edu.washington.cs.knowitall.extractor.mapper
Class ReVerbRelationDictionaryFilter

java.lang.Object
  edu.washington.cs.knowitall.extractor.mapper.Mapper<T>
      edu.washington.cs.knowitall.extractor.mapper.FilterMapper<ChunkedExtraction>
          edu.washington.cs.knowitall.extractor.mapper.ReVerbRelationDictionaryFilter

public class ReVerbRelationDictionaryFilter
extends FilterMapper<ChunkedExtraction>

Filters relations based on how many distinct arg2 values it takes in a large corpus. These frequencies must be pre-computed and stored in a flat file in the tab-delimited format (#arg2s, relation). The relations should be normalized using the VerbalRelationNormalizer class. By default, this class searches the classpath for a file called freq_rel.txt.gz and reads the relations with at least 20 distinct arg2s from it.

Author:

afader

Field Summary

static int

defaultMinFreq

Constructor Summary

ReVerbRelationDictionaryFilter()
Constructs a new dictionary filter using the data in the file freq_rel.txt.gz, and loads all relations with at least 20 distinct arg2s.

ReVerbRelationDictionaryFilter(InputStream in, int minFreq)
Constructs a new dictionary filter from the data in in.

ReVerbRelationDictionaryFilter(int minFreq)
Constructs a new dictionary filter using the data in the file freq_rel.txt.gz, which is found on the classpath.

Method Summary

boolean

doFilter(ChunkedExtraction extr)

Methods inherited from class edu.washington.cs.knowitall.extractor.mapper.FilterMapper

doMap

Methods inherited from class edu.washington.cs.knowitall.extractor.mapper.Mapper

disable, enable, isEnabled, map

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

defaultMinFreq

public static final int defaultMinFreq

See Also:

Constant Field Values

Constructor Detail

ReVerbRelationDictionaryFilter

public ReVerbRelationDictionaryFilter(InputStream in,
                                      int minFreq)
                               throws IOException

Constructs a new dictionary filter from the data in in. This data should be in the tab-delimited format (#arg2, relation).

Parameters:

in - the dictionary of relations and their number of distinct arg2s.

minFreq - the minimum number of distinct arg2s a relation must have to be included.

Throws:

IOException

ReVerbRelationDictionaryFilter

public ReVerbRelationDictionaryFilter(int minFreq)
                               throws IOException

Constructs a new dictionary filter using the data in the file freq_rel.txt.gz, which is found on the classpath.

Parameters:

minFreq - the minimum number of distinct arg2s a relation must have to be included.

Throws:

IOException

ReVerbRelationDictionaryFilter

public ReVerbRelationDictionaryFilter()
                               throws IOException

Constructs a new dictionary filter using the data in the file freq_rel.txt.gz, and loads all relations with at least 20 distinct arg2s.

Throws:

IOException

Method Detail

doFilter

public boolean doFilter(ChunkedExtraction extr)

Specified by:

doFilter in class FilterMapper<ChunkedExtraction>