Python 敏感词过滤的实现

一个简单的实现

class NaiveFilter():

    '''Filter Messages from keywords

    very simple filter implementation

    >>> f = NaiveFilter()
    >>> f.add("sexy")
    >>> 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

其中strip() 函数 删除附近的一些空格，解码采用utf-8的形式，然后将其转为小写。
parse()函数就是打开文件，然后从中取各个关键词，然后将其存在关键词集合中。

filter()函数是一个过滤器函数，其中将消息转化为小写，然后将关键词替换成*。、

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

在上面的实现例子中，对于搜索查找进行了优化，对于英语单词，直接进行了按词索引字典查找。对于其他语言模式，我们采用逐字符查找匹配的一种模式。
BFS：宽度优先搜索方式。

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)

DFA即Deterministic Finite Automaton，也就是确定有穷自动机。
使用了嵌套的字典来实现。

参考

Github:敏感词过滤系统