org.apache.solr.handler.component

Class PhrasesIdentificationComponent.Phrase

java.lang.Object

org.apache.solr.handler.component.PhrasesIdentificationComponent.Phrase

Enclosing class:

PhrasesIdentificationComponent

public static final class PhrasesIdentificationComponent.Phrase
extends Object

Model the data known about a single (candidate) Phrase -- which may or may not be indexed

NOTE: This API is for internal purposes only and might change in incompatible ways in the next release.

Method Summary

Modifier and Type	Method and Description
static List<PhrasesIdentificationComponent.Phrase>	extractPhrases(String input, SchemaField analysisField, int maxIndexedPositionLength, int maxQueryPositionLength) Factory method for constructing a list of Phrases given the specified input and using the analyzer for the specified field.
static List<NamedList<Object>>	formatShardResponse(List<PhrasesIdentificationComponent.Phrase> phrases) Format the phrases suitable for returning in a shard response
long	getConjunctionDocCount(String field) Returns the number of documents that contain all of the getIndividualIndexedTerms() that make up this Phrase, in the specified field.
NamedList	getDetails()
long	getDocFreq(String field) Returns the number of documents that contain this (indexed) Phrase as term in the specified field.
double	getFieldScore(String field) Returns the score for this Phrase in this given field.
List<PhrasesIdentificationComponent.Phrase>	getIndexedSuperPhrases() Returns all phrases larger then this phrase, which fully include this phrase, and are indexed.
List<PhrasesIdentificationComponent.Phrase>	getIndividualIndexedTerms() Returns the list of "individual" (ie: getPositionLength()==1 terms.
List<PhrasesIdentificationComponent.Phrase>	getLargestIndexedSubPhrases() Returns the list of (overlapping) sub phrases that have the largest possible size based on the effective value of PhrasesIdentificationComponent.PhrasesContextData.maxIndexedPositionLength.
int	getOffsetEnd()
int	getOffsetStart()
int	getPositionEnd() NOTE: positions start at '1'
int	getPositionLength()
BitSet	getPositionsBitSet() Each set bit identifies a position filled by this Phrase
int	getPositionStart() NOTE: positions start at '1'
CharSequence	getSubSequence() The characters from the original input that corrispond with this Phrase
double	getTotalScore() Returns the overall score for this Phrase.
long	getTTF(String field) Returns the number of total TTF of this (indexed) Phrase as term in the specified field.
static void	populateScores(List<PhrasesIdentificationComponent.Phrase> phrases, Map<String,Double> fieldWeights, int maxIndexedPositionLength, int maxQueryPositionLength) Public for testing purposes
static void	populateScores(PhrasesIdentificationComponent.PhrasesContextData contextData) Uses the previously popuated stats to populate each Phrase with it's scores for the specified fields, and it's over all (weighted) total score.
static void	populateStats(List<PhrasesIdentificationComponent.Phrase> phrases, Collection<String> fieldNames, SolrIndexSearcher searcher) Populates the phrases with stats from the local index for the specified fields
static void	populateStats(List<PhrasesIdentificationComponent.Phrase> phrases, List<NamedList<Object>> shardData) Populates the phrases with (merged) stats from a remote shard
String	toString()

Methods inherited from class java.lang.Object

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

Method Detail

extractPhrases

public static List<PhrasesIdentificationComponent.Phrase> extractPhrases(String input,
                                                                         SchemaField analysisField,
                                                                         int maxIndexedPositionLength,
                                                                         int maxQueryPositionLength)

Factory method for constructing a list of Phrases given the specified input and using the analyzer for the specified field. The maxIndexedPositionLength and maxQueryPositionLength provided *must* match the effective values used by respective analyzers.

formatShardResponse

public static List<NamedList<Object>> formatShardResponse(List<PhrasesIdentificationComponent.Phrase> phrases)

Format the phrases suitable for returning in a shard response

See Also:

populateStats(List,List)

populateStats

public static void populateStats(List<PhrasesIdentificationComponent.Phrase> phrases,
                                 List<NamedList<Object>> shardData)

Populates the phrases with (merged) stats from a remote shard

See Also:

formatShardResponse(java.util.List<org.apache.solr.handler.component.PhrasesIdentificationComponent.Phrase>)

populateStats

public static void populateStats(List<PhrasesIdentificationComponent.Phrase> phrases,
                                 Collection<String> fieldNames,
                                 SolrIndexSearcher searcher)
                          throws IOException

Populates the phrases with stats from the local index for the specified fields

Throws:

IOException

populateScores

public static void populateScores(PhrasesIdentificationComponent.PhrasesContextData contextData)

Uses the previously popuated stats to populate each Phrase with it's scores for the specified fields, and it's over all (weighted) total score. This is not needed on shard requests.

