论坛首页 入门技术论坛

How to make nutch run in eclipse ?

浏览 1729 次
精华帖 (0) :: 良好帖 (0) :: 新手帖 (0) :: 隐藏帖 (0)
作者 正文
   发表时间:2014-05-20  
Nutch是一个优秀的开源的数据爬取框架,我们只需要简单的配置,就可以完成数据爬取,当然,Nutch里面也提供了很灵活的的插件机制,我们随时都可以对它进行二次开发,以满足我们的需求,本篇散仙,先来介绍下,如何在eclipse里面以local模式调试nutch,只有在eclipse里面把它弄清楚了,那么,我们学习起来,才会更加容易,因为,目前大多数人,使用nutch,都是基于命令行的操作,虽然很简单方便,但是想深入定制开发,就很困难,所以,散仙在本篇里,会介绍下nutch基本的调试,以及编译。


下面进入正题,我们先来看下基本的步骤。
<table class="bbcode"><tr><td>序号</td><td>名称</td><td>描述<tr><td>1</td><td>安装部署ant</td><td>编译nutch编码使用<tr><td>2</td><td>下载nutch源码</td><td>必须步骤<tr><td>3</td><td>在nutch源码根目录下,执行ant等待编译完成</td><td>构建nutch<tr><td>4</td><td>配置nutch-site.xml</td><td>必须步骤<tr><td>5</td><td>ant eclipse 构建eclipse项目</td><td>导入eclipse中,调试<tr><td>6</td><td>conf目录置顶</td><td>nutch加载时,会读取配置文件<tr><td>7</td><td>执行org.apache.nutch.crawl.Injector注入种子</td><td>local调试<tr><td>8</td><td>执行org.apache.nutch.crawl.Generator生成一个抓取列表</td><td>local调试<tr><td>9</td><td>执行org.apache.nutch.fetcher.Fetcher生成一个抓取队列</td><td>local调试<tr><td>10</td><td>执行org.apache.nutch.parse.ParseSegment执行contet生一个段文件</td><td>local调试<tr><td>11</td><td>配置好solr服务</td><td>检索服务查询<tr><td>12</td><td>执行org.apache.nutch.indexer.IndexingJob映射solr索引</td><td>local调试<tr><td>13</td><td>映射完成后,就可以solr里面执行查询了</td><td>校验结果</table>

编译完,导入eclipse的中如下图所示,注意conf文件夹置顶:


nutch-site.xml里面的配置如下:
<pre name="code" class="xml"><?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>

<!-- Put site-specific property overrides in this file. -->

<configuration>

<property>
  <name>http.agent.name</name>
  <value>mynutch</value>
</property>


<property>
  <name>http.robots.agents</name>
  <value>*</value>
  <description>The agent strings we'll look for in robots.txt files,
  comma-separated, in decreasing order of precedence. You should
  put the value of http.agent.name as the first agent name, and keep the
  default * at the end of the list. E.g.: BlurflDev,Blurfl,*
  </description>
</property>

<property>
  <name>plugin.folders</name>
  <value>./src/plugin</value>
  <description>Directories where nutch plugins are located.  Each
  element may be a relative or absolute path.  If absolute, it is used
  as is.  If relative, it is searched for on the classpath.</description>
</property>

</configuration>
</pre>
下面简单介绍下,在各个类里运行,需要做的一些改动,首先运行nutch,是基于Hadoop的local模式调试的,所以,你得改变下hadoop的权限,否则在运行过程中,会报错。散仙在这里提供一个简单的方法,拷贝hadoop的FileUtils类进行eclipse中,修改它的权限校验即可,如果你是在linux上运行,就不需要考虑这个问题了。

在开始调试之前,你需要在项目的根目录下建一个urls文件夹,并新建一个种子文件放入你要抓取的网址。

在Injector类里面,run方法里,改成

<pre name="code" class="java">  public int run(String[] args) throws Exception {
//    if (args.length < 2) {
//      System.err.println("Usage: Injector <crawldb> <url_dir>");
//      return -1;
//    }
  args=new String[]{"mydir","urls"};//urls
    try {
      inject(new Path(args[0]), new Path(args[1]));
      return 0;
    } catch (Exception e) {
      LOG.error("Injector: " + StringUtils.stringifyException(e));
      return -1;
    }
  }</pre>
在Generator里面的run方法改成
<pre name="code" class="java">public int run(String[] args) throws Exception {
//    if (args.length < 2) {
//      System.out
//          .println("Usage: Generator <crawldb> <segments_dir> [-force] [-topN N] [-numFetchers numFetchers] [-adddays numDays] [-noFilter] [-noNorm][-maxNumSegments num]");
//      return -1;
//    }
 
  args=new String[]{"mydir","myseg","6","7",""};

    Path dbDir = new Path(args[0]);
    Path segmentsDir = new Path(args[1]);
    long curTime = System.currentTimeMillis();
    long topN = Long.MAX_VALUE;
    int numFetchers = -1;
    boolean filter = true;
    boolean norm = true;
    boolean force = false;
    int maxNumSegments = 1;

    for (int i = 2; i < args.length; i++) {
      if ("-topN".equals(args[i])) {
        topN = Long.parseLong(args[i + 1]);
        i++;
      } else if ("-numFetchers".equals(args[i])) {
        numFetchers = Integer.parseInt(args[i + 1]);
        i++;
      } else if ("-adddays".equals(args[i])) {
        long numDays = Integer.parseInt(args[i + 1]);
        curTime += numDays * 1000L * 60 * 60 * 24;
      } else if ("-noFilter".equals(args[i])) {
        filter = false;
      } else if ("-noNorm".equals(args[i])) {
        norm = false;
      } else if ("-force".equals(args[i])) {
        force = true;
      } else if ("-maxNumSegments".equals(args[i])) {
        maxNumSegments = Integer.parseInt(args[i + 1]);
      }

