http://qindongliang.iteye.com/blog/2164583
Lucene内置很多的分词器工具包，几乎涵盖了全球所有的国家和地区，最近散仙，在搞多语言分词的一个处理，主要国家有西班牙，葡萄牙，德语，法语，意大利，其实这些语系都与英语非常类似，都是以空格为分割的语种。
那么首先，探讨下分词器的词形还原和词干提取的对搜索的意义？在这之前，先看下两者的概念：
词形还原（lemmatization），是把一个任何形式的语言词汇还原为一般形式（能表达完整语义），而词干提取
（stemming）是抽取词的词干或词根形式（不一定能够表达完整语义）。词形还原和词干提取是词形规范化的两类
重要方式，都能够达到有效归并词形的目的，二者既有联系也有区别
详细介绍，请参考这篇文章
在电商搜索里，词干的抽取，和单复数的还原比较重要（这里主要针对名词来讲），因为这有关搜索的查准率，和查全率的命中，如果我们的分词器没有对这些词做过处理，会造成什么影响呢？那么请看如下的一个例子？
句子： i have two cats
分词器如果什么都没有做：
这时候我们搜cat，就会无命中结果，而必须搜cats才能命中到一条数据，而事实上cat和cats是同一个东西，只不过单词的形式不一样，这样以来，如果不做处理，我们的查全率和查全率都会下降，会涉及影响到我们的搜索体验，所以stemming这一步，在某些场合的分词中至关重要。
本篇，散仙，会参考源码分析一下，关于德语分词中中如何做的词干提取，先看下德语的分词声明：
Listlist=new ArrayList (); list.add("player");//这里面的词，不会被做词干抽取，词形还原 CharArraySet ar=new CharArraySet(Version.LUCENE_43,list , true); //分词器的第二个参数是禁用词参数，第三个参数是排除不做词形转换，或单复数的词 GermanAnalyzer sa=new GermanAnalyzer(Version.LUCENE_43,null,ar);
接着，我们具体看下，在德语的分词器中，都经过了哪几部分的过滤处理：
protected TokenStreamComponents createComponents(String fieldName,
Reader reader) {
//标准分词器过滤
final Tokenizer source = new StandardTokenizer(matchVersion, reader);
TokenStream result = new StandardFilter(matchVersion, source);
//转小写过滤
result = new LowerCaseFilter(matchVersion, result);
//禁用词过滤
result = new StopFilter( matchVersion, result, stopwords);
//排除词过滤
result = new SetKeywordMarkerFilter(result, exclusionSet);
if (matchVersion.onOrAfter(Version.LUCENE_36)) {
//在lucene3.6以后的版本，采用如下filter过滤
//规格化，将德语中的特殊字符，映射成英语
result = new GermanNormalizationFilter(result);
//stem词干抽取，词性还原
result = new GermanLightStemFilter(result);
} else if (matchVersion.onOrAfter(Version.LUCENE_31)) {
//在lucene3.1至3.6的版本中，采用SnowballFilter处理
result = new SnowballFilter(result, new German2Stemmer());
} else {
//在lucene3.1之前的采用兼容的GermanStemFilter处理
result = new GermanStemFilter(result);
}
return new TokenStreamComponents(source, result);
}
OK，我们从源码中得知，在Lucene4.x中对德语的分词也做了向前和向后兼容，现在我们主要关注在lucene4.x之后的版本如何的词形转换，下面分别看下
result = new GermanNormalizationFilter(result);
result = new GermanLightStemFilter(result);
这两个类的功能：
package org.apache.lucene.analysis.de; /* * Licensed to the Apache Software Foundation (ASF) under one or more * contributor license agreements. See the NOTICE file distributed with * this work for additional information regarding copyright ownership. * The ASF licenses this file to You under the Apache License, Version 2.0 * (the "License"); you may not use this file except in compliance with * the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ import java.io.IOException; import org.apache.lucene.analysis.TokenFilter; import org.apache.lucene.analysis.TokenStream; import org.apache.lucene.analysis.tokenattributes.CharTermAttribute; import org.apache.lucene.analysis.util.StemmerUtil; /** * Normalizes German characters according to the heuristics * of the * German2 snowball algorithm. * It allows for the fact that ä, ö and ü are sometimes written as ae, oe and ue. * * [list] *
* This is useful if you want this normalization without using
* the German2 stemmer, or perhaps no stemming at all.
