Package nzilbb.labbcat.server.db

Class OneQuerySearch

java.lang.Object

java.lang.Thread

nzilbb.labbcat.server.task.Task

nzilbb.labbcat.server.search.SearchTask

nzilbb.labbcat.server.db.OneQuerySearch

All Implemented Interfaces:

Runnable, nzilbb.util.MonitorableTask

public class OneQuerySearch
extends SearchTask

Implementation of search that uses a single monolithic SQL query to identify matches.

Author:

Robert Fromont robert@fromont.net.nz

Nested Class Summary

Nested classes/interfaces inherited from class java.lang.Thread

Thread.State, Thread.UncaughtExceptionHandler

Field Summary

Fields inherited from class java.lang.Thread

MAX_PRIORITY, MIN_PRIORITY, NORM_PRIORITY

Constructor Summary

Constructors

Constructor	Description
OneQuerySearch()

Method Summary

Modifier and Type	Method	Description
void	cancel()	Override to allow update-statement cancellation
String	generateOneSpanSql(Vector<Object> parameters, nzilbb.ag.Schema schema, nzilbb.ag.Layer spanLayer, LayerMatch layerMatch)	Create an SQL query that identifies results that match the search matrix patterns, optimised for a matrix that only searches one span layer, filling in the given lists with information about parameters to set.
String	generateOrthographySql(Vector<Object> parameters, nzilbb.ag.Schema schema)	Create an SQL query that identifies results that match the search matrix patterns, optimised for a matrix that only searches the orthography layer, filling in the given lists with information about parameters to set.
String	generateSql(Vector<Object> parameters, nzilbb.ag.Schema schema, Predicate<nzilbb.ag.Layer> layerIsSpanningAndWordAnchored, UnaryOperator<String> participantCondition, UnaryOperator<String> transcriptCondition)	Create an SQL query that identifies results that match the search matrix patterns, filling in the given lists with information about parameters to set.
static boolean	IsPhraseLayer(nzilbb.ag.Layer layer, nzilbb.ag.Schema schema)	Determines whether the given layer is a 'phrase' layer.
static boolean	IsSegmentLayer(nzilbb.ag.Layer layer, nzilbb.ag.Schema schema)	Determines whether the given layer is a 'segment' layer.
static boolean	IsSpanLayer(nzilbb.ag.Layer layer, nzilbb.ag.Schema schema)	Determines whether the given layer is a 'span' layer.
static boolean	IsWordLayer(nzilbb.ag.Layer layer, nzilbb.ag.Schema schema)	Determines whether the given layer is a 'word' layer.
void	keepAlive()	Somebody is still interested in the thread, so keep it from dying.

Methods inherited from class nzilbb.labbcat.server.search.SearchTask

getAnchorConfidenceThreshold, getDescription, getMainParticipantOnly, getMatchesPerTranscript, getMatrix, getMaxMatches, getOverlapThreshold, getRestrictByUser, getResults, getSuppressResults, getWordsContext, run, setAnchorConfidenceThreshold, setDescription, setMainParticipantOnly, setMatchesPerTranscript, setMatrix, setMaxMatches, setOverlapThreshold, setRestrictByUser, setResults, setSuppressResults, setWordsContext, validate

Methods inherited from class nzilbb.labbcat.server.task.Task

defaultThreadName, findTask, findTask, getCreationTime, getDuration, getLastException, getLastKeepAlive, getLog, getMaxLogSize, getPercentComplete, getResources, getResultTarget, getResultText, getResultUrl, getRunning, getStatus, getStore, getStoreCache, getTaskThreadGroup, getWaitToDieMilliseconds, getWho, release, setLastException, setLastKeepAlive, setMaxLogSize, setResources, setResultTarget, setResultText, setResultUrl, setStatus, setStore, setStoreCache, setWaitToDieMilliseconds, setWho, waitToDie, waitToDie

Methods inherited from class java.lang.Thread

activeCount, checkAccess, countStackFrames, currentThread, dumpStack, enumerate, getAllStackTraces, getContextClassLoader, getDefaultUncaughtExceptionHandler, getId, getName, getPriority, getStackTrace, getState, getThreadGroup, getUncaughtExceptionHandler, holdsLock, interrupt, interrupted, isAlive, isDaemon, isInterrupted, join, join, join, onSpinWait, resume, setContextClassLoader, setDaemon, setDefaultUncaughtExceptionHandler, setName, setPriority, setUncaughtExceptionHandler, sleep, sleep, start, stop, suspend, toString, yield

Methods inherited from class java.lang.Object

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

Methods inherited from interface nzilbb.util.MonitorableTask

getTaskId

Constructor Detail

OneQuerySearch

public OneQuerySearch()

Method Detail

generateSql

public String generateSql(Vector<Object> parameters,
                          nzilbb.ag.Schema schema,
                          Predicate<nzilbb.ag.Layer> layerIsSpanningAndWordAnchored,
                          UnaryOperator<String> participantCondition,
                          UnaryOperator<String> transcriptCondition)
                   throws Exception

Create an SQL query that identifies results that match the search matrix patterns, filling in the given lists with information about parameters to set.

