Package com.tibco.patterns.learn.recpair

Class RecPairScorer

java.lang.Object

com.tibco.patterns.learn.recpair.RecPairScorer

All Implemented Interfaces:

java.lang.AutoCloseable

public final class RecPairScorer
extends java.lang.Object
implements java.lang.AutoCloseable

Manages calculation of feature scores and communication with Patterns server. Does approximate text matching between the two records in the pair. Runs specific queries on the server using the feature query, maintains the necessary thesauri and an optional model on the server. The input of Record Pair Scorer is predefined record pairs (from file, other class, etc.) Features are defined in FeatureQuery object. Output is a feature vector for the RecPairExample. It contains match scores for each feature.

Constructor Summary

Constructors

Constructor	Description
RecPairScorer(com.netrics.likeit.NetricsServerInterface server, com.netrics.likeit.NetricsTable tableDef, FeatureQuery featureQuery, java.nio.file.Path thesDir)	Creates scorer object for scoring records.
RecPairScorer(com.netrics.likeit.NetricsServerInterface server, java.lang.String table, FeatureQuery featureQuery, java.nio.file.Path thesDir)	Creates scorer object for scoring records and getting field values from the existing table on the server.

Method Summary

Modifier and Type	Method	Description
double	calcAndScore(java.util.List<java.lang.String> tableRecValues, java.util.List<java.lang.String> queryValues)	Calculates symmetric match score for the given pair of records using all feature querylets (that compute symmetric querylet scores) combined with an AND score combiner.
double	calcQltScore(java.lang.String featureName, int qltIndex, java.util.List<java.lang.String> tableRecValues, java.util.List<java.lang.String> queryValues)	Calculates score of a single querylet of the given feature.
void	close()	Deletes thesauri and model (if it exists) from server.
protected void	finalize()	Provided for safety, but client code must call close() explicitly.
java.util.List<com.netrics.likeit.NetricsSearchResult>	getLowConfMatches(java.util.List<java.lang.String> queryValues, double confThreshold, Confidence conf)	Runs a query on server to get the low confidence matches in the data table for the given query record.The query uses the model and thesauri on server.
boolean	isModelLoaded()	Checks if the model was loaded to server by this object.
void	loadModel(java.lang.String projectName, java.nio.file.Path modelFile)	Loads the model to server.
void	safeDeleteModel()	Deletes the model from server if it has been loaded.
double[]	scoreRecPair(RecPairId recPairId)	Computes symmetric querylet scores for the pair of records defined by key0 and key1.
double[]	scoreRecPair(java.util.List<java.lang.String> tableRecValues, java.util.List<java.lang.String> queryValues)	Computes symmetric querylet scores for the given pair of records.
java.lang.String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

RecPairScorer

public RecPairScorer(com.netrics.likeit.NetricsServerInterface server,
                     com.netrics.likeit.NetricsTable tableDef,
                     FeatureQuery featureQuery,
                     java.nio.file.Path thesDir)
              throws java.io.IOException,
                     com.netrics.likeit.NetricsException

Creates scorer object for scoring records. Does not use existing table on the server. Methods that search or get field values from the table on server cannot be used.

Parameters:

server - - refers to the server instance used to calculate feature scores. Not null.

tableDef - - definition of the table that all records come from. Not null.

featureQuery - - defines the query with model features (querylets). Not null.

thesDir - - the directory with all thesaurus files used in feature query. May be null if no thesauri are used.

Throws:

com.netrics.likeit.NetricsException - if error occured on server.

java.io.IOException - if the server does not exist, is not accessible, or if an I/O error occurs while communicating with server

RecPairScorer

public RecPairScorer(com.netrics.likeit.NetricsServerInterface server,
                     java.lang.String table,
                     FeatureQuery featureQuery,
                     java.nio.file.Path thesDir)
              throws java.io.IOException,
                     com.netrics.likeit.NetricsException

Creates scorer object for scoring records and getting field values from the existing table on the server. Loads thesauri to server.

Parameters:

server - - refers to the server instance used to calculate feature scores.

table - - the table that contains all records. Must exist on server.

featureQuery - - defines the query with model features (querylets)

thesDir - - the directory with all thesaurus files used in feature query. May be null if no thesauri are used.

Throws:

com.netrics.likeit.NetricsException - if the given table is not found on server (TBLNOTFOUND), or if another error occured on server.

java.io.IOException - if the server does not exist, is not accessible or if an I/O error occurs while communicating with server.

Method Detail

close

public void close()
           throws java.io.IOException,
                  com.netrics.likeit.NetricsException

Deletes thesauri and model (if it exists) from server. Must be called when this object will no longer be used.

Specified by:

close in interface java.lang.AutoCloseable

Throws:

com.netrics.likeit.NetricsException - if the server indicates that an error has occured

java.io.IOException - if an I/O error occurs while communicating with server

finalize

protected void finalize()
                 throws java.io.IOException,
                        com.netrics.likeit.NetricsException

Provided for safety, but client code must call close() explicitly.

Overrides:

finalize in class java.lang.Object

Throws:

com.netrics.likeit.NetricsException - if the server indicates that an error has occured

java.io.IOException - if an I/O error occurs while communicating with server

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

Returns:

information on table, query, thesauri and model on server.

scoreRecPair

public double[] scoreRecPair(RecPairId recPairId)
                      throws java.io.IOException,
                             com.netrics.likeit.NetricsException

Computes symmetric querylet scores for the pair of records defined by key0 and key1. The two records must exist in the server table which must be provided to the constructor.

