Nutch2.x 集成ElasticSearch 抓取+索引

javascript:void(0)

 

使用https://github.com/apache/nutch.git导入nutch项目到intellij

 

配置ivy.xml和conf下的gora.properties、nutch-site.xml

修改ivy/ivy.xml

     修改elasticsearch版本

[html] view plaincopy

 

1. <dependency org="org.elasticsearch" name="elasticsearch" rev="0.90.5" conf="*->default"/>  

     去掉如下内容注解

[html] view plaincopy

 

1. <dependency org="org.apache.gora" name="gora-hbase" rev="0.3" conf="*->default" />  

     修改软件版本，从1.2.15改成1.2.16，解决部分包导入失败问题

[html] view plaincopy

 

1. <dependency org="log4j" name="log4j" rev="1.2.16" conf="*->master" />  

修改gora.properties

     注掉如下几行

[plain] view plaincopy

 

1. #gora.sqlstore.jdbc.driver=org.hsqldb.jdbc.JDBCDriver  
2. #gora.sqlstore.jdbc.url=jdbc:hsqldb:hsql://localhost/nutchtest  
3. #gora.sqlstore.jdbc.user=sa  
4. #gora.sqlstore.jdbc.password=  

     添加一行

[html] view plaincopy

 

1. gora.datastore.default=org.apache.gora.hbase.store.HBaseStore  

 

修改nutch-site.xml，增加如下配置项

[html] view plaincopy

 

1. <property>  
2.     <name>storage.data.store.class</name>  
3.     <value>org.apache.gora.hbase.store.HBaseStore</value>  
4. </property>  
5. <property>  
6.     <name>http.agent.name</name>  
7.     <value>NutchCrawler</value>  
8. </property>  
9. <property>  
10.     <name>parser.character.encoding.default</name>  
11.     <value>utf-8</value>  
12. </property>  
13. <property>  
14.     <name>http.accept.language</name>  
15.     <value>ja-jp, en-us, zh-cn,en-gb,en;q=0.7,*;q=0.3</value>  
16. </property>  
17. <property>    
18.     <name>generate.batch.id</name>    
19.     <value>1</value>    
20. </property>  

 

增加hbase配置文件hbase-site.xml到nutch/conf中  

[html] view plaincopy

 

1. <configuration>  
2.     <property>  
3.         <name>hbase.rootdir</name>  
4.         <value>file:///data/hbase</value>  
5.     </property>  
6.     <property>  
7.         <name>hbase.zookeeper.property.dataDir</name>  
8.         <value>/data/zookeeper</value>  
9.     </property>  
10. </configuration>     

 

修改nutch/src/bin/nutch，文件开头增加

      NUTCH_JAVA_HOME=/usr/local/jdk

 

修改src下org.apache.nutch.indexer.elastic.ElasticWriter 109行，使支持es0.90.5

      item.isFailed()

 

删除nutch/conf下所有template文件

 

编译nutch

      ant clean 

      ant runtime

 

修改nutch-site.xml

[html] view plaincopy

 

1. <property>  
2.     <name>plugin.folders</name>  
3.     <value>/home/eryk/workspace/nutch/runtime/local/plugins</value>  
4. </property>  

设置intelil，增加nutch/conf和nutch/runtime/lib到classpath

File->Project Structure->Dependencies 增加nutch/conf和nutch/runtime/local/lib目录

 

增加pom.xml的依赖库

[html] view plaincopy

 

1. <dependency>  
2.     <groupId>net.sourceforge.nekohtml</groupId>  
3.     <artifactId>nekohtml</artifactId>  
4.     <version>1.9.15</version>  
5. </dependency>  
6. <dependency>  
7.     <groupId>org.ccil.cowan.tagsoup</groupId>  
8.     <artifactId>tagsoup</artifactId>  
9.     <version>1.2</version>  
10. </dependency>  
11. <dependency>  
12.     <groupId>rome</groupId>  
13.     <artifactId>rome</artifactId>  
14.     <version>1.0</version>  
15. </dependency>  

 

修改pom.xml中es版本

[html] view plaincopy

 

1. <dependency>  
2.      <groupId>org.elasticsearch</groupId>  
3.      <artifactId>elasticsearch</artifactId>  
4.      <version>0.90.5</version>  
5.      <optional>true</optional>  
6. </dependency>  

 

修正依赖库的版本冲突

[html] view plaincopy

 

1. <dependency>  
2.     <groupId>org.restlet.jse</groupId>  
3.     <artifactId>org.restlet.ext.jackson</artifactId>  
4.     <version>2.0.5</version>  
5.                <exclusions>  
6.                    <exclusion>  
7.                        <artifactId>jackson-core-asl</artifactId>  
8.                        <groupId>org.codehaus.jackson</groupId>  
9.                    </exclusion>  
10.                    <exclusion>  
11.                        <artifactId>jackson-mapper-asl</artifactId>  
12.                        <groupId>org.codehaus.jackson</groupId>  
13.                    </exclusion>  
14.                </exclusions>  
15.                <optional>true</optional>  
16.  </dependency>  
17.      <dependency>  
18.                    <groupId>org.apache.gora</groupId>  
19.                    <artifactId>gora-core</artifactId>  
20.                    <version>0.3</version>  
21.                <exclusions>  
22.                    <exclusion>  
23.                        <artifactId>jackson-mapper-asl</artifactId>  
24.                        <groupId>org.codehaus.jackson</groupId>  
25.                    </exclusion>  
26.                </exclusions>  
27.                <optional>true</optional>  
28.            </dependency>  