Parameters:

parameters - List of parameter values, which must be Double, Integer, or String.

schema - The layer schema.

layerIsSpanningAndWordAnchored - A predicate that determines whether the given layer is a phrase or span layer for which all annotations share anchors with words.

participantCondition - Supplies the WHERE condition that identifies the target participants, presumably based on Matrix.participantQuery.

transcriptCondition - Supplies the WHERE condition that identifies the target transcripts, presumably based on Matrix.transcriptQuery.

Returns:

An SQL query to implement the query.

Throws:

Exception - If the search should be halted for any reason - e.g. the Matrix.participantQuery identifies no participants.

generateOrthographySql

public String generateOrthographySql(Vector<Object> parameters,
                                     nzilbb.ag.Schema schema)
                              throws Exception

Create an SQL query that identifies results that match the search matrix patterns, optimised for a matrix that only searches the orthography layer, filling in the given lists with information about parameters to set.

This implementation makes speed gains over generateSql(java.util.Vector<java.lang.Object>, nzilbb.ag.Schema, java.util.function.Predicate<nzilbb.ag.Layer>, java.util.function.UnaryOperator<java.lang.String>, java.util.function.UnaryOperator<java.lang.String>) by supporting only the orthography layer, and no border conditions, main-participant or transcript-type filtering. Allowing only plain, full-database searches of the orthography layer means that the initial search query can be completed with no SQL JOINs, meaning that searches can complete in something like one sixth of the time, on very large databases.

Parameters:

parameters - List of parameter values, which must be Double, Integer, or String.

schema - The layer schema.

Returns:

An SQL query to implement the query.

Throws:

Exception - If the search should be halted for any reason - e.g. the Matrix.participantQuery identifies no participants.

generateOneSpanSql

public String generateOneSpanSql(Vector<Object> parameters,
                                 nzilbb.ag.Schema schema,
                                 nzilbb.ag.Layer spanLayer,
                                 LayerMatch layerMatch)
                          throws Exception

Create an SQL query that identifies results that match the search matrix patterns, optimised for a matrix that only searches one span layer, filling in the given lists with information about parameters to set.

This implementation makes speed gains over generateSql(java.util.Vector<java.lang.Object>, nzilbb.ag.Schema, java.util.function.Predicate<nzilbb.ag.Layer>, java.util.function.UnaryOperator<java.lang.String>, java.util.function.UnaryOperator<java.lang.String>) by supporting only a single span (freeform) layer, identifying the matching spans first, and then identifying the first word in each to provide the match token.

Parameters:

parameters - List of parameter values, which must be Double, Integer, or String.

schema - The layer schema.

spanLayer - The one span layer being searched.

layerMatch - The one span's match condition.

Returns:

An SQL query to implement the query.

Throws:

Exception - If the search should be halted for any reason - e.g. the Matrix.participantQuery identifies no participants.

keepAlive

public void keepAlive()

Somebody is still interested in the thread, so keep it from dying.

Overrides:

keepAlive in class Task

cancel

public void cancel()

Override to allow update-statement cancellation

Specified by:

cancel in interface nzilbb.util.MonitorableTask

Overrides:

cancel in class Task

IsSegmentLayer

public static boolean IsSegmentLayer(nzilbb.ag.Layer layer,
                                     nzilbb.ag.Schema schema)

Determines whether the given layer is a 'segment' layer.

Parameters:

layer - The layer to test.

schema - The schema the layer comes from.

Returns:

true if the given layer is the "segment" layer or a child of it.

IsWordLayer

public static boolean IsWordLayer(nzilbb.ag.Layer layer,
                                  nzilbb.ag.Schema schema)

Determines whether the given layer is a 'word' layer.

Parameters:

layer - The layer to test.

schema - The schema the layer comes from.

Returns:

true if the given layer is the word layer or a child of it.

IsPhraseLayer

public static boolean IsPhraseLayer(nzilbb.ag.Layer layer,
                                    nzilbb.ag.Schema schema)

Determines whether the given layer is a 'phrase' layer.

Parameters:

layer - The layer to test.

schema - The schema the layer comes from.

Returns:

true if the given layer is the turn layer or a non-word child of it.

IsSpanLayer

public static boolean IsSpanLayer(nzilbb.ag.Layer layer,
                                  nzilbb.ag.Schema schema)

Determines whether the given layer is a 'span' layer.

Parameters:

layer - The layer to test.

schema - The schema the layer comes from.

Returns:

true if the given layer is an aligned top-level layer - i.e. parent is root.