*上面的解释说得很清楚，主要是对德文的一些特殊字母，转换成对应的英文处理
*
*/
public final class GermanNormalizationFilter extends TokenFilter {
// FSM with 3 states:
private static final int N = 0; /* ordinary state */
private static final int V = 1; /* stops 'u' from entering umlaut state */
private static final int U = 2; /* umlaut state, allows e-deletion */
private final CharTermAttribute termAtt = addAttribute(CharTermAttribute.class);
public GermanNormalizationFilter(TokenStream input) {
super(input);
}
@Override
public boolean incrementToken() throws IOException {
if (input.incrementToken()) {
int state = N;
char buffer[] = termAtt.buffer();
int length = termAtt.length();
for (int i = 0; i < length; i++) {
final char c = buffer[i];
switch(c) {
case 'a':
case 'o':
state = U;
break;
case 'u':
state = (state == N) ? U : V;
break;
case 'e':
if (state == U)
length = StemmerUtil.delete(buffer, i--, length);
state = V;
break;
case 'i':
case 'q':
case 'y':
state = V;
break;
case 'ä':
buffer[i] = 'a';
state = V;
break;
case 'ö':
buffer[i] = 'o';
state = V;
break;
case 'ü':
buffer[i] = 'u';
state = V;
break;
case 'ß':
buffer[i++] = 's';
buffer = termAtt.resizeBuffer(1+length);
if (i < length)
System.arraycopy(buffer, i, buffer, i+1, (length-i));
buffer[i] = 's';
length++;
state = N;
break;
default:
state = N;
}
}
termAtt.setLength(length);
return true;
} else {
return false;
}
}
}
package org.apache.lucene.analysis.de;
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
import java.io.IOException;
import org.apache.lucene.analysis.TokenFilter;
import org.apache.lucene.analysis.TokenStream;
import org.apache.lucene.analysis.miscellaneous.SetKeywordMarkerFilter;
import org.apache.lucene.analysis.tokenattributes.CharTermAttribute;
import org.apache.lucene.analysis.tokenattributes.KeywordAttribute;
/**
* A {@link TokenFilter} that applies {@link GermanLightStemmer} to stem German
* words.
*
* To prevent terms from being stemmed use an instance of
* {@link SetKeywordMarkerFilter} or a custom {@link TokenFilter} that sets
* the {@link KeywordAttribute} before this {@link TokenStream}.
*
*
*
*这个类，主要做Stemmer（词干提取），而我们主要关注
*GermanLightStemmer这个类的作用
*
*
*/
public final class GermanLightStemFilter extends TokenFilter {
private final GermanLightStemmer stemmer = new GermanLightStemmer();
private final CharTermAttribute termAtt = addAttribute(CharTermAttribute.class);
private final KeywordAttribute keywordAttr = addAttribute(KeywordAttribute.class);
public GermanLightStemFilter(TokenStream input) {
super(input);
}
@Override
public boolean incrementToken() throws IOException {
if (input.incrementToken()) {
if (!keywordAttr.isKeyword()) {
final int newlen = stemmer.stem(termAtt.buffer(), termAtt.length());
termAtt.setLength(newlen);
}
return true;
} else {
return false;
}
}
}
下面看下，在GermanLightStemmer中，如何做的词干提取：源码如下：
