哈工大ltp应用java版

ゝ一世哀愁。 2023-02-14 10:51 15阅读 0赞

哈工大的ltp机器学习python版本很多，现在用java版本实现以下，网上有很多的方法是需要自己编译，我这边编译完成了

model下载：  
链接: https://pan.baidu.com/s/1HDaZpsrPHDcu8P15ho41VQ 提取码: dw9x  
ltp编译后的文件：  
https://download.csdn.net/download/qq\_16613311/12489534

如果不想打赏积分也可以参考进行自己编译：  
http://codepub.cn/2015/05/07/Compile-the-Language-Technology-Platform(C++)-and-LTP4J(Java)source-code/

下边是我写的相关类：

package com.dbapp.database.scanning.util;
    
    import edu.hit.ir.ltp4j.NER;
    import edu.hit.ir.ltp4j.Postagger;
    import edu.hit.ir.ltp4j.Segmentor;
    import lombok.extern.slf4j.Slf4j;
    
    import java.math.BigDecimal;
    import java.util.ArrayList;
    import java.util.List;
    
    /**
     * @ClassName: ltp4jUtil
     * 哈工大ltp机器学习引入
     * @Author: yongtao.ding on 2020/5/25 16:41
     */
    @Slf4j
    public class Ltp4jUtil {
    
    	//中文地名
    	public static String ADDRESS_C="ns";
    	//地名
    	public static String ADDRESS_L="nl";
    	//人名
    	public static String  PERSONNAME_C="nh";
    	//机构团体名
    	public static String GROUPS_G="ni";
    	//公检法
    	public static String GOVERNMENT="j";
    
    	/**
    	 * 判断一句话中是否存在地名，先使用segmentor将内容分词，使用postagger判断是否为地名
    	 * @param postagger
    	 * @param segmentor
    	 * @param content 内容
    	 * @return
    	 */
    	public static Boolean isAddress(Postagger postagger, Segmentor segmentor,String content){
    		List<String> words = new ArrayList<String>();
    		segmentor.segment(content, words);
    		List<String> postags = new ArrayList<String>();
    		int size = postagger.postag(words,postags);
    		boolean isAddress=false;
    		for (int i = 0; i < size; i++) {
    			if (postags.get(i).equals(ADDRESS_C)||postags.get(i).equals(ADDRESS_L)){
    				isAddress=true;
    				break;
    			}
    		}
    		return isAddress;
    	}
    
    	/**
    	 * 判断一句话中是否存在企业团体，先使用segmentor将内容分词，使用postagger判断是否为企业团体
    	 * @param postagger
    	 * @param segmentor
    	 * @param content 内容
    	 * @return
    	 */
    	public static Boolean isCorporateGroups(Postagger postagger, Segmentor segmentor,NER ner,String content){
    		Boolean isGroups=false;
    		List<String> words = new ArrayList<String>();
    		List<String> tags = new ArrayList<String>();
    		List<String> ners = new ArrayList<String>();
    		segmentor.segment(content, words);
    		List<String> postags = new ArrayList<String>();
    		int size = postagger.postag(words,postags);
    		for (int i = 0; i < size; i++) {
    			tags.add(postags.get(i));
    		}
    		ner.recognize(words, tags, ners);
    		for (int i = 0; i < words.size(); i++) {
    			System.out.println(words.get(i)+"/"+ners.get(i));
    		}
    		if (ners.contains("E-Ni")){
    			isGroups=true;
    			return isGroups;
    		}
    		return isGroups;
    	}
    
    	/**
    	 * 判断一句话中是否存在公检法，先使用segmentor将内容分词，使用postagger判断是否为公检法
    	 * @param postagger
    	 * @param segmentor
    	 * @param content 内容
    	 * @return
    	 */
    	public static Boolean isAbbreviation(Postagger postagger, Segmentor segmentor,String content){
    //		List<String> words = new ArrayList<String>();
    //		segmentor.segment(content, words);
    		List<String> list = new ArrayList<>();
    		list.add(content);
    		List<String> postags = new ArrayList<String>();
    		int size = postagger.postag(list,postags);
    		boolean isGovernment=false;
    		for (int i = 0; i < size; i++) {
    			if (postags.get(i).equals(GOVERNMENT)){
    				isGovernment=true;
    				break;
    			}
    		}
    		return isGovernment;
    	}
    
    	/**
    	 * 判断一句话中是否存在人名，先使用segmentor将内容分词，使用postagger判断是否为人名
    	 * @param postagger
    	 * @param segmentor
    	 * @param content 内容
    	 * @return
    	 */
    	public static Boolean isPersonName(Postagger postagger, Segmentor segmentor,String content){
    		List<String> words = new ArrayList<String>();
    		segmentor.segment(content, words);
    		List<String> postags = new ArrayList<String>();
    		int size = postagger.postag(words,postags);
    		boolean isPersonName=false;
    		for (int i = 0; i < size; i++) {
    			if (postags.get(i).equals(PERSONNAME_C)){
    				isPersonName=true;
    				break;
    			}
    		}
    		return isPersonName;
    	}
    