    }

    try {
      Path[] segs = generate(dbDir, segmentsDir, numFetchers, topN, curTime, filter,
          norm, force, maxNumSegments);
      if (segs == null) return -1;
    } catch (Exception e) {
      LOG.error("Generator: " + StringUtils.stringifyException(e));
      return -1;
    }
    return 0;
  }</pre>

在Fetcher的run方法里面改动:
<pre name="code" class="java">  public int run(String[] args) throws Exception {

    String usage = "Usage: Fetcher <segment> [-threads n]";

     args=new String[]{"D:\\20140520nutchplugin\\apache-nutch-1.8\\myseg\\20140520120541","4"};
//    if (args.length < 1) {
//      System.err.println(usage);
//      return -1;
//    }

    Path segment = new Path(args[0]);

    int threads = getConf().getInt("fetcher.threads.fetch", 10);
    boolean parsing = false;

    for (int i = 1; i < args.length; i++) {       // parse command line
      if (args[i].equals("-threads")) {           // found -threads option
        threads =  Integer.parseInt(args[++i]);
      }
    }

    getConf().setInt("fetcher.threads.fetch", threads);

    try {
      fetch(segment, threads);
      return 0;
    } catch (Exception e) {
      LOG.error("Fetcher: " + StringUtils.stringifyException(e));
      return -1;
    }

  }</pre>
在ParseSegment里面的run方法改动:

<pre name="code" class="java"> public int run(String[] args) throws Exception {
    Path segment;

    String usage = "Usage: ParseSegment segment [-noFilter] [-noNormalize]";

//    if (args.length == 0) {
//      System.err.println(usage);
//      System.exit(-1);
//    }

     args=new String[]{"D:\\20140520nutchplugin\\apache-nutch-1.8\\myseg\\20140520120541"};
    if(args.length > 1) {
      for(int i = 1; i < args.length; i++) {
        String param = args[i];

        if("-nofilter".equalsIgnoreCase(param)) {
          getConf().setBoolean("parse.filter.urls", false);
        } else if ("-nonormalize".equalsIgnoreCase(param)) {
          getConf().setBoolean("parse.normalize.urls", false);
        }
      }
    }

    segment = new Path(args[0]);
    parse(segment);
    return 0;
  }</pre>
在IndexingJob的run方法里面改动:
<pre name="code" class="java">  public int run(String[] args) throws Exception {
    args=new String[]{"mydir","D:\\20140520nutchplugin\\apache-nutch-1.8\\myseg\\20140520120541"};
        if (args.length < 2) {
            System.err
                    .println("Usage: Indexer <crawldb> [-linkdb <linkdb>] [-params k1=v1&k2=v2...] (<segment> ... | -dir <segments>) [-noCommit] [-deleteGone] [-filter] [-normalize]");
            IndexWriters writers = new IndexWriters(getConf());
            System.err.println(writers.describe());
            return -1;
        }

        final Path crawlDb = new Path(args[0]);
        Path linkDb = null;

        final List<Path> segments = new ArrayList<Path>();
        String params = null;

        boolean noCommit = false;
        boolean deleteGone = false;
        boolean filter = false;
        boolean normalize = false;

        for (int i = 1; i < args.length; i++) {
            if (args[i].equals("-linkdb")) {
                linkDb = new Path(args[++i]);
            } else if (args[i].equals("-dir")) {
                Path dir = new Path(args[++i]);
                FileSystem fs = dir.getFileSystem(getConf());
                FileStatus[] fstats = fs.listStatus(dir,
                        HadoopFSUtil.getPassDirectoriesFilter(fs));
                Path[] files = HadoopFSUtil.getPaths(fstats);
                for (Path p : files) {
                    segments.add(p);
                }
            } else if (args[i].equals("-noCommit")) {
                noCommit = true;
            } else if (args[i].equals("-deleteGone")) {
                deleteGone = true;
            } else if (args[i].equals("-filter")) {
                filter = true;
            } else if (args[i].equals("-normalize")) {
                normalize = true;
            } else if (args[i].equals("-params")) {
                params = args[++i];
            } else {
                segments.add(new Path(args[i]));
            }
        }

        try {
            index(crawlDb, linkDb, segments, noCommit, deleteGone, params,
                    filter, normalize);
            return 0;
        } catch (final Exception e) {
            LOG.error("Indexer: " + StringUtils.stringifyException(e));
            return -1;
        }
    }</pre>

除此之外,还需要,在SolrIndexWriter的187行和SolrUtils的54行分别添加如下代码,修改solr的映射地址:
<pre name="code" class="java">String serverURL = conf.get(SolrConstants.SERVER_URL);
        serverURL="http://localhost:8983/solr/";</pre>
<pre name="code" class="java">// String serverURL = job.get(SolrConstants.SERVER_URL);
    String serverURL ="http://localhost:8983/solr";</pre>



按上面几个步骤,每执行一个类的时候,就修改其的运行参数,因为nutch的作业具有依赖性,这一个作业的输入,往往是上一个作业的输出,手动依次运行修改上面的5个类,最终我们的索引就可以生成在solr里,截图如下:




当然,我们还可以,配置分词策略,来使我们检索更加通用,准确.
  • 大小: 397.5 KB
  • 大小: 229.1 KB
论坛首页 入门技术版

跳转论坛:
Global site tag (gtag.js) - Google Analytics