AC多模匹配算法过滤敏感词实例

本文章摘编、转载需要注明来源 http://blog.csdn.net/shadowsick/article/details/8891939

在应用项目中很多时候都需要用到过滤敏感词的功能,自己写个遍历明显在小数据的时候还能凑合下,但是对于大数据的时候就有点力不从心了,这里推荐使用ac多模匹配算法


先来写个应用类

/**
 * AC多模匹配敏感字符工具类实现类
 * 
 * @author shadow
 * @email 124010356@qq.com
 * @create 2012.04.28
 */
public class AcUtilImpl implements AcUtil {

	public String contrast(String filters, String word, String regex) {

		if (null == filters || "".equals(filters) || null == word
				|| "".equals(word))
			return "";

		AhoCorasick ac = new AhoCorasick();
		String[] strings = StringUtils.split(filters, regex);
		for (String string : strings)
			ac.add(string.getBytes(), string);
		ac.prepare();
		return matching(ac, word);
	}

	public String contrast(String[] filters, String word) {

		if (null == filters || filters.length <= 0 || null == word
				|| "".equals(word))
			return "";

		AhoCorasick ac = new AhoCorasick();
		for (int i = 0, len = filters.length; i < len; i++) {
			ac.add(filters[i].getBytes(), filters[i]);
		}
		ac.prepare();
		return matching(ac, word);
	}

	public String contrast(List<String> filters, String word) {

		if (null == filters || filters.size() <= 0 || null == word
				|| "".equals(word))
			return "";

		AhoCorasick ac = new AhoCorasick();
		for (int i = 0, len = filters.size(); i < len; i++) {
			ac.add(filters.get(i).getBytes(), filters.get(i));
		}
		ac.prepare();
		return matching(ac, word);
	}

	private String matching(AhoCorasick ac, String word) {
		StringBuffer buffer = new StringBuffer();
		Iterator<?> iterator = ac.search(word.getBytes());
		while (iterator.hasNext()) {
			SearchResult result = (SearchResult) iterator.next();
			buffer.append(result.getOutputs()).append(",");
		}
		return buffer.length() > 0 ? buffer.substring(0, buffer.length() - 1)
				: "";
	}

	public static void main(String[] args) {
		String filters = "or,world,33,dd,test";
		String word = "hello world, how are you!";
		String regex = ",";
		String result = new AcUtilImpl().contrast(filters, word, regex);
		System.out.println(result);
	}
}

然后运行main函数测试下,获得的结果是

[or],[world]


这个插件的性能,匹配度也灰常不错,AhoCorasick这个类自己下载放到项目里就可以了

上一篇:comet4j java服务端推送消息到web页面实例


下一篇:在windows上秒开应用程序