solr 自定义分词器

今天需要将一个以逗号分隔的字段建立到索引库中去，没找到有现成的逗号分隔符分词器，于是看了看源码里空格分词器WhitespaceTokenizerFactory的写法。照葫芦画瓢写了一个逗号分词器：

package com.besttone.analyzer;

import java.io.Reader;
import java.util.Map;

import org.apache.solr.analysis.BaseTokenizerFactory;

public class CommaTokenizerFactory extends BaseTokenizerFactory {

	@Override
	public void init(Map<String, String> args) {
		super.init(args);
		assureMatchVersion();
	}

	public CommaTokenizer create(Reader input) {
		return new CommaTokenizer(luceneMatchVersion, input);
	}

}

package com.besttone.analyzer;

import java.io.Reader;

import org.apache.lucene.analysis.CharTokenizer;
import org.apache.lucene.util.AttributeSource;
import org.apache.lucene.util.Version;

public class CommaTokenizer extends CharTokenizer {

	/**
	 * Construct a new WhitespaceTokenizer. * @param matchVersion Lucene version
	 * to match See {@link <a href="#version">above</a>}
	 * 
	 * @param in
	 *            the input to split up into tokens
	 */
	public CommaTokenizer(Version matchVersion, Reader in) {
		super(matchVersion, in);
	}

	/**
	 * Construct a new WhitespaceTokenizer using a given {@link AttributeSource}
	 * .
	 * 
	 * @param matchVersion
	 *            Lucene version to match See
	 *            {@link <a href="#version">above</a>}
	 * @param source
	 *            the attribute source to use for this {@link Tokenizer}
	 * @param in
	 *            the input to split up into tokens
	 */
	public CommaTokenizer(Version matchVersion, AttributeSource source,
			Reader in) {
		super(matchVersion, source, in);
	}

	/**
	 * Construct a new WhitespaceTokenizer using a given
	 * {@link org.apache.lucene.util.AttributeSource.AttributeFactory}.
	 * 
	 * @param matchVersion
	 *            Lucene version to match See
	 *            {@link <a href="#version">above</a>}
	 * @param factory
	 *            the attribute factory to use for this {@link Tokenizer}
	 * @param in
	 *            the input to split up into tokens
	 */
	public CommaTokenizer(Version matchVersion, AttributeFactory factory,
			Reader in) {
		super(matchVersion, factory, in);
	}

	/**
	 * Construct a new CommaTokenizer.
	 * 
	 * @deprecated use {@link #CommaTokenizer(Version, Reader)} instead. This
	 *             will be removed in Lucene 4.0.
	 */
	@Deprecated
	public CommaTokenizer(Reader in) {
		super(in);
	}

	/**
	 * Construct a new CommaTokenizer using a given {@link AttributeSource}.
	 * 
	 * @deprecated use {@link #CommaTokenizer(Version, AttributeSource, Reader)}
	 *             instead. This will be removed in Lucene 4.0.
	 */
	@Deprecated
	public CommaTokenizer(AttributeSource source, Reader in) {
		super(source, in);
	}

	/**
	 * Construct a new CommaTokenizer using a given
	 * {@link org.apache.lucene.util.AttributeSource.AttributeFactory}.
	 * 
	 * @deprecated use
	 *             {@link #CommaTokenizer(Version, AttributeSource.AttributeFactory, Reader)}
	 *             instead. This will be removed in Lucene 4.0.
	 */
	@Deprecated
	public CommaTokenizer(AttributeFactory factory, Reader in) {
		super(factory, in);
	}

	/**
	 * Collects only characters which do not satisfy
	 * {@link Character#isWhitespace(int)}.
	 */
	@Override
	protected boolean isTokenChar(int c) {
		// return !Character.isWhitespace(c);
		// 44表示逗号
		return !(c == 44);
	}

}

其实很简单，只要继承一下solr提供的通用字符分词器：CharTokenizer，然后实现自己的isTokenChar方法：

protected boolean isTokenChar(int c) {
		// return !Character.isWhitespace(c);
		// 44表示逗号
		return !(c == 44);
	}

判断是否等于44，如果等于就返回false,否则返回true。返回false表示分词。44是逗号的asc码值，比如a的asc码值为97,如果不知道一个字符对应的值为多少，可以这样：

char[] c = new char[]{'a',',','b'};

Character.codePointAt(c, 1);

获得char数组里index为1的字符的asc码值。

然后打包成jar,放到solr_home/lib下面，或者其他地方也行，但是要在solrconfig.xml里配置lib的路径或者solr.xml里配置sharelib的路径都行，总之就是要solr启动时加载这个jar包。

然后就可以在solr控制台的analysis页面测试一下分词效果了。

 

    <fieldType name="text_comma" class="solr.TextField" positionIncrementGap="100">
      <analyzer>
        <tokenizer class="com.besttone.analyzer.CommaTokenizerFactory"/>
      </analyzer>
    </fieldType>