VectorSearchBedrockRerankingConfigurationProperty Struct
Configuration for using Amazon Bedrock foundation models to rerank Knowledge Base vector search results.
This enables more sophisticated relevance ranking using large language models.
Example
ts
// The code below shows an example of how to instantiate this type.
// The values are placeholders you should change.
import { aws_bedrock as bedrock } from 'aws-cdk-lib';
declare const additionalModelRequestFields: any;
const vectorSearchBedrockRerankingConfigurationProperty: bedrock.CfnFlow.VectorSearchBedrockRerankingConfigurationProperty = {
modelConfiguration: {
modelArn: 'modelArn',
// the properties below are optional
additionalModelRequestFields: additionalModelRequestFields,
},
// the properties below are optional
metadataConfiguration: {
selectionMode: 'selectionMode',
// the properties below are optional
selectiveModeConfiguration: {
fieldsToExclude: [{
fieldName: 'fieldName',
}],
fieldsToInclude: [{
fieldName: 'fieldName',
}],
},
},
numberOfRerankedResults: 123,
};
Properties
metadataConfigurationoptionalIResolvable | CfnFlow.MetadataConfigurationForRerankingPropertymetadata_configurationoptionaltyping.Union[IResolvable, MetadataConfigurationForRerankingProperty]metadataConfigurationoptionalIResolvableMetadataConfigurationoptionalIResolvableMetadataConfigurationoptionalinterface{}#Configuration for how document metadata should be used during the reranking process.
This determines which metadata fields are included when reordering search results.
modelConfigurationIResolvable | CfnFlow.VectorSearchBedrockRerankingModelConfigurationPropertymodel_configurationtyping.Union[IResolvable, VectorSearchBedrockRerankingModelConfigurationProperty]modelConfigurationIResolvableModelConfigurationIResolvableModelConfigurationinterface{}required#Configuration for the Amazon Bedrock foundation model used for reranking.
This includes the model ARN and any additional request fields required by the model.
numberOfRerankedResultsoptionalnumbernumber_of_reranked_resultsoptionalint | floatnumberOfRerankedResultsoptionalNumberNumberOfRerankedResultsoptionaldoubleNumberOfRerankedResultsoptional*float64#The maximum number of results to rerank.
This limits how many of the initial vector search results will be processed by the reranking model. A smaller number improves performance but may exclude potentially relevant results.