【python 走进NLP】两种高效过滤敏感词算法--DFA算法和AC自动机算法

一道bat面试题：快速替换10亿条标题中的5万个敏感词，有哪些解决思路？
有十亿个标题，存在一个文件中，一行一个标题。有5万个敏感词，存在另一个文件。写一个程序过滤掉所有标题中的所有敏感词，保存到另一个文件中。

1、DFA过滤敏感词算法

在实现文字过滤的算法中，DFA是比较好的实现算法。DFA即Deterministic Finite Automaton，也就是确定有穷自动机。
算法核心是建立了以敏感词为基础的许多敏感词树。

python 实现DFA算法：

# -*- coding:utf-8 -*-

import time
time1=time.time()

# DFA算法
class DFAFilter():
    def __init__(self):
        self.keyword_chains = {}
        self.delimit = '\x00'

    def add(self, keyword):
        keyword = keyword.lower()
        chars = keyword.strip()
        if not chars:
            return
        level = self.keyword_chains
        for i in range(len(chars)):
            if chars[i] in level:
                level = level[chars[i]]
            else:
                if not isinstance(level, dict):
                    break
                for j in range(i, len(chars)):
                    level[chars[j]] = {}
                    last_level, last_char = level, chars[j]
                    level = level[chars[j]]
                last_level[last_char] = {self.delimit: 0}
                break
        if i == len(chars) - 1:
            level[self.delimit] = 0

    def parse(self, path):
        with open(path,encoding='utf-8') as f:
            for keyword in f:
                self.add(str(keyword).strip())

    def filter(self, message, repl="*"):
        message = message.lower()
        ret = []
        start = 0
        while start < len(message):
            level = self.keyword_chains
            step_ins = 0
            for char in message[start:]:
                if char in level:
                    step_ins += 1
                    if self.delimit not in level[char]:
                        level = level[char]
                    else:
                        ret.append(repl * step_ins)
                        start += step_ins - 1
                        break
                else:
                    ret.append(message[start])
                    break
            else:
                ret.append(message[start])
            start += 1

        return ''.join(ret)


if __name__ == "__main__":
    gfw = DFAFilter()
    path="F:/文本反垃圾算法/sensitive_words.txt"
    gfw.parse(path)
    text="新疆骚乱苹果新品发布会雞八"
    result = gfw.filter(text)

    print(text)
    print(result)
    time2 = time.time()
    print('总共耗时：' + str(time2 - time1) + 's')

运行效果：

E:\laidefa\python.exe "E:/Program Files/pycharmproject/敏感词过滤算法/敏感词过滤算法DFA.py"
新疆骚乱苹果新品发布会雞八
****苹果新品发布会**
总共耗时：0.0010344982147216797s

Process finished with exit code 0

2、AC自动机过滤敏感词算法

AC自动机：一个常见的例子就是给出n个单词，再给出一段包含m个字符的文章，让你找出有多少个单词在文章里出现过。
简单地讲，AC自动机就是字典树+kmp算法+失配指针

# -*- coding:utf-8 -*-

import time
time1=time.time()

# AC自动机算法
class node(object):
    def __init__(self):
        self.next = {}
        self.fail = None
        self.isWord = False
        self.word = ""

class ac_automation(object):

    def __init__(self):
        self.root = node()

    # 添加敏感词函数
    def addword(self, word):
        temp_root = self.root
        for char in word:
            if char not in temp_root.next:
                temp_root.next[char] = node()
            temp_root = temp_root.next[char]
        temp_root.isWord = True
        temp_root.word = word

    # 失败指针函数
    def make_fail(self):
        temp_que = []
        temp_que.append(self.root)
        while len(temp_que) != 0:
            temp = temp_que.pop(0)
            p = None
            for key,value in temp.next.item():
                if temp == self.root:
                    temp.next[key].fail = self.root
                else:
                    p = temp.fail
                    while p is not None:
                        if key in p.next:
                            temp.next[key].fail = p.fail
                            break
                        p = p.fail
                    if p is None:
                        temp.next[key].fail = self.root
                temp_que.append(temp.next[key])

    # 查找敏感词函数
    def search(self, content):
        p = self.root
        result = []
        currentposition = 0

        while currentposition < len(content):
            word = content[currentposition]
            while word in p.next == False and p != self.root:
                p = p.fail

            if word in p.next:
                p = p.next[word]
            else:
                p = self.root

            if p.isWord:
                result.append(p.word)
                p = self.root
            currentposition += 1
        return result

    # 加载敏感词库函数
    def parse(self, path):
        with open(path,encoding='utf-8') as f:
            for keyword in f:
                self.addword(str(keyword).strip())

    # 敏感词替换函数
    def words_replace(self, text):
        """
        :param ah: AC自动机
        :param text: 文本
        :return: 过滤敏感词之后的文本
        """
        result = list(set(self.search(text)))
        for x in result:
            m = text.replace(x, '*' * len(x))
            text = m
        return text





if __name__ == '__main__':

    ah = ac_automation()
    path='F:/文本反垃圾算法/sensitive_words.txt'
    ah.parse(path)
    text1="新疆骚乱苹果新品发布会雞八"
    text2=ah.words_replace(text1)

    print(text1)
    print(text2)

    time2 = time.time()
    print('总共耗时：' + str(time2 - time1) + 's')

E:\laidefa\python.exe "E:/Program Files/pycharmproject/敏感词过滤算法/AC自动机过滤敏感词算法.py"
新疆骚乱苹果新品发布会雞八
****苹果新品发布会**
总共耗时：0.0010304450988769531s

Process finished with exit code 0

3、java 实现参考链接：
https://www.cnblogs.com/AlanLee/p/5329555.html

4、敏感词生成

# -*- coding:utf-8 -*-

path = 'F:/文本反垃圾算法/sensitive_worlds7.txt'
from 敏感词过滤算法.langconv import *
import pandas as pd
import pypinyin


# 文本转拼音
def pinyin(text):
    """
    :param text: 文本
    :return: 文本转拼音
    """
    gap = ' '
    piny = gap.join(pypinyin.lazy_pinyin(text))
    return piny


# 繁体转简体
def tradition2simple(text):
    """
    :param text: 要过滤的文本
    :return: 繁体转简体函数
    """
    line = Converter('zh-hans').convert(text)
    return line



data=pd.read_csv(path,sep='\t')

chinise_lable=[]
chinise_type=data['type']

for i in data['lable']:
    line=tradition2simple(i)
    chinise_lable.append(line)


chg_data=pd.DataFrame({'lable':chinise_lable,'type':chinise_type})



eng_lable=[]
eng_type=data['type']
for i in data['lable']:
    # print(i)
    piny=pinyin(i)
    # print(piny)
    eng_lable.append(piny)

eng_data=pd.DataFrame({'lable':eng_lable,'type':eng_type})
# print(eng_data)
# 合并
result=chg_data.append(eng_data,ignore_index=True)

# 数据框去重

res = result.drop_duplicates()
print(res)

# 输出
res.to_csv('F:/文本反垃圾算法/中英混合的敏感词10.txt',header=True,index=False,sep='\t',encoding='utf-8')