Class SingleCellExpressionExperimentCreateSubSetsAndAggregateServiceImpl

java.lang.Object

ubic.gemma.core.analysis.singleCell.aggregate.SingleCellExpressionExperimentCreateSubSetsAndAggregateServiceImpl

All Implemented Interfaces:

SingleCellExpressionExperimentCreateSubSetsAndAggregateService

@Service
public class SingleCellExpressionExperimentCreateSubSetsAndAggregateServiceImpl
extends Object
implements SingleCellExpressionExperimentCreateSubSetsAndAggregateService

Constructor Summary

Constructors

Constructor	Description
SingleCellExpressionExperimentCreateSubSetsAndAggregateServiceImpl()

Method Summary

Modifier and Type	Method	Description
QuantitationType	createSubSetsAndAggregate(ExpressionExperiment expressionExperiment, QuantitationType scQt, CellLevelCharacteristics clc, ExperimentalFactor cellTypeFactor, Map<Characteristic, FactorValue> c2f, SingleCellExperimentSubSetsCreationConfig singleCellExperimentSubSetsCreationConfig, SingleCellAggregationConfig config)	See createSubSetsAndAggregateByCellType(ExpressionExperiment, SingleCellExperimentSubSetsCreationConfig, SingleCellAggregationConfig) for why this is Propagation.NEVER.
QuantitationType	createSubSetsAndAggregateByCellType(ExpressionExperiment expressionExperiment, SingleCellExperimentSubSetsCreationConfig singleCellExperimentSubSetsCreationConfig, SingleCellAggregationConfig config)	Not transactional: SingleCellExpressionExperimentSubSetService.createSubSets(ExpressionExperiment, SingleCellDimension, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellExperimentSubSetsCreationConfig) commits the subset/BioAssay/BioMaterial rows it creates — and with them the ACL rows AclEventListener inserts alongside each — BEFORE SingleCellExpressionExperimentAggregateService.aggregateVectors(ExpressionExperiment, QuantitationType, List, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellAggregationConfig) opens its own, separate, potentially hour-long transaction to stream and aggregate the single-cell data.
QuantitationType	redoAggregate(ExpressionExperiment expressionExperiment, QuantitationType scQt, CellLevelCharacteristics clc, ExperimentalFactor factor, Map<Characteristic, FactorValue> c2f, BioAssayDimension dimension, QuantitationType previousQt, SingleCellAggregationConfig config)	Re-aggregate a dataset by any cell-level characteristics.
QuantitationType	redoAggregateByCellType(ExpressionExperiment expressionExperiment, BioAssayDimension dimension, QuantitationType previousQt, SingleCellAggregationConfig config)	Re-aggregate a dataset by cell type.

Methods inherited from class Object

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

Constructor Details

SingleCellExpressionExperimentCreateSubSetsAndAggregateServiceImpl

public SingleCellExpressionExperimentCreateSubSetsAndAggregateServiceImpl()

Method Details

createSubSetsAndAggregateByCellType

@Transactional(propagation=NEVER)
public QuantitationType createSubSetsAndAggregateByCellType(ExpressionExperiment expressionExperiment,
 SingleCellExperimentSubSetsCreationConfig singleCellExperimentSubSetsCreationConfig,
 SingleCellAggregationConfig config)

Not transactional: SingleCellExpressionExperimentSubSetService.createSubSets(ExpressionExperiment, SingleCellDimension, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellExperimentSubSetsCreationConfig) commits the subset/BioAssay/BioMaterial rows it creates — and with them the ACL rows AclEventListener inserts alongside each — BEFORE SingleCellExpressionExperimentAggregateService.aggregateVectors(ExpressionExperiment, QuantitationType, List, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellAggregationConfig) opens its own, separate, potentially hour-long transaction to stream and aggregate the single-cell data.

Wrapping both in one transaction (as this method did until 2026-09-30) holds those ACL inserts' locks for the aggregation's entire duration. Every ACL insert anywhere in Gemma references one of a handful of shared acl_sid rows and a shared acl_class row per entity type, so a long-held lock here blocks unrelated ACL-creating writers elsewhere in the system — measured on frink 2026-09-30: diffExAnalyze failed 14+ times over roughly an hour with Lock wait timeout exceeded on its own acl_entry insert while this method's transaction, opened by a concurrent aggregateSingleCellData run, was still open and had not touched an ACL table in well over an hour.

