I already integrated carrot2 with solr-4.x with my customerized chinese tokenizer successfully.
But I run some errors following my series of blogs http://ylzhj02.iyunv.com/blog/2152348 to adopt carrot2 to solr-5.1.0
The error is
org.carrot2.util.factory.FallbackFactory; Tokenizer for Chinese Simplified (zh_cn) is not available. This may degrade clustering quality of Chinese Simplified content. Cause: java.lang.NoSuchMethodError: org.apache.lucene.analysis.Tokenizer.<init>(Ljava/io/Reader;)V
The reason is that solr-5.2.1 adopted lucene 5.1.0, however carrot2-3.10.0 used lucene 4.6.0. So the cause is jars uncompatible.
So, the solution is to download the latest version of carrot2
#git clone git://github.com/carrot2/carrot2.git
(3.11.0)
the lucene version is now 5.1.0
#cd carrot2
step 1:
#vi core/carrot2-util-text/src/org/carrot2/text/linguistic/DefaultTokenizerFactory.java
add
import org.carrot2.text.linguistic.lucene.InokChineseTokenizerAdapter;
change
100 map.put(LanguageCode.CHINESE_SIMPLIFIED,
101 new NewClassInstanceFactory<ITokenizer>(ChineseTokenizerAdapter.class));
to
map.put(LanguageCode.CHINESE_SIMPLIFIED,
new NewClassInstanceFactory<ITokenizer>(InokChineseTokenizerAdapter.class));
step 2:
#vi InokChineseTokenizerAdapter.java
#cp chineseTokenizer/InokChineseTokenizerAdapter.java ./core/carrot2-util-text/src/org/carrot2/text/linguistic/lucene/
step 3:
#mkdir lib/org.lionsoul.jcseg
├── build.properties
├── jcseg-core-1.9.6.jar
├── jcseg.LICENSE
└── META-INF
└── MANIFEST.MF
the file and jars is
build.properties
bin.includes = META-INF/,\
jcseg-core-1.9.6.jar,\
jcseg.LICENSE
META-INF/MANIFEST.MF
Manifest-Version: 1.0
Bundle-ManifestVersion: 2
Bundle-Name: Jcseg Tokenizer
Bundle-SymbolicName: org.lionsoul.jcseg
Bundle-Version: 1.9.6
Bundle-ClassPath: jcseg-core-1.9.6.jar
Bundle-Vendor: INokNok Inc.
Bundle-RequiredExecutionEnvironment: JavaSE-1.6
step 4:
modify build.xml
141 <patternset id="lib.test">
142 <include name="core/**/*.jar" />
143 <include name="lib/**/*.jar" />
144 <include name="lib/org.lionsoul.jcseg/*.jar" />
145 <exclude name="lib/org.slf4j/slf4j-nop*" />
146 <include name="applications/carrot2-dcs/**/*.jar" />
147 <include name="applications/carrot2-webapp/lib/*.jar" />
148 <include name="applications/carrot2-benchmarks/lib/*.jar" />
149 </patternset>
173 <patternset id="lib.core">
174 <include name="lib/**/*.jar" />
175 <include name="lib/org.lionsoul.jcseg/*.jar" />
176 <include name="core/carrot2-util-matrix/lib/*.jar" />
177 <patternset refid="lib.core.excludes" />
178 </patternset>
180 <patternset id="lib.core.mini">
181 <include name="lib/**/mahout-*.jar" />
182 <include name="lib/**/jcseg*.jar" />
183 <include name="lib/**/mahout.LICENSE" />
184 <include name="lib/**/colt.LICENSE" />
185 <include name="lib/**/commons-lang*" />
186 <include name="lib/**/guava*" />
187 <include name="lib/**/jackson*" />
188 <include name="lib/**/lucene-snowball*" />
189 <include name="lib/**/lucene.LICENSE" />
190 <include name="lib/**/hppc-*.jar" />
191 <include name="lib/**/hppc*.LICENSE" />
192
193 <include name="lib/**/slf4j-api*.jar" />
194 <include name="lib/**/slf4j-nop*.jar" />
195 <include name="lib/**/slf4j.LICENSE" />
196
197 <include name="lib/**/attributes-binder-*.jar" />
198 </patternset>
199
906 <target name="core" depends="jar, jar.src, lib-no-jar.flattened" description="Builds Carrot2 Java API JAR with dependencies">
907 <delete dir="${api.dir}" failonerror="false" />
908 <mkdir dir="${api.dir}" />
909 <mkdir dir="${api.dir}/lib" />
910 <mkdir dir="${api.dir}/examples" />
911 <mkdir dir="${api.dir}/resources" />
912
913 <patternset id="carrot2.required">
914 <include name="**/jcseg*" />
915 <include name="**/commons-lang*" />
step 6:
#ant jar
#scp tmp/jar/carrot2-core-3.11.0-SNAPSHOT.jar root@192.168.0.135:/opt/solr/contrib/clustering/lib
carrot2-core-3.11.0-SNAPSHOT.jar
restart solr server to test clustering
-----------------------------
An error happans
org.apache.solr.common.SolrException; null:java.lang.RuntimeException: java.lang.NoClassDefFoundError: com
/carrotsearch/hppc/ObjectHashSet
Solution :
#scp lib/com.carrotsearch.hppc/hppc-0.7.1.jar root@192.168.0.135:/opt/solr/contrib/clustering/lib/
hppc-0.7.1.jar
#rm -f opt/solr/contrib/clustering/lib/hppc-0.5.2.jar
------
another error is
java.lang.RuntimeException: java.lang.IllegalAccessError: class
com.carrotsearch.hppc.ObjectHashSet cannot access its superclass com.carrotsearch.hppc.AbstractObjectCollection
The reason is that there is an old hppc-0.5.2.jar in /opt/solr/server/webapps/solr.war
so, Solution is to
#cd /opt/solr/server/solr-webapp/webapp
#rm -f WEB-INF/lib/hppc-0.5.2.jar
#cp hppc-0.7.1.jar WEB-INF/lib
#jar cf solr.war ./
#mv solr.war /opt/solr/server/webapps
restart solr
the error disappers
运维网声明
1、欢迎大家加入本站运维交流群:群②:261659950 群⑤:202807635 群⑦870801961 群⑧679858003
2、本站所有主题由该帖子作者发表,该帖子作者与运维网 享有帖子相关版权
3、所有作品的著作权均归原作者享有,请您和我们一样尊重他人的著作权等合法权益。如果您对作品感到满意,请购买正版
4、禁止制作、复制、发布和传播具有反动、淫秽、色情、暴力、凶杀等内容的信息,一经发现立即删除。若您因此触犯法律,一切后果自负,我们对此不承担任何责任
5、所有资源均系网友上传或者通过网络收集,我们仅提供一个展示、介绍、观摩学习的平台,我们不对其内容的准确性、可靠性、正当性、安全性、合法性等负责,亦不承担任何法律责任
6、所有作品仅供您个人学习、研究或欣赏,不得用于商业或者其他用途,否则,一切后果均由您自己承担,我们对此不承担任何法律责任
7、如涉及侵犯版权等问题,请您及时通知我们,我们将立即采取措施予以解决
8、联系人Email:admin@iyunv.com 网址:www.yunweiku.com