ASF_20_SRC_AUTO/src/blocks/lucene/samples/statistic-index.xsp

<?xml version="1.0" encoding="ISO-8859-1"?>

<xsp:page language="java" xmlns:xsp="http://apache.org/xsp">

  <xsp:structure>
    <xsp:include>org.apache.avalon.framework.context.ContextException</xsp:include>
    <xsp:include>org.apache.cocoon.components.search.*</xsp:include>
    
    <xsp:include>org.apache.lucene.document.Document</xsp:include>
    <xsp:include>org.apache.lucene.index.*</xsp:include>
    <xsp:include>org.apache.lucene.document.*</xsp:include>
    <xsp:include>org.apache.lucene.store.*</xsp:include>
    
    <xsp:include>java.io.*</xsp:include>
    <xsp:include>java.util.*</xsp:include>
    <xsp:include>java.net.*</xsp:include>
  </xsp:structure>
  
  <xsp:logic>
    File workDir = null;
    /** Contextualize this class */
    public void contextualize(Context context) throws ContextException {
      super.contextualize( context );
      workDir = (File) context.get(Constants.CONTEXT_WORK_DIR);
    }
    
    Directory directory;
    
    void init(String indexName) throws ProcessingException {
      try {
        directory = LuceneCocoonHelper.getDirectory( new File( workDir, indexName ), false );
      } catch (Exception e) {
        throw new ProcessingException( "Exception in init()!", e );
      }
    }

    void closeReader( IndexReader reader ) {
      if (reader != null) {
        try {
          reader.close();
        } catch (IOException ioe) {
        }
      }
    }
    
    int numDocs() throws ProcessingException {
      IndexReader reader = null;
      try {
        reader = IndexReader.open( directory );
        int num_docs;
        num_docs = reader.numDocs();
        return num_docs;
      } catch (Exception e) {
        throw new ProcessingException( "Exception in numDocs()!", e );
      } finally {
        closeReader(reader);
      }
    }
    
    Map allDocuments() throws ProcessingException {
      IndexReader reader = null;
      try {
        reader = IndexReader.open( directory );
        
        HashMap fieldsStatistic = new HashMap();
        
        for (int i = 0; i &lt; reader.maxDoc(); i++ ) {
          if (!reader.isDeleted(i)) {
            Document document = reader.document(i);
            Enumeration fields = document.fields();
            while (fields.hasMoreElements()) {
              Field f = (Field)fields.nextElement();
              String name = f.name();
              String value = f.stringValue();
              if (value == null) value = "--void--";
              
              String fieldStatistic = name + "/" + value;
              if (fieldsStatistic.get( fieldStatistic ) == null) {
                fieldsStatistic.put( fieldStatistic, new Integer(1) );
              } else {
                Integer sum = (Integer)fieldsStatistic.get( fieldStatistic );
                int sum_plus = sum.intValue() + 1;
                fieldsStatistic.put( fieldStatistic, new Integer( sum_plus ) );
              }
            }
          }
        }
        return fieldsStatistic;
        //map.keySet();
      } catch (Exception e) {
        throw new ProcessingException( "Exception allDocuments()!", e );
      } finally {
        closeReader(reader);
      }
    }

    Map allTerms() throws ProcessingException {
      IndexReader reader = null;
      try {
        reader = IndexReader.open( directory );
        
        TermEnum te = reader.terms();
        HashMap all_fields = new HashMap();
        while (te.next()) {
          Term term = te.term();
          int docfreq = te.docFreq();
          String field = term.field();
          if (field != null) {
            if (all_fields.containsKey( field )) {
              Integer sum = (Integer)all_fields.get( field );
              int sum_plus = sum.intValue() + docfreq;
              all_fields.put( field, new Integer( sum_plus ) );
            } else {
              all_fields.put( field, new Integer( docfreq ) );
            }
          }
        }
        te.close();
        return all_fields;
      } catch (Exception e) {
        throw new ProcessingException( "Exception allDocuments()!", e );
      } finally {
        closeReader(reader);
      }
    }
    Map sort( Map map ) {
      TreeMap treeMap = new TreeMap( map );
      return treeMap;
    }
  </xsp:logic>
  
  <page>
    <xsp:logic>
      String indexName = request.getParameter("indexName");
      if (indexName == null) indexName = "index";
      init(indexName);
    </xsp:logic>
    <title>Index Statistics</title>
    <content>
      <para>
        <font size="-1">
          <a href="welcome">Welcome</a>
        </font>
      </para>
      <para>
        Statistics:
      </para>
      <para>
        Total Count Of Documents
        <xsp:expr>String.valueOf(numDocs())</xsp:expr>
      </para>
      <para>
        <table>
          <tr>
            <td>Count Of Terms</td><td>Fieldname/Fieldvalue</td>
          </tr>
          <xsp:logic>
            Map all_docs = sort(allDocuments());
            Iterator it1 = all_docs.keySet().iterator();
            while (it1.hasNext()) {
              String k = (String)it1.next();
              Integer v = (Integer)all_docs.get( k );
              <xsp:content>
                <tr>
                  <td> <xsp:expr>v.toString()</xsp:expr> </td>
                  <td> <xsp:expr>k</xsp:expr> </td>
                </tr>
              </xsp:content>
            }
          </xsp:logic>
        </table>
      </para>
      <para>
        All Terms
      </para>
      <para>
        <table>
          <tr>
            <td>Count Of Terms</td><td>Term</td>
          </tr>
          <xsp:logic>
            Map all_terms = sort(allTerms());
            Iterator it2 = all_terms.keySet().iterator();
            while (it2.hasNext()) {
              String k = (String)it2.next();
              Integer v = (Integer)all_terms.get( k );
              <xsp:content>
                <tr>
                  <td> <xsp:expr>v.toString()</xsp:expr> </td>
                  <td> <xsp:expr>k</xsp:expr> </td> 
                </tr>
              </xsp:content>
            }
          </xsp:logic>
        </table>
      </para>
    </content>
  </page>

</xsp:page>