Examples with LSMInvertedIndexSearchOperatorDescriptor org.apache.hyracks.storage.am.lsm.invertedindex.dataflow.LSMInvertedIndexSearchOperatorDescriptor used on opensource projects

Example 1 with LSMInvertedIndexSearchOperatorDescriptor

use of org.apache.hyracks.storage.am.lsm.invertedindex.dataflow.LSMInvertedIndexSearchOperatorDescriptor in project asterixdb by apache.

the class InvertedIndexPOperator method buildInvertedIndexRuntime.

public Pair<IOperatorDescriptor, AlgebricksPartitionConstraint> buildInvertedIndexRuntime(MetadataProvider metadataProvider, JobGenContext context, JobSpecification jobSpec, AbstractUnnestMapOperator unnestMap, IOperatorSchema opSchema, boolean retainInput, boolean retainMissing, String datasetName, Dataset dataset, String indexName, ATypeTag searchKeyType, int[] keyFields, SearchModifierType searchModifierType, IAlgebricksConstantValue similarityThreshold, int[] minFilterFieldIndexes, int[] maxFilterFieldIndexes, boolean isFullTextSearchQuery) throws AlgebricksException {
    try {
        IAObject simThresh = ((AsterixConstantValue) similarityThreshold).getObject();
        int numPrimaryKeys = dataset.getPrimaryKeys().size();
        Index secondaryIndex = MetadataManager.INSTANCE.getIndex(metadataProvider.getMetadataTxnContext(), dataset.getDataverseName(), dataset.getDatasetName(), indexName);
        if (secondaryIndex == null) {
            throw new AlgebricksException("Code generation error: no index " + indexName + " for dataset " + datasetName);
        }
        IVariableTypeEnvironment typeEnv = context.getTypeEnvironment(unnestMap);
        RecordDescriptor outputRecDesc = JobGenHelper.mkRecordDescriptor(typeEnv, opSchema, context);
        Pair<IFileSplitProvider, AlgebricksPartitionConstraint> secondarySplitsAndConstraint = metadataProvider.getSplitProviderAndConstraints(dataset, indexName);
        // TODO: Here we assume there is only one search key field.
        int queryField = keyFields[0];
        // Get tokenizer and search modifier factories.
        IInvertedIndexSearchModifierFactory searchModifierFactory = InvertedIndexAccessMethod.getSearchModifierFactory(searchModifierType, simThresh, secondaryIndex);
        IBinaryTokenizerFactory queryTokenizerFactory = InvertedIndexAccessMethod.getBinaryTokenizerFactory(searchModifierType, searchKeyType, secondaryIndex);
        IIndexDataflowHelperFactory dataflowHelperFactory = new IndexDataflowHelperFactory(metadataProvider.getStorageComponentProvider().getStorageManager(), secondarySplitsAndConstraint.first);
        LSMInvertedIndexSearchOperatorDescriptor invIndexSearchOp = new LSMInvertedIndexSearchOperatorDescriptor(jobSpec, outputRecDesc, queryField, dataflowHelperFactory, queryTokenizerFactory, searchModifierFactory, retainInput, retainMissing, context.getMissingWriterFactory(), dataset.getSearchCallbackFactory(metadataProvider.getStorageComponentProvider(), secondaryIndex, ((JobEventListenerFactory) jobSpec.getJobletEventListenerFactory()).getJobId(), IndexOperation.SEARCH, null), minFilterFieldIndexes, maxFilterFieldIndexes, isFullTextSearchQuery, numPrimaryKeys, false);
        return new Pair<>(invIndexSearchOp, secondarySplitsAndConstraint.second);
    } catch (MetadataException e) {
        throw new AlgebricksException(e);
    }
}