:]
                        kif[cnt] = y
                    else:
                        break
            kif[cnt] = kif[cnt].replace("\u3000","")
            for z in range(len(y)):
                    if y[z] == "(":
                        kif[cnt] = y[:z]
                        break
        kifs = ""
        for i in kif:
            kifs += i.replace("\u3000","")
        df["kif"].iloc[x] = kifs
    return df

def nomalize_comment(df):
    #文章中のword省略処理
    for cnt in range(len(df["output"])):
        x = df["output"].iloc[cnt]
        read = x.split("。")
        #print(read)
        line = ""
        for z in read:
            if "期" in z or "出身" in z or "優勝" in z or "受賞" in z or "回" in z or "記録" in z or "棋士番号" in z or "勝" in z or "敗" in z or "名人" in z:
                pass
            elif "時" in z or "分" in z or "成績" in z or "棋戦" in z or "段" in z or "本日" in z or "立会" in z or "ABEMA" in z or "第" in z or "本局" in z:
                pass
            elif "対局" in z or "永世" in z:
                pass
            elif z == "":
                pass
            else:
                #print(z)
                line += z+"。"
        df["output"].iloc[cnt] = line
    return df

def accuracy_bestlist(df):
    cnt2 = 0
    num = 0
    for z in range(len(df)):
        blist = eval(df["bestlist"].iloc[z])
        b2list = eval(df["best2list"].iloc[z])
        te = eval(df["kif"].iloc[z])
        #print(blist[0][0])    
        #print(b2list[0][0])
        cnt = 0
        for x in range(1,len(te)):
            try:
                if blist[x-1][0] in te[x] or b2list[x-1][0] in te[x]:
                    cnt += 1
                #print(te[x],blist[x][0],b2list[x][0])
            except Exception as e:
                pass
        if cnt == 0:
            print("accuracy = 0",z)
        print("z = ",z," accuracy = ",cnt/len(te))
        cnt2 += cnt/len(te)
        num += 1
    print("mean_acuuracy",cnt2/num)

def nomalize_sfen(s):
    flag = 0
    movelist = []
    for x in range(len(s)):
            if x < 2:
                continue
            if len(s[x]) < 30 and flag == 0:
                #半角の指し手を全角に変換する
                temp = s[x].split()
                num = temp[1][0] + temp[1][1]
                for y in range(len(KIFU_FROM_SQUARE_NAMES)):
                    if num == KIFU_FROM_SQUARE_NAMES[y]:
                        sq = KIFU_TO_SQUARE_NAMES[y]
                word = sq+temp[1][2:]
                word = word.replace("竜","龍").replace("成銀","全").replace("成桂","圭").replace("成香","杏")
                if s[x].split()[1] not in ["投了" , "千日手" , "持将棋" , "反則勝ち"]:
                    movelist.append(word)
                else:
                    movelist.append(s[x].split()[1])
                    flag = 1
    return movelist

def make_triplets(df, column):
    # 重複を除いたユニークな文章リストを作成
    triplets = []
    for x in range(len(df)):
        anchor = df.iloc[x]
        # Anchorと同じではない文章をPositiveとして選択
        num = df.loc[(df[column] == anchor[column]) & (df["kif"] != anchor["kif"])].sample(n=1).index
        # print(df.loc[num])
        positive = df.loc[num]["kif"].values[0]
        
        # Anchorと異なる文章をNegativeとして選択
        num2 = df.loc[(df[column] != anchor[column]) & (df["kif"] != anchor["kif"])].sample(n=1).index
        # print(df.loc[num2])
        negative = df.loc[num2]["kif"].values[0]

        triplets.append((anchor["kif"], positive, negative,df.loc[num][column].values[0],df.loc[num2][column].values[0]))

def add_symbol(df,column):
    teban ="▲"
    kif = ""
    for x in range(len(df)):
        for y in df[column].iloc[x]:
            if y in ["０","１","２","３","４","５","６","７","８","９","同",0,1,2,3,4,5,6,7,8,9]:
                kif += teban + y
                if teban =="▲":
                    teban = "△"
                else:
                    teban = "▲"
            else:
                kif += y
        df[column].iloc[x] = kif
        kif = ""
    return df