    	/**
    	 * 识别地名,返回地名所占比例
    	 * @param words 字段内容
    	 * @return
    	 */
    	public static double identificationAddress( Postagger postagger,Segmentor segmentor,List<String> words){
    		int n=0;
    		for (int i = 0; i < words.size(); i++) {
    			String s = words.get(i);
    			Boolean address = isAddress(postagger, segmentor, s);
    			if (address){
    				n+=1;
    			}
    		}
    		double proportion = new BigDecimal((float)n/words.size()).setScale(2, BigDecimal.ROUND_HALF_UP).doubleValue();
    		return proportion;
    	}
    	/**
    	 * 识别人名,返回人名所占比例
    	 * @param words 字段内容
    	 * @return
    	 */
    	public static double identificationPersonName( Postagger postagger,Segmentor segmentor,List<String> words){
    		int n=0;
    		for (int i = 0; i < words.size(); i++) {
    			String s = words.get(i);
    			Boolean address = isPersonName(postagger, segmentor, s);
    			if (address){
    				n+=1;
    			}
    		}
    		double proportion = new BigDecimal((float)n/words.size()).setScale(2, BigDecimal.ROUND_HALF_UP).doubleValue();
    		return proportion;
    	}
    
    	/**
    	 * 识别企业团体,返回企业团体所占比例
    	 * @param words 字段内容
    	 * @return
    	 */
    	public static double identificationGroups( Postagger postagger,Segmentor segmentor,NER ner,List<String> words){
    		int n=0;
    		for (int i = 0; i < words.size(); i++) {
    			String s = words.get(i);
    			Boolean groups = isCorporateGroups(postagger, segmentor, ner,s);
    			if (groups){
    				n+=1;
    			}
    		}
    		double proportion = new BigDecimal((float)n/words.size()).setScale(2, BigDecimal.ROUND_HALF_UP).doubleValue();
    		return proportion;
    	}
    	/**
    	 * 识别公检法,返回公检法所占比例
    	 * @param words 字段内容
    	 * @return
    	 */
    	public static double identificationGovernment( Postagger postagger,Segmentor segmentor,List<String> words){
    		int n=0;
    		for (int i = 0; i < words.size(); i++) {
    			String s = words.get(i);
    			Boolean groups = isAbbreviation(postagger, segmentor, s);
    			if (groups){
    				n+=1;
    			}
    		}
    		double proportion = new BigDecimal((float)n/words.size()).setScale(2, BigDecimal.ROUND_HALF_UP).doubleValue();
    		return proportion;
    	}
    
    	public static void main(String[] args) {
    		/*System.load("D:\\test\\ltp\\ltp4j-0.1.0-SNAPSHOT.dll");
    		Postagger postagger = new Postagger();
    		Segmentor segmentor = new Segmentor();
    		if (postagger.create("D:\\test\\ltp\\pos.model") < 0) {
    			log.error("pos.model加载失败");
    		}
    		if (segmentor.create("D:\\test\\ltp\\cws.model") < 0) {
    			log.error("load cws.model failed");
    		}
    		List<String> words = new ArrayList<String>();
    		words.add("上海澜海实业有限公司");
    		words.add("上海丰临进出口有限公司");
    		words.add("上海宝钢国际经济贸易有限公司");
    		words.add("上海诚齐机械制造有限公司");
    		double identificationName = Ltp4jUtil.identificationPersonName(postagger,segmentor,words);
    		double identificationAddress = Ltp4jUtil.identificationAddress(postagger,segmentor,words);
    		double identificationGroups = Ltp4jUtil.identificationGroups(postagger,segmentor,words);
    		double identificationGovernment = Ltp4jUtil.identificationGovernment(postagger,segmentor,words);
    		postagger.release();
    		segmentor.release();
    		System.out.println("地点："+identificationAddress);
    		System.out.println("人名："+identificationName);
    		System.out.println("企业："+identificationGroups);
    		System.out.println("公检法："+identificationGovernment);
    
    */
    		String content="上海丰临进出口有限公司";
    		LoadLibsUtil.addDirToPath("libs");
    		NER ner = new NER();
    		Postagger postagger = new Postagger();
    		Segmentor segmentor = new Segmentor();
    		if (ner.create("D:\\test\\ltp\\ner.model") < 0) {
    			log.error("pos.model加载失败");
    		}
    		if (postagger.create("D:\\test\\ltp\\pos.model") < 0) {
    			log.error("pos.model加载失败");
    		}
    		if (segmentor.create("D:\\test\\ltp\\cws.model") < 0) {
    			log.error("load cws.model failed");
    		}
    
    		Boolean s=isCorporateGroups(postagger,segmentor,ner,content);
    		System.out.println(s);
    
    
    		postagger.release();
    		segmentor.release();
    		ner.release();
    
    	}
    
    }

package com.dbapp.database.scanning.util;
    
    import java.lang.reflect.Field;
    
    /**
     * @ClassName: LoadLibsUtil
     * @Author: yongtao.ding on 2020/5/26 9:09
     */
    public class LoadLibsUtil {
    
    	public static void addDirToPath(String s){
    		try {
    			//获取系统path变量对象
    			Field field=ClassLoader.class.getDeclaredField("sys_paths");
    			//设置此变量对象可访问
    			field.setAccessible(true);
    			//获取此变量对象的值
    			String[] path=(String[])field.get(null);
    			//创建字符串数组，在原来的数组长度上增加一个，用于存放增加的目录
    			String[] tem=new String[path.length+1];
    			//将原来的path变量复制到tem中
    			System.arraycopy(path,0,tem,0,path.length);
    			//将增加的目录存入新的变量数组中
    			tem[path.length]=s;
    			//将增加目录后的数组赋给path变量对象
    			field.set(null,tem);
    		} catch (Exception e) {
    			e.printStackTrace();
    		}
    	}
    }

关于加载dll：  
代码中

LoadLibsUtil.addDirToPath("libs");

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzE2NjEzMzEx_size_16_color_FFFFFF_t_70]  
有问题评论，能不能看到随缘~~~~

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzE2NjEzMzEx_size_16_color_FFFFFF_t_70]: https://img-blog.csdnimg.cn/20200602174249287.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzE2NjEzMzEx,size_16,color_FFFFFF,t_70