Python 敏感词过滤的三种实现方式：Naive、BSF 与 DFA

山间有风

发布时间：2026-10-14 | 4人浏览过

原创

本文介绍了 Python 中实现敏感词过滤的三种常见方法。首先展示了基于循环和 replace 的简单实现；其次详解了利用宽度优先搜索（BSF）优化替换次数的方案，区分中英文处理逻辑；最后深入解析了使用确定有穷自动机（DFA）通过嵌套字典保持恒定性能的实现方式。

一个简单的实现

主要思路是通过遍历循环并结合字符串的 replace 方法来完成敏感词的替换操作。


class NaiveFilter():

    '''Filter Messages from keywords

    very simple filter implementation

    >>> f = NaiveFilter()
    >>> f.parse("filepath")
    >>> f.filter("hello sexy baby")
    hello **** baby
    '''

    def __init__(self):
        self.keywords = set([])

    def parse(self, path):
        for keyword in open(path):
            self.keywords.add(keyword.strip().decode('utf-8').lower())

    def filter(self, message, repl="*"):
        message = str(message).lower()
        for kw in self.keywords:
            message = message.replace(kw, repl)
        return message

使用 BSF（宽度优先搜索）进行实现

该方案对搜索查找过程进行了优化。针对英语单词，直接通过按词索引字典进行查找；而对于其他语言模式，则采用逐字符查找匹配的策略。

BFS：宽度优先搜索方式


class BSFilter:

    '''Filter Messages from keywords

    Use Back Sorted Mapping to reduce replacement times

    >>> f = BSFilter()
    >>> f.add("sexy")
    >>> f.filter("hello sexy baby")
    hello **** baby
    '''

    def __init__(self):
        self.keywords = []
        self.kwsets = set([])
        self.bsdict = defaultdict(set)
        self.pat_en = re.compile(r'^[0-9a-zA-Z]+$')  # english phrase or not

    def add(self, keyword):
        if not isinstance(keyword, str):
            keyword = keyword.decode('utf-8')
        keyword = keyword.lower()
        if keyword not in self.kwsets:
            self.keywords.append(keyword)
            self.kwsets.add(keyword)
            index = len(self.keywords) - 1
            for word in keyword.split():
                if self.pat_en.search(word):
                    self.bsdict[word].add(index)
                else:
                    for char in word:
                        self.bsdict[char].add(index)

    def parse(self, path):
        with open(path, "r") as f:
            for keyword in f:
                self.add(keyword.strip())

    def filter(self, message, repl="*"):
        if not isinstance(message, str):
            message = message.decode('utf-8')
        message = message.lower()
        for word in message.split():
            if self.pat_en.search(word):
                for index in self.bsdict[word]:
                    message = message.replace(self.keywords[index], repl)
            else:
                for char in word:
                    for index in self.bsdict[char]:
                        message = message.replace(self.keywords[index], repl)
        return message

使用 DFA（Deterministic Finite Automaton）进行实现

DFA 即 Deterministic Finite Automaton，也就是确定有穷自动机。该实现使用了嵌套的字典结构来构建状态机。


class DFAFilter():

    '''Filter Messages from keywords

    Use DFA to keep algorithm perform constantly

    >>> f = DFAFilter()
    >>> f.add("sexy")
    >>> f.filter("hello sexy baby")
    hello **** baby
    '''

    def __init__(self):
        self.keyword_chains = {}
        self.delimit = '\x00'

    def add(self, keyword):
        if not isinstance(keyword, str):
            keyword = keyword.decode('utf-8')
        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(keyword.strip())

    def filter(self, message, repl="*"):
        if not isinstance(message, str):
            message = message.decode('utf-8')
        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)

到此这篇关于 Python 敏感词过滤的实现示例的文章就介绍到这了，更多相关 Python 敏感词过滤内容请搜索本站以前的文章或继续浏览下面的相关文章希望大家以后多多支持本站！