Package org.terrier.matching

Class QueryResultSet

java.lang.Object

org.terrier.matching.CollectionResultSet

org.terrier.matching.QueryResultSet

All Implemented Interfaces:

java.io.Serializable, ResultSet

Direct Known Subclasses:

FatQueryResultSet, FeaturedQueryResultSet

public class QueryResultSet
extends CollectionResultSet

A result set for a given query. This result set is created for a given number of documents, usually the number of retrieved documents for a query. Initially, it is created by cropping an instance of the CollectionResultSet, that is used in the Matching classes.
This class has support for adding metadata as well.

Author:

Vassilis Plachouras, Craig Macdonald

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
protected java.lang.String[][]	metadata	The structure holding the metadata.
protected gnu.trove.TObjectIntHashMap<java.lang.String>	metaMap	The structure that holds the metadata about the results.

Fields inherited from class org.terrier.matching.CollectionResultSet

arraysInitialised, docids, exactResultSize, lock, occurrences, resultSize, scores, statusCode

Constructor Summary

Constructors

Constructor	Description
QueryResultSet(int numberOfDocuments)	A default constructor for the result set with a given number of documents.
QueryResultSet(int[] docids, double[] ds, short[] occurrences)

Method Summary

Modifier and Type	Method	Description
void	addMetaItem(java.lang.String name, int index, java.lang.String value)	Adds a metadata value for a given document
void	addMetaItems(java.lang.String name, java.lang.String[] values)	Adds the metadata values for all the documents in the result set.
protected int	addMetaType(java.lang.String name)	Get the metadata index for the given name
java.lang.String[][]	allMetaItems()	returns all metadata, indexed first by metadata type, then by rank.
java.lang.String	getMetaItem(java.lang.String name, int index)	Gets a metadata value for a given document.
java.lang.String[]	getMetaItems(java.lang.String name)	Gets the metadata information for all documents.
java.lang.String[]	getMetaKeys()	Returns the names of the meta keys which this resultset has
ResultSet	getResultSet(int[] positions)	Extracts a subset of the resultset given by the list parameter, which contains a list of positions in the resultset that should be saved.
ResultSet	getResultSet(int startPosition, int length)	Crops the existing result file and extracts a subset from the given starting point to the ending point.
boolean	hasMetaItems(java.lang.String name)	does the metaindex have a particular type of metadata?
void	initialise()	Initialises the arrays prior of retrieval.
protected QueryResultSet	makeNewResultSet(int length)
void	sort()	Sorts all documents in this resultset by descending score
void	sort(int topDocs)	Sorts the top topDocs document in this resultset be first.

Methods inherited from class org.terrier.matching.CollectionResultSet

getDocids, getExactResultSize, getLock, getOccurrences, getResultSize, getScores, getStatusCode, initialise, setExactResultSize, setResultSize, setStatusCode

Methods inherited from class java.lang.Object

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

Field Detail

metaMap

protected gnu.trove.TObjectIntHashMap<java.lang.String> metaMap

The structure that holds the metadata about the results. It maps the metadata name, a string, to an array index of the metadata array that contains the particular metadata.

metadata

protected java.lang.String[][] metadata

The structure holding the metadata. Each column is associated with a named metavalue type, as known by metaMap. Each row is for a given document.

Constructor Detail

QueryResultSet

public QueryResultSet(int numberOfDocuments)

A default constructor for the result set with a given number of documents.

Parameters:

numberOfDocuments - the number of documents contained in the result set.

QueryResultSet

public QueryResultSet(int[] docids,
                      double[] ds,
                      short[] occurrences)

Method Detail

initialise

public void initialise()

Initialises the arrays prior of retrieval. Only the first time it is called, it will allocate memory for the arrays.