Propagation.NEVER rather than simply dropping @Transactional: it fails loudly if this is ever called from within an existing transaction instead of silently joining one and reintroducing the problem. SingleCellExpressionExperimentSubSetService.createSubSets(ExpressionExperiment, SingleCellDimension, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellExperimentSubSetsCreationConfig), SingleCellExpressionExperimentAggregateService.aggregateVectors(ExpressionExperiment, QuantitationType, List, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellAggregationConfig) and SingleCellExpressionExperimentAggregateService.aggregateVectorsByCellType(ExpressionExperiment, List, SingleCellAggregationConfig) remain independently @Transactional — each opens its own transaction exactly as it already does when called directly from outside this class.

Specified by:

createSubSetsAndAggregateByCellType in interface SingleCellExpressionExperimentCreateSubSetsAndAggregateService

See Also:

• SingleCellExpressionExperimentSubSetService.createSubSetsByCellType(ExpressionExperiment, SingleCellExperimentSubSetsCreationConfig)
• SingleCellExpressionExperimentAggregateService.aggregateVectorsByCellType(ExpressionExperiment, List, SingleCellAggregationConfig)

createSubSetsAndAggregate

@Transactional(propagation=NEVER)
public QuantitationType createSubSetsAndAggregate(ExpressionExperiment expressionExperiment,
 QuantitationType scQt,
 CellLevelCharacteristics clc,
 ExperimentalFactor cellTypeFactor,
 Map<Characteristic, FactorValue> c2f,
 SingleCellExperimentSubSetsCreationConfig singleCellExperimentSubSetsCreationConfig,
 SingleCellAggregationConfig config)

See createSubSetsAndAggregateByCellType(ExpressionExperiment, SingleCellExperimentSubSetsCreationConfig, SingleCellAggregationConfig) for why this is Propagation.NEVER.

Specified by:

createSubSetsAndAggregate in interface SingleCellExpressionExperimentCreateSubSetsAndAggregateService

See Also:

• SingleCellExpressionExperimentSubSetService.createSubSets(ExpressionExperiment, SingleCellDimension, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellExperimentSubSetsCreationConfig)
• SingleCellExpressionExperimentAggregateService.aggregateVectors(ExpressionExperiment, QuantitationType, List, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellAggregationConfig)

redoAggregateByCellType

@Transactional
public QuantitationType redoAggregateByCellType(ExpressionExperiment expressionExperiment,
 BioAssayDimension dimension,
 @Nullable
 QuantitationType previousQt,
 SingleCellAggregationConfig config)

Description copied from interface: SingleCellExpressionExperimentCreateSubSetsAndAggregateService

Re-aggregate a dataset by cell type.

Specified by:

redoAggregateByCellType in interface SingleCellExpressionExperimentCreateSubSetsAndAggregateService

Parameters:

dimension - a dimension to reuse for aggregating

previousQt - a previous quantitation type the re-aggregated one is replacing, it will be removed prior to the new one being added. Ignored if null.

See Also:

• SingleCellExpressionExperimentAggregateService.aggregateVectorsByCellType(ExpressionExperiment, List, SingleCellAggregationConfig)

redoAggregate

@Transactional
public QuantitationType redoAggregate(ExpressionExperiment expressionExperiment,
 QuantitationType scQt,
 CellLevelCharacteristics clc,
 ExperimentalFactor factor,
 Map<Characteristic, FactorValue> c2f,
 BioAssayDimension dimension,
 @Nullable
 QuantitationType previousQt,
 SingleCellAggregationConfig config)

Description copied from interface: SingleCellExpressionExperimentCreateSubSetsAndAggregateService

Re-aggregate a dataset by any cell-level characteristics.

Specified by:

redoAggregate in interface SingleCellExpressionExperimentCreateSubSetsAndAggregateService

Parameters:

dimension - a dimension to reuse for aggregating

previousQt - a previous quantitation type the re-aggregated one is replacing, it will be removed prior to the new one being added. Ignored if null.

See Also:

• SingleCellExpressionExperimentAggregateService.aggregateVectors(ExpressionExperiment, QuantitationType, List, CellLevelCharacteristics, ExperimentalFactor, Map, SingleCellAggregationConfig)