README.md

Amazon OpenSearch Vector Index Construct Library

---

All classes are under active development and subject to non-backward compatible changes or removal in any future version. These are not subject to the Semantic Versioning model. This means that while you may use them, you may need to update your source code when upgrading to a newer version of this package.

---

Language	Package
TypeScript	@cdklabs/generative-ai-cdk-constructs
Python	cdklabs.generative_ai_cdk_constructs

This construct library provides a resource that creates a vector index on an Amazon OpenSearch Domain. It currently only supports Amazon OpenSearch Serverless.

Table of contents

• API
• Vector Index

API

See the API documentation.

Vector Index

The VectorIndex resource connects to OpenSearch and creates an index suitable for use with Amazon Bedrock Knowledge Bases.

TypeScript

import {
  opensearchserverless,
  opensearch_vectorindex,
} from '@cdklabs/generative-ai-cdk-constructs';

const vectorStore = new opensearchserverless.VectorCollection(
  this,
  'VectorCollection'
);

new opensearch_vectorindex.VectorIndex(this, 'VectorIndex', {
  collection: vectorStore,
  indexName: 'bedrock-knowledge-base-default-index',
  vectorField: 'bedrock-knowledge-base-default-vector',
  vectorDimensions: 1536,
  mappings: [
    {
      mappingField: 'AMAZON_BEDROCK_TEXT_CHUNK',
      dataType: 'text',
      filterable: true,
    },
    {
      mappingField: 'AMAZON_BEDROCK_METADATA',
      dataType: 'text',
      filterable: false,
    },
  ],
  analyzer: {
    characterFilters: [opensearchserverless.CharacterFilterType.ICU_NORMALIZER],
    tokenizer: opensearchserverless.TokenizerType.KUROMOJI_TOKENIZER,
    tokenFilters: [
      opensearchserverless.TokenFilterType.KUROMOJI_BASEFORM,
      opensearchserverless.TokenFilterType.JA_STOP,
    ],
  },
});

Python

from cdklabs.generative_ai_cdk_constructs import (
    opensearchserverless,
    opensearch_vectorindex,
)

vectorCollection = opensearchserverless.VectorCollection(self, "VectorCollection")

vectorIndex = opensearch_vectorindex.VectorIndex(self, "VectorIndex",
    vector_dimensions= 1536,
    collection=vectorCollection,
    index_name='bedrock-knowledge-base-default-index',
    vector_field='bedrock-knowledge-base-default-vector',
    mappings= [
        opensearch_vectorindex.MetadataManagementFieldProps(
            mapping_field='AMAZON_BEDROCK_TEXT_CHUNK',
            data_type='text',
            filterable=True
        ),
        opensearch_vectorindex.MetadataManagementFieldProps(
            mapping_field='AMAZON_BEDROCK_METADATA',
            data_type='text',
            filterable=False
        )
    ],
    analyzer=opensearchserverless.AnalyzerProps(
        character_filters=[opensearchserverless.CharacterFilterType.ICU_NORMALIZER],
        tokenizer=opensearchserverless.TokenizerType.KUROMOJI_TOKENIZER,
        token_filters=[
            opensearchserverless.TokenFilterType.KUROMOJI_BASEFORM,
            opensearchserverless.TokenFilterType.JA_STOP,
        ],
    )
)