import * as iam from 'aws-cdk-lib/aws-iam';
import { IGrantable } from 'aws-cdk-lib/aws-iam';
import { Construct } from 'constructs';
export interface KnowledgeBaseProps {
    /**
      * The name of the knowledge base to create.
      */
    readonly knowledgeBaseName: string;
    /**
      * Contains details about the vector embeddings configuration of the knowledge base.
      */
    readonly knowledgeBaseConfiguration: KnowledgeBaseConfiguration;
    /**
      * The ARN (Amazon Resource Name) of the S3 bucket where vector embeddings are stored.
      * This bucket contains the vector data used by the knowledge base.
      */
    readonly vectorBucketArn: string;
    /**
      * The ARN (Amazon Resource Name) of the vector index used for the knowledge base.
      * This ARN identifies the specific vector index resource within Amazon Bedrock.
      */
    readonly indexArn: string;
    /**
      * A description of the knowledge base.
      */
    readonly description?: string;
    /**
      * A unique, case-sensitive identifier to ensure that the API request completes no more than one time.
      * Must have length greater than or equal to 33.
      *
      * If this token matches a previous request, Amazon Bedrock ignores the request, but does not return an error.
      * For more information, see [Ensuring Idempotency](https://docs.aws.amazon.com/AWSEC2/latest/APIReference/Run_Instance_Idempotency.html).
      */
    readonly clientToken?: string;
}
export interface KnowledgeBaseConfiguration {
    /**
      * The ARN (Amazon Resource Name) of the model used to create vector embeddings for the knowledge base.
      */
    readonly embeddingModelArn: string;
    /**
      * The data type for the vectors when using a model to convert text into vector embeddings.
      * The model must support the specified data type for vector embeddings.
      *
      * Floating-point (float32) is the default data type, and is supported by most models for vector embeddings.
      * See [Supported embeddings models](https://docs.aws.amazon.com/bedrock/latest/userguide/knowledge-base-supported.html)
      * for information on the available models and their vector data types.
      */
    readonly embeddingDataType?: 'BINARY' | 'FLOAT32';
    /**
      * The dimensions details for the vector configuration used on the Bedrock embeddings model.
      *
      * Must be supported by the chosen embedding model.
      */
    readonly dimensions?: string;
    /**
      * Multi model supplemental data storage configuration
      *
      * See https://docs.aws.amazon.com/bedrock/latest/APIReference/API_agent_SupplementalDataStorageConfiguration.html.
      */
    readonly supplementalDataStorageConfiguration?: SupplementalDataStorageConfiguration;
}
export interface SupplementalDataStorageConfiguration {
    /**
      * The S3 URI for the supplemental data storage.
      */
    readonly s3Location: string;
}
/**
 * Creates a Amazon Bedrock knowledge base with S3 Vectors as the underlying vector store.
 *
 * To create a knowledge base, you must first set up and configure a S3 Vectors bucket and index.
 * For more information, see [Set up a knowledge base](https://docs.aws.amazon.com/bedrock/latest/userguide/knowlege-base-prereq.html).
 */
export declare class KnowledgeBase extends Construct {
    /**
      * The ID of the knowledge base.
      */
    readonly knowledgeBaseId: string;
    /**
      * The Amazon Resource Name (ARN) of the knowledge base.
      */
    readonly knowledgeBaseArn: string;
    /**
     * The IAM role for the knowledge base.
     */
    readonly role: iam.Role;
    /**
      * @summary Creates a new Bedrock knowledge base construct with S3 Vectors as the vector store.
      * @param {cdk.App} scope - Represents the scope for all resources.
      * @param {string} id - Scope-unique id.
      * @param {KnowledgeBaseProps} props - User provided props for the construct.
      * @access public
      */
    constructor(scope: Construct, id: string, props: KnowledgeBaseProps);
    /**
      * Grants permission to start an ingestion job for the knowledge base.
      * @param grantee The principal to grant permissions to.
      */
    grantIngestion(grantee: IGrantable): void;
}