See Also:

populateStats(java.util.List<org.apache.solr.handler.component.PhrasesIdentificationComponent.Phrase>, java.util.List<org.apache.solr.common.util.NamedList<java.lang.Object>>), getFieldScore(String), getTotalScore()

populateScores

public static void populateScores(List<PhrasesIdentificationComponent.Phrase> phrases,
                                  Map<String,Double> fieldWeights,
                                  int maxIndexedPositionLength,
                                  int maxQueryPositionLength)

Public for testing purposes

See Also:

populateScores(PhrasesIdentificationComponent.PhrasesContextData)

NOTE: This API is for internal purposes only and might change in incompatible ways in the next release.

toString

public String toString()

Overrides:

toString in class Object

getDetails

public NamedList getDetails()

getSubSequence

public CharSequence getSubSequence()

The characters from the original input that corrispond with this Phrase

getIndividualIndexedTerms

public List<PhrasesIdentificationComponent.Phrase> getIndividualIndexedTerms()

Returns the list of "individual" (ie: getPositionLength()==1 terms. NOTE: Indexed phrases of length 1 are the (sole) individual terms of themselves

getLargestIndexedSubPhrases

public List<PhrasesIdentificationComponent.Phrase> getLargestIndexedSubPhrases()

Returns the list of (overlapping) sub phrases that have the largest possible size based on the effective value of PhrasesIdentificationComponent.PhrasesContextData.maxIndexedPositionLength. NOTE: Indexed phrases of length less then the max indexed length are the (sole) largest sub-phrases of themselves.

getIndexedSuperPhrases

public List<PhrasesIdentificationComponent.Phrase> getIndexedSuperPhrases()

Returns all phrases larger then this phrase, which fully include this phrase, and are indexed. NOTE: A Phrase is never the super phrase of itself.

getPositionStart

public int getPositionStart()

NOTE: positions start at '1'

getPositionEnd

public int getPositionEnd()

NOTE: positions start at '1'

getPositionLength

public int getPositionLength()

getPositionsBitSet

public BitSet getPositionsBitSet()

Each set bit identifies a position filled by this Phrase

getOffsetStart

public int getOffsetStart()

getOffsetEnd

public int getOffsetEnd()

getTotalScore

public double getTotalScore()

Returns the overall score for this Phrase. In the current implementation, the only garuntee made regarding the range of possible values is that 0 (or less) means it is not a good phrase.

Returns:

A numeric value indicating the confidence in this Phrase, higher numbers are higher confidence.

getFieldScore

public double getFieldScore(String field)

Returns the score for this Phrase in this given field. In the current implementation, the only garuntee made regarding the range of possible values is that 0 (or less) means it is not a good phrase.

Returns:

A numeric value indicating the confidence in this Phrase for this field, higher numbers are higher confidence.

getTTF

public long getTTF(String field)

Returns the number of total TTF of this (indexed) Phrase as term in the specified field. NOTE: behavior of calling this method is undefined unless one of the populateStats(java.util.List<org.apache.solr.handler.component.PhrasesIdentificationComponent.Phrase>, java.util.List<org.apache.solr.common.util.NamedList<java.lang.Object>>) methods has been called with this field.

getConjunctionDocCount

public long getConjunctionDocCount(String field)

Returns the number of documents that contain all of the getIndividualIndexedTerms() that make up this Phrase, in the specified field. NOTE: behavior of calling this method is undefined unless one of the populateStats(java.util.List<org.apache.solr.handler.component.PhrasesIdentificationComponent.Phrase>, java.util.List<org.apache.solr.common.util.NamedList<java.lang.Object>>) methods has been called with this field.

getDocFreq

public long getDocFreq(String field)

Returns the number of documents that contain this (indexed) Phrase as term in the specified field. NOTE: behavior of calling this method is undefined unless one of the populateStats(java.util.List<org.apache.solr.handler.component.PhrasesIdentificationComponent.Phrase>, java.util.List<org.apache.solr.common.util.NamedList<java.lang.Object>>) methods has been called with this field.