修改src下org.apache.nutch.crawl.Crawler代码，增加-elasticindex和-batchId参数

 

[html] view plaincopy

 

1.  Map<String,Object> argMap = ToolUtil.toArgMap(  
2.     Nutch.ARG_THREADS, threads,  
3.     Nutch.ARG_DEPTH, depth,  
4.     Nutch.ARG_TOPN, topN,  
5.     Nutch.ARG_SOLR, solrUrl,  
6.     ElasticConstants.CLUSTER,elasticSearchAddr,      //使用es建立索引  
7.     Nutch.ARG_SEEDDIR, seedDir,  
8.     Nutch.ARG_NUMTASKS, numTasks,  
9.     Nutch.ARG_BATCH,batchId,      //解决NullPointerException问题  
10.     GeneratorJob.BATCH_ID,batchId);       //解决NullPointerException问题，貌似没用  
11. run(argMap);  

 

修改org.apache.nutch.indexer.elastic.ElasticWriter代码，支持-elasticindex ip:port传参

[html] view plaincopy

 

1. public void open(TaskAttemptContext job) throws IOException {  
2.     String clusterName = job.getConfiguration().get(ElasticConstants.CLUSTER);  
3.     if (clusterName != null && !clusterName.contains(":")) {  
4.       node = nodeBuilder().clusterName(clusterName).client(true).node();  
5.     } else {  
6.       node = nodeBuilder().client(true).node();  
7.     }  
8.     LOG.info(String.format("clusterName=[%s]",clusterName));  
9.   
10.     if(clusterName.contains(":")){  
11.         String[] addr = clusterName.split(":");  
12.         client = new TransportClient()  
13.                 .addTransportAddress(new InetSocketTransportAddress(addr[0],Integer.parseInt(addr[1])));  
14.   
15.     }else{  
16.         client = node.client();  
17.     }  
18.   
19.     bulk = client.prepareBulk();  
20.     defaultIndex = job.getConfiguration().get(ElasticConstants.INDEX, "index");  
21.     maxBulkDocs = job.getConfiguration().getInt(  
22.         ElasticConstants.MAX_BULK_DOCS, DEFAULT_MAX_BULK_DOCS);  
23.     maxBulkLength = job.getConfiguration().getInt(  
24.         ElasticConstants.MAX_BULK_LENGTH, DEFAULT_MAX_BULK_LENGTH);  
25.   }  

 

在nutch目录下增加urls目录，在url目录下新建seed.txt，写入要爬的种子地址

 

运行Crawler

     传入参数

     urls -elasticindex a2:9300 -threads 10 -depth 3 -topN 5 -batchId 1

     观察nutch/hadoop.log日志

[html] view plaincopy

 

1.            2013-11-03 22:57:36,682 INFO  elasticsearch.node - [Ikonn] started  
2. 2013-11-03 22:57:36,682 INFO  elastic.ElasticWriter - clusterName=[a2:9300]  
3. 2013-11-03 22:57:36,692 INFO  elasticsearch.plugins - [Electron] loaded [], sites []  
4. 2013-11-03 22:57:36,863 INFO  basic.BasicIndexingFilter - Maximum title length for indexing set to: 100  
5. 2013-11-03 22:57:36,864 INFO  indexer.IndexingFilters - Adding org.apache.nutch.indexer.basic.BasicIndexingFilter  
6. 2013-11-03 22:57:36,864 INFO  anchor.AnchorIndexingFilter - Anchor deduplication is: off  
7. 2013-11-03 22:57:36,865 INFO  indexer.IndexingFilters - Adding org.apache.nutch.indexer.anchor.AnchorIndexingFilter  
8. 2013-11-03 22:57:37,946 INFO  elastic.ElasticWriter - Processing remaining requests [docs = 86, length = 130314, total docs = 86]  
9. 2013-11-03 22:57:37,988 INFO  elastic.ElasticWriter - Processing to finalize last execute  
10. 2013-11-03 22:57:41,986 INFO  elastic.ElasticWriter - Previous took in ms 1590, including wait 3998  
11. 2013-11-03 22:57:42,020 INFO  elasticsearch.node - [Ikonn] stopping ...  
12. 2013-11-03 22:57:42,032 INFO  elasticsearch.node - [Ikonn] stopped  
13. 2013-11-03 22:57:42,032 INFO  elasticsearch.node - [Ikonn] closing ...  
14. 2013-11-03 22:57:42,039 INFO  elasticsearch.node - [Ikonn] closed  
15. 2013-11-03 22:57:42,041 WARN  mapred.FileOutputCommitter - Output path is null in cleanup  
16. 2013-11-03 22:57:42,057 INFO  elastic.ElasticIndexerJob - Done  

查询es

     http://a2:9200/_search?q=%E7%BE%8E%E5%A5%B3&pretty=true

     返回结果，说明已经跑通了，观察hbase中，表已经自动建好，并存入了已经爬到的数据

     

 

参考

http://www.blogjava.net/paulwong/archive/2013/08/31/403513.html

javascript:void(0)

http://www.searchtech.pro/nutch2.1-elasticsearch-mysql-local-Integrate

javascript:void(0)