Specified by:

initialise in interface ResultSet

Overrides:

initialise in class CollectionResultSet

addMetaItem

public void addMetaItem(java.lang.String name,
                        int index,
                        java.lang.String value)

Adds a metadata value for a given document

Specified by:

addMetaItem in interface ResultSet

Overrides:

addMetaItem in class CollectionResultSet

Parameters:

name - the name of the metadata type. For example, it can be url for adding the URLs of documents.

index - the position in the resultset array of the given document

value - the metadata value.

addMetaItems

public void addMetaItems(java.lang.String name,
                         java.lang.String[] values)

Adds the metadata values for all the documents in the result set. The length of the metadata array values should be equal to the length of the docids array. The array must be sorted in the same way as the resultset (descending score)

Specified by:

addMetaItems in interface ResultSet

Overrides:

addMetaItems in class CollectionResultSet

Parameters:

name - the name of the metadata type. For example, it can be url for adding the URLs of documents.

values - the metadata values.

getMetaItem

public java.lang.String getMetaItem(java.lang.String name,
                                    int index)

Gets a metadata value for a given document. If the requested metadata information is not specified, then we return null.

Specified by:

getMetaItem in interface ResultSet

Overrides:

getMetaItem in class CollectionResultSet

Parameters:

name - the name of the metadata type.

index - the postition in the array

Returns:

a string with the metadata information, or null of the metadata is not available.

getMetaItems

public java.lang.String[] getMetaItems(java.lang.String name)

Gets the metadata information for all documents. If the requested metadata information is not specified, then we return null.

Specified by:

getMetaItems in interface ResultSet

Overrides:

getMetaItems in class CollectionResultSet

Parameters:

name - the name of the metadata type.

Returns:

an array of strings with the metadata information, or null of the metadata is not available.

allMetaItems

public java.lang.String[][] allMetaItems()

returns all metadata, indexed first by metadata type, then by rank. The order of keys is as in getMetaKeys()

Specified by:

allMetaItems in interface ResultSet

Overrides:

allMetaItems in class CollectionResultSet

hasMetaItems

public boolean hasMetaItems(java.lang.String name)

does the metaindex have a particular type of metadata?

Specified by:

hasMetaItems in interface ResultSet

Overrides:

hasMetaItems in class CollectionResultSet

Parameters:

name - of the desired metaitem set

Returns:

true if the set exists.

getMetaKeys

public java.lang.String[] getMetaKeys()

Returns the names of the meta keys which this resultset has

Specified by:

getMetaKeys in interface ResultSet

Overrides:

getMetaKeys in class CollectionResultSet

Returns:

the list of key names

addMetaType

protected int addMetaType(java.lang.String name)

Get the metadata index for the given name

Parameters:

name - The name of the metadata type to add on

Returns:

Integer of the place in the metadata array it should be stored in

makeNewResultSet

protected QueryResultSet makeNewResultSet(int length)

getResultSet

public ResultSet getResultSet(int startPosition,
                              int length)

Crops the existing result file and extracts a subset from the given starting point to the ending point.

Specified by:

getResultSet in interface ResultSet

Overrides:

getResultSet in class CollectionResultSet

Parameters:

startPosition - the beginning of the subset.

length - the number of entries to get.

Returns:

a subset of the current result set.

getResultSet

public ResultSet getResultSet(int[] positions)

Extracts a subset of the resultset given by the list parameter, which contains a list of positions in the resultset that should be saved.

Specified by:

getResultSet in interface ResultSet

Overrides:

getResultSet in class CollectionResultSet

Parameters:

positions - int[] the list of elements in the current list that should be kept.

Returns:

a subset of the current result set specified by the list.

sort

public void sort()

Description copied from interface: ResultSet

Sorts all documents in this resultset by descending score

Specified by:

sort in interface ResultSet

Overrides:

sort in class CollectionResultSet

sort

public void sort(int topDocs)

Description copied from interface: ResultSet

Sorts the top topDocs document in this resultset be first. The order of the remaining documents is undefined.

Specified by:

sort in interface ResultSet

Overrides:

sort in class CollectionResultSet

Parameters:

topDocs - number of documents to top-rank