textractor.didyoumean
Class DidYouMean

java.lang.Object
  textractor.didyoumean.DidYouMean

All Implemented Interfaces:

DidYouMeanI

public class DidYouMean
extends Object
implements DidYouMeanI

DidYouMean class for suggesting new search terms, a la 'Did You Mean' on Google.com.

Constructor Summary

DidYouMean(DocumentIndexManager documentManager)
Initialize the DidYouMean engine.

Method Summary

String	getDidYouMeanBasename()
void	setDidYouMeanBasename(String basename)
List<ScoredResult>	suggest(String term, boolean orderWithVignaScore, float cutoff) Returns "Did you mean" suggestions based on a search term.
List<ScoredResult>	suggest(String term, float cutoff) Returns "Did you mean" suggestions based on a search term.
List<ScoredResult>	suggestPaiceHusk(String term, float cutoff) Suggest terms that belong to the same stemmed class as the query term.
List<ScoredResult>	suggestRelated(String term, float cutoff) Suggests terms related to this term.
List<ScoredResult>	suggestRelated2(String term, float cutoff) Deprecated. This method will be removed in the near future. Use suggestRelated for biomedical corpora.

Methods inherited from class java.lang.Object

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

Constructor Detail

DidYouMean

public DidYouMean(DocumentIndexManager documentManager)
           throws IOException,
                  NoSuchMethodException,
                  IllegalAccessException,
                  ConfigurationException,
                  InvocationTargetException,
                  InstantiationException,
                  ClassNotFoundException,
                  URISyntaxException

Initialize the DidYouMean engine.

Parameters:

documentManager - DocumenManager for the main index (the engine will find the associated DidYouMean index)

Throws:

IOException - If an error occured reading the did you mean index data.

NoSuchMethodException

IllegalAccessException

ConfigurationException

InvocationTargetException

InstantiationException

ClassNotFoundException

URISyntaxException

Method Detail

suggest

public List<ScoredResult> suggest(String term,
                                  float cutoff)
                           throws IOException,
                                  ConfigurationException,
                                  ParseException,
                                  ClassNotFoundException,
                                  QueryParserException,
                                  QueryBuilderVisitorException

Returns "Did you mean" suggestions based on a search term.

Specified by:

suggest in interface DidYouMeanI

Parameters:

term - - the search term

Returns:

an ArrayList of suggestions, of null if no suggestions are found

Throws:

IOException

ConfigurationException

ParseException

ClassNotFoundException

QueryParserException

QueryBuilderVisitorException

suggest

public List<ScoredResult> suggest(String term,
                                  boolean orderWithVignaScore,
                                  float cutoff)
                           throws IOException,
                                  ParseException,
                                  QueryParserException,
                                  QueryBuilderVisitorException

Returns "Did you mean" suggestions based on a search term.

Specified by:

suggest in interface DidYouMeanI

Parameters:

term - the search term

orderWithVignaScore - If true, terms are ordered first by vigna score, and then by similarity to the search term

Returns:

a List of suggestions, of null if no suggestions are found

Throws:

IOException

ParseException

QueryParserException

QueryBuilderVisitorException

suggestRelated2

@Deprecated
public List<ScoredResult> suggestRelated2(String term,
                                                     float cutoff)
                                   throws ConfigurationException,
                                          IOException,
                                          ParseException,
                                          ClassNotFoundException,
                                          QueryParserException,
                                          QueryBuilderVisitorException

Deprecated. This method will be removed in the near future. Use suggestRelated for biomedical corpora.

Suggests terms related to this term. Terms suggested by this method have the same stem as the input term. We use a stemming model tuned for Medline. Terms whose prefix suffix probability are below the model's are not returned. Try experimenting with cutoff values 1E-3 to 1E-4.

Parameters:

term - Term suggestions are sought for.

cutoff - Probability cutoff. Try 1E-3 or 1E-4.

Returns:

Terms syntactically similar to query that share the same stem.

Throws:

ConfigurationException

IOException

ParseException

ClassNotFoundException

QueryParserException

QueryBuilderVisitorException

suggestRelated

public List<ScoredResult> suggestRelated(String term,
                                         float cutoff)
                                  throws ConfigurationException,
                                         IOException,
                                         ParseException,
                                         ClassNotFoundException,
                                         QueryParserException,
                                         QueryBuilderVisitorException

Suggests terms related to this term. Terms suggested by this method are morphologically related to the input term. Each suggestion is associated with a score that indicates how similar the suggestion is to the input term. We use a stemming method tuned for Medline. Terms whose prefix suffix probability are below the model's are not returned. On the Medline corpus, cutoff values around 1E-7 seem to work reasonably well, but a higher cutoff can be used if more stringency is required.

Specified by:

suggestRelated in interface DidYouMeanI

Parameters:

term - Term suggestions are sought for.

cutoff - Probability cutoff. Try 1E-7 for the Medline corpus.

Returns:

Terms syntactically similar to query that share the same stem.

Throws:

ConfigurationException

IOException

ParseException

ClassNotFoundException

QueryParserException

QueryBuilderVisitorException

suggestPaiceHusk

public List<ScoredResult> suggestPaiceHusk(String term,
                                           float cutoff)
                                    throws ConfigurationException,
                                           IOException,
                                           ParseException,
                                           ClassNotFoundException,
                                           QueryParserException,
                                           QueryBuilderVisitorException

Suggest terms that belong to the same stemmed class as the query term. Use the PaiceHusk stemmer to define the stemming class.

Specified by:

suggestPaiceHusk in interface DidYouMeanI

Parameters:

term -

cutoff -

Returns:

Throws:

ConfigurationException

IOException

ParseException

ClassNotFoundException

QueryParserException

QueryBuilderVisitorException

getDidYouMeanBasename

public String getDidYouMeanBasename()

setDidYouMeanBasename

public void setDidYouMeanBasename(String basename)