CYK 算法 Python 實現

CYK 算法 Python 實現

原理參考：https://blog.csdn.net/ssaalkjhgf/article/details/80435676

實現過程

類說明

CNF

由一個字典初始化，輸入爲規範的範式

find_product方法實現輸入一個變元組合返回其生成變元

CYK

由一個CNF初始化，find方法用於判斷字符串是否在CNF中

並保存推導過程，即動態規劃數組

Input

實現自己的輸入函數，一直等待輸入直到輸入EOF（Ctrl + Z）

並對輸入字符串進行簡單處理，切分

輸入字符串格式爲：A -> BC|AD|b

其中空格可含可不含

Visualization

可視化類，將推導過程輸出到.csv文件中

將.csv文件使用Excel打開便可查看完整推導過程

代碼中已有完整註釋和使用說明

代碼

'''

Used for judging a string whether be in the CFL (presented in CNF) using CYK algorithm
Author: JOke-Lin
Time: 2019-4-27
Vision: Python 3.7
Introduction: 
    For reading CNF you should use the format "A -> BC|a|b" (the blank space is allowed except "->")
    The first character will be considered as CFL's Start Symbol
    And end inputing by input "Ctrl + Z"
    For testing strings, you can input as many as strings you want
    And exit by using "ctrl + Z"
    For Output, the output format is the .csv file whcih should be checked by Excel
    The output file name can be changed in codes (the default is "result.csv")
Attention:
    When you test a new string, you must closed the last .csv file
    Otherwise you will meet a I/O Error because the read-write conflict

'''


class CNF:

    def __init__(self, cnf: dict):
        ''' cnf: [str:list] '''
        self.cnf = cnf
        self.S = list(cnf.keys())[0]

    def find_product(self, target: str):
        ''' find the productor  '''
        res = []
        for key, value in self.cnf.items():
            if target in value:
                res.append(key)
        return res


class CYK:

    def __init__(self, cnf: CNF):
        self.cnf = cnf

    def find(self, target: str):
        ''' judge if the target is in the CNF '''
        self.dp_mat = [[""]*len(target) for i in range(len(target))]
        # initialize
        for i in range(len(target)):
            self.dp_mat[i][i] = list(set(self.cnf.find_product(target[i])))

        for length in range(2, len(target) + 1):
            # i,j are the position in dp_mat
            for i in range(len(target) - length + 1):
                j = i + length - 1
                temp = []
                # split the sub_target
                for k in range(i, j):
                    list1 = self.dp_mat[i][k]
                    list2 = self.dp_mat[k+1][j]
                    temp += self.find_product(list1, list2)
                self.dp_mat[i][j] = list(set(temp))

        if self.cnf.S in self.dp_mat[0][len(target) - 1]:
            return True
        return False

    def find_product(self, list1: list, list2: list):
        ''' find all productions from Cartesian product of two lists '''
        res = []
        for i in range(len(list1)):
            for j in range(len(list2)):
                temp = self.cnf.find_product(list1[i]+list2[j])
                if temp != None:
                    res += temp
        return res


class Input:
    ''' my input class for reading the CNF '''

    def getDictItemFromInput(self, s: str):
        s = s.replace(' ', '')  # delete extra blank spacing
        temp = s.split("->")
        return temp[0], temp[1].split("|")

    def getInput(self):
        ''' return a dict CNF '''
        res = {}
        while True:
            try:
                s = input()
                dict_item = self.getDictItemFromInput(s)
                res[dict_item[0]] = dict_item[1]
            except:
                break
        return res


class Visualization:
    '''
    make the deduction visualized
    use dp_mat from class CYK to initialize
    and creat .csv file to show in Excel
    '''

    def __init__(self, data: list):
        self.data = [[""]*len(data) for i in range(len(data))]
        for i in range(len(data)):
            temp = data[i][i:]
            for j in range(len(data) - i):
                self.data[len(data) - j - 1][i] = temp[j]

    def seeAsCsv(self, filename: str):
        ''' write data to filename as .csv '''
        with open(filename, "w") as f:
            for i in range(len(self.data)):
                line = ""
                for j in range(len(self.data[i])):
                    temp = ""
                    if type(self.data[i][j]) == str:
                        temp = self.data[i][j]
                    else:
                        temp = "\"X{},{}=".format(j+1,len(self.data)-i)+"{"
                        for k in range(len(self.data[i][j])):
                            temp = temp + self.data[i][j][k]
                            if k != len(self.data[i][j]) - 1:
                                temp = temp + ","
                        temp = temp + "}\""
                    line = line + temp + ","
                f.write(line + "\n")


if __name__ == "__main__":
    print("Please Input the CNF: (The First Character must be Start symbol until Ctrl+Z )")
    myinput = Input()
    # get the CNF read until miss ctrl + Z
    data_dict = myinput.getInput()
    # creat my CYK algorithm
    cyk = CYK(CNF(data_dict))
    print("input the string: (until Ctrl+Z)")
    # test the input string until read the ctrl + z
    while True:
        try:
            target_str = input()
            if cyk.find(target_str):
                print("Yes")
            else:
                print("No")
            visual = Visualization(cyk.dp_mat)
            try:
                visual.seeAsCsv("result.csv")
            except:
                print("Write Error: Please Close Last csv file")
        except:
            break