org.wltea.analyzer.dic.Dictionary Maven / Gradle / Ivy
The newest version!
package org.wltea.analyzer.dic;
import lombok.extern.slf4j.Slf4j;
import org.wltea.analyzer.cfg.Configuration;
import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.util.Collection;
import java.util.List;
/**
* 词典管理类,单子模式
*/
@Slf4j
public class Dictionary {
/*
* 词典单子实例
*/
private static Dictionary singleton;
/*
* 主词典对象
*/
private DictSegment _MainDict;
/*
* 停止词词典
*/
private DictSegment _StopWordDict;
/*
* 量词词典
*/
private DictSegment _QuantifierDict;
/**
* 配置对象
*/
private final Configuration cfg;
private Dictionary(Configuration cfg) {
this.cfg = cfg;
this.loadMainDict();
this.loadStopWordDict();
this.loadQuantifierDict();
}
/**
* 词典初始化
* 由于IK Analyzer的词典采用Dictionary类的静态方法进行词典初始化
* 只有当Dictionary类被实际调用时，才会开始载入词典，
* 这将延长首次分词操作的时间
* 该方法提供了一个在应用加载阶段就初始化字典的手段
*
* @param cfg cfg
* @return Dictionary
*/
public static Dictionary initial(Configuration cfg) {
if (singleton == null) {
synchronized (Dictionary.class) {
if (singleton == null) {
singleton = new Dictionary(cfg);
return singleton;
}
}
}
return singleton;
}
/**
* 获取词典单子实例
*
* @return Dictionary 单例对象
*/
public static Dictionary getSingleton() {
if (singleton == null) {
throw new IllegalStateException("词典尚未初始化，请先调用initial方法");
}
return singleton;
}
/**
* 批量加载新词条
*
* @param words 词条列表
*/
public void addWords(Collection words) {
if (words != null) {
//批量加载词条到主内存词典中
words.stream().filter(word -> word != null).forEach(word -> {
//批量加载词条到主内存词典中
singleton._MainDict.fillSegment(word.trim().toLowerCase().toCharArray());
});
}
}
/**
* 批量移除（屏蔽）词条
*
* @param words words
*/
public void disableWords(Collection words) {
if (words != null) {
//批量屏蔽词条
words.stream().filter(word -> word != null).forEach(word -> {
//批量屏蔽词条
singleton._MainDict.disableSegment(word.trim().toLowerCase().toCharArray());
});
}
}
/**
* 检索匹配主词典
*
* @param charArray charArray
* @return Hit 匹配结果描述
*/
public Hit matchInMainDict(char[] charArray) {
return singleton._MainDict.match(charArray);
}
/**
* 检索匹配主词典
*
* @param charArray charArray
* @param begin begin
* @param length length
* @return Hit 匹配结果描述
*/
public Hit matchInMainDict(char[] charArray, int begin, int length) {
return singleton._MainDict.match(charArray, begin, length);
}
/**
* 检索匹配量词词典
*
* @param charArray charArray
* @param begin begin
* @param length length
* @return Hit 匹配结果描述
*/
public Hit matchInQuantifierDict(char[] charArray, int begin, int length) {
return singleton._QuantifierDict.match(charArray, begin, length);
}
/**
* 从已匹配的Hit中直接取出DictSegment，继续向下匹配
*
* @param charArray charArray
* @param currentIndex currentIndex
* @param matchedHit matchedHit
* @return Hit
*/
public Hit matchWithHit(char[] charArray, int currentIndex, Hit matchedHit) {
DictSegment ds = matchedHit.getMatchedDictSegment();
return ds.match(charArray, currentIndex, 1, matchedHit);
}
/**
* 判断是否是停止词
*
* @param charArray charArray
* @param begin begin
* @param length length
* @return boolean
*/
public boolean isStopWord(char[] charArray, int begin, int length) {
return singleton._StopWordDict.match(charArray, begin, length).isMatch();
}
/**
* 加载主词典及扩展词典
*/
private void loadMainDict() {
//建立一个主词典实例
_MainDict = new DictSegment((char) 0);
//读取主词典文件
InputStream is = this.getClass().getClassLoader().getResourceAsStream(cfg.getMainDictionary());