Parameters:

recPairId - - the keys for the record pair to be scored. Not null.

Returns:

Feature vector, i.e. querylet scores for the record pair.

Throws:

com.netrics.likeit.NetricsException - if a record does not exist in the server table, or other errors on server.

java.io.IOException - if I/O error occurs while communicating with server

java.lang.IllegalArgumentException - if number of query field values does not match the number of all fields in feature query.

java.lang.IllegalStateException - if server table was not provided to the constructor, or if the object has been closed.

scoreRecPair

public double[] scoreRecPair(java.util.List<java.lang.String> tableRecValues,
                             java.util.List<java.lang.String> queryValues)
                      throws java.io.IOException,
                             com.netrics.likeit.NetricsException

Computes symmetric querylet scores for the given pair of records.

Parameters:

tableRecValues - - field values of record 0. Does not include the key field. Not null.

queryValues - - field values of record 1. Does not include the key field. Not null.

Returns:

Feature vector, i.e. querylet scores for the record pair.

Throws:

com.netrics.likeit.NetricsException - - errors on the server

java.io.IOException - if I/O error occurs while communicating with server

java.lang.IllegalArgumentException - if numbers of values in table record and query do not match, or if number of query field values does not match the number of all fields in feature query.

java.lang.IllegalStateException - if this object has been closed.

calcAndScore

public double calcAndScore(java.util.List<java.lang.String> tableRecValues,
                           java.util.List<java.lang.String> queryValues)
                    throws java.io.IOException,
                           com.netrics.likeit.NetricsException

Calculates symmetric match score for the given pair of records using all feature querylets (that compute symmetric querylet scores) combined with an AND score combiner.

Parameters:

tableRecValues - - field values of record 0. Does not include the key field. Not null.

queryValues - - field values of record 1. Does not include the key field. Not null.

Returns:

the match score for the record pair from the AND score combiner.

Throws:

com.netrics.likeit.NetricsException - - errors on the server

java.io.IOException - if I/O error occurs while communicating with server

java.lang.IllegalArgumentException - if numbers of values in table record and query do not match, or if number of query field values does not match the number of all fields in feature query.

java.lang.IllegalStateException - if this object has been closed.

calcQltScore

public double calcQltScore(java.lang.String featureName,
                           int qltIndex,
                           java.util.List<java.lang.String> tableRecValues,
                           java.util.List<java.lang.String> queryValues)
                    throws java.io.IOException,
                           com.netrics.likeit.NetricsException

Calculates score of a single querylet of the given feature.

Parameters:

featureName - - name of the feature. If several features have the same name, the first such feature is used.

qltIndex - - index of the querylet within the feature.

tableRecValues - - field values of record 0. Does not include the key field. Not null.

queryValues - - field values of record 1. Does not include the key field. Not null.

Returns:

the match score of the querylet.

Throws:

java.lang.IllegalArgumentException - if feature with the given name is not found, or if number of query field values does not match the number of all fields.

java.lang.IllegalStateException - if this object has been closed.

com.netrics.likeit.NetricsException - - errors on the server

java.io.IOException - if I/O error occurs while communicating with server

loadModel

public void loadModel(java.lang.String projectName,
                      java.nio.file.Path modelFile)
               throws com.netrics.likeit.NetricsException,
                      java.io.IOException

Loads the model to server. Deletes any previous model.

Parameters:

projectName - - project name to be prepended to the model name on server.

modelFile - - full path to the model file to be loaded.

Throws:

com.netrics.likeit.NetricsException - if the server indicates that an error has occured

java.io.IOException - if an I/O error occurs

java.lang.IllegalArgumentException - if modelFile does not exist.

java.lang.IllegalStateException - if this object has been closed.

safeDeleteModel

public void safeDeleteModel()
                     throws com.netrics.likeit.NetricsException,
                            java.io.IOException

Deletes the model from server if it has been loaded.

Throws:

com.netrics.likeit.NetricsException - if the server indicates that an error has occured.

java.io.IOException - if an I/O error occurs.

isModelLoaded

public boolean isModelLoaded()

Checks if the model was loaded to server by this object. (It does not check the server if the loaded model is still there.)

Returns:

true if the trained model has been leaded to server.

getLowConfMatches

public java.util.List<com.netrics.likeit.NetricsSearchResult> getLowConfMatches(java.util.List<java.lang.String> queryValues,
                                                                                double confThreshold,
                                                                                Confidence conf)
                                                                         throws com.netrics.likeit.NetricsException,
                                                                                java.io.IOException

Runs a query on server to get the low confidence matches in the data table for the given query record.The query uses the model and thesauri on server.

Parameters:

queryValues - - field values for the query record. Not null.

confThreshold - - return records with confidence less than or equal to this threshold.

conf - - contains feature confidence parameters to be used in the query. Not null.

Returns:

the low confidence matches in the data table returned by the query.

Throws:

com.netrics.likeit.NetricsException - if the server indicates that an error has occured

java.io.IOException - if an I/O error occurs

java.lang.IllegalStateException - if this object has been closed, or if the data table or the model was not loaded to server, or if conf is not FeatureQConf.