package org.apache.lucene.analysis.de; /* * Licensed to the Apache Software Foundation (ASF) under one or more * contributor license agreements. See the NOTICE file distributed with * this work for additional information regarding copyright ownership. * The ASF licenses this file to You under the Apache License, Version 2.0 * (the "License"); you may not use this file except in compliance with * the License. You may obtain a copy of the License at * * http://www.apache.org/licenses/LICENSE-2.0 * * Unless required by applicable law or agreed to in writing, software * distributed under the License is distributed on an "AS IS" BASIS, * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. * See the License for the specific language governing permissions and * limitations under the License. */ /* * This algorithm is updated based on code located at: * http://members.unine.ch/jacques.savoy/clef/ * * Full copyright for that code follows: */ /* * Copyright (c) 2005, Jacques Savoy * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are met: * * Redistributions of source code must retain the above copyright notice, this * list of conditions and the following disclaimer. Redistributions in binary * form must reproduce the above copyright notice, this list of conditions and * the following disclaimer in the documentation and/or other materials * provided with the distribution. Neither the name of the author nor the names * of its contributors may be used to endorse or promote products derived from * this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. */ /** * Light Stemmer for German. ** This stemmer implements the "UniNE" algorithm in: * Light Stemming Approaches for the French, Portuguese, German and Hungarian Languages * Jacques Savoy */ public class GermanLightStemmer { //处理特殊字符映射 public int stem(char s[], int len) { for (int i = 0; i < len; i++) switch(s[i]) { case 'ä': case 'à': case 'á': case 'â': s[i] = 'a'; break; case 'ö': case 'ò': case 'ó': case 'ô': s[i] = 'o'; break; case 'ï': case 'ì': case 'í': case 'î': s[i] = 'i'; break; case 'ü': case 'ù': case 'ú': case 'û': s[i] = 'u'; break; } len = step1(s, len); return step2(s, len); } private boolean stEnding(char ch) { switch(ch) { case 'b': case 'd': case 'f': case 'g': case 'h': case 'k': case 'l': case 'm': case 'n': case 't': return true; default: return false; } } //处理基于以下规则的词干抽取和缩减 private int step1(char s[], int len) { if (len > 5 && s[len-3] == 'e' && s[len-2] == 'r' && s[len-1] == 'n') return len - 3; if (len > 4 && s[len-2] == 'e') switch(s[len-1]) { case 'm': case 'n': case 'r': case 's': return len - 2; } if (len > 3 && s[len-1] == 'e') return len - 1; if (len > 3 && s[len-1] == 's' && stEnding(s[len-2])) return len - 1; return len; } //处理基于以下规则est，er，en等的词干抽取和缩减 private int step2(char s[], int len) { if (len > 5 && s[len-3] == 'e' && s[len-2] == 's' && s[len-1] == 't') return len - 3; if (len > 4 && s[len-2] == 'e' && (s[len-1] == 'r' || s[len-1] == 'n')) return len - 2; if (len > 4 && s[len-2] == 's' && s[len-1] == 't' && stEnding(s[len-3])) return len - 2; return len; } }
具体的分析结果如下：
搜索技术交流群：324714439 大数据hadoop交流群：376932160 0，将一些德语特殊字符，替换成对应的英文表示 1，将所有词干元音还原 a ，o，i，u ste（2）（按先后顺序，符合以下任意一项，就完成一次校验（return）） 2，单词长度大于5的词，以ern结尾的，直接去掉 3，单词长度大于4的词，以em，en，es，er结尾的，直接去掉 4，单词长度大于3的词，以e结尾的直接去掉 5，单词长度大于3的词，以bs，ds，fs，gs，hs，ks，ls，ms，ns，ts结尾的，直接去掉s step（3）（按先后顺序，符合以下任意一项，就完成一次校验（return）） 6，单词长度大于5的词，以est结尾的，直接去掉 7，单词长度大于4的词，以er或en结尾的直接去掉 8，单词长度大于4的词，bst，dst，fst，gst，hst，kst，lst，mst，nst，tst，直接去掉后两位字母st
最后，结合网上资料分析，基于er，en，e，s结尾的是做单复数转换的，其他的几条规则主要是对非名词的单词，做词干抽取。
原创不易，转载请务必注明，原创地址，谢谢配合!
http://qindongliang.iteye.com/blog/2164583