if (is == null) {
throw new RuntimeException("Main Dictionary not found!!!");
}
try {
BufferedReader br = new BufferedReader(new InputStreamReader(is, "UTF-8"), 512);
String theWord;
do {
theWord = br.readLine();
if (theWord != null && !"".equals(theWord.trim())) {
_MainDict.fillSegment(theWord.trim().toLowerCase().toCharArray());
}
} while (theWord != null);
} catch (IOException ioe) {
log.error("Main Dictionary loading exception.", ioe);
} finally {
try {
if (is != null) {
is.close();
}
} catch (IOException e) {
log.error(e.getMessage(), e);
}
}
//加载扩展词典
this.loadExtDict();
}
/**
* 加载用户配置的扩展词典到主词库表
*/
private void loadExtDict() {
//加载扩展词典配置
List extDictFiles = cfg.getExtDictionarys();
if (extDictFiles != null) {
InputStream is;
for (String extDictName : extDictFiles) {
//读取扩展词典文件
if (log.isDebugEnabled()) {
log.debug("加载扩展词典：" + extDictName);
}
is = this.getClass().getClassLoader().getResourceAsStream(extDictName);
//如果找不到扩展的字典，则忽略
if (is == null) {
continue;
}
try {
BufferedReader br = new BufferedReader(new InputStreamReader(is, "UTF-8"), 512);
String theWord;
do {
theWord = br.readLine();
if (theWord != null && !"".equals(theWord.trim())) {
//加载扩展词典数据到主内存词典中
_MainDict.fillSegment(theWord.trim().toLowerCase().toCharArray());
}
} while (theWord != null);
} catch (IOException ioe) {
log.error("Extension Dictionary loading exception.", ioe);
} finally {
try {
if (is != null) {
is.close();
}
} catch (IOException e) {
log.error(e.getMessage(), e);
}
}
}
}
}
/**
* 加载用户扩展的停止词词典
*/
private void loadStopWordDict() {
//建立一个主词典实例
_StopWordDict = new DictSegment((char) 0);
//加载扩展停止词典
List extStopWordDictFiles = cfg.getExtStopWordDictionarys();
if (extStopWordDictFiles != null) {
InputStream is;
for (String extStopWordDictName : extStopWordDictFiles) {
if (log.isDebugEnabled()) {
log.debug("加载扩展停止词典：" + extStopWordDictName);
}
//读取扩展词典文件
is = this.getClass().getClassLoader().getResourceAsStream(extStopWordDictName);
//如果找不到扩展的字典，则忽略
if (is == null) {
continue;
}
try {
BufferedReader br = new BufferedReader(new InputStreamReader(is, "UTF-8"), 512);
String theWord;
do {
theWord = br.readLine();
if (theWord != null && !"".equals(theWord.trim())) {
//加载扩展停止词典数据到内存中
_StopWordDict.fillSegment(theWord.trim().toLowerCase().toCharArray());
}
} while (theWord != null);
} catch (IOException ioe) {
log.error("Extension Stop word Dictionary loading exception.", ioe);
} finally {
try {
if (is != null) {
is.close();
}
} catch (IOException e) {
log.error(e.getMessage(), e);
}
}
}
}
}
/**
* 加载量词词典
*/
private void loadQuantifierDict() {
//建立一个量词典实例
_QuantifierDict = new DictSegment((char) 0);
//读取量词词典文件
InputStream is = this.getClass().getClassLoader().getResourceAsStream(cfg.getQuantifierDicionary());
if (is == null) {
throw new RuntimeException("Quantifier Dictionary not found!!!");
}
try {
BufferedReader br = new BufferedReader(new InputStreamReader(is, "UTF-8"), 512);
String theWord;
do {
theWord = br.readLine();
if (theWord != null && !"".equals(theWord.trim())) {
_QuantifierDict.fillSegment(theWord.trim().toLowerCase().toCharArray());
}
} while (theWord != null);
} catch (IOException ioe) {
log.error("Quantifier Dictionary loading exception.", ioe);
} finally {
try {
if (is != null) {
is.close();
}
} catch (IOException e) {
log.error(e.getMessage(), e);
}
}
}
}