149 lines
4.8 KiB
TypeScript
149 lines
4.8 KiB
TypeScript
import { INode, INodeData, INodeOutputsValue, INodeParams } from '../../../src/Interface'
|
|
import { OpenSearchVectorStore } from 'langchain/vectorstores/opensearch'
|
|
import { Embeddings } from 'langchain/embeddings/base'
|
|
import { Document } from 'langchain/document'
|
|
import { Client, RequestParams } from '@opensearch-project/opensearch'
|
|
import { flatten } from 'lodash'
|
|
import { getBaseClasses } from '../../../src/utils'
|
|
import { buildMetadataTerms } from './core'
|
|
|
|
class OpenSearchUpsert_VectorStores implements INode {
|
|
label: string
|
|
name: string
|
|
version: number
|
|
description: string
|
|
type: string
|
|
icon: string
|
|
category: string
|
|
baseClasses: string[]
|
|
inputs: INodeParams[]
|
|
outputs: INodeOutputsValue[]
|
|
|
|
constructor() {
|
|
this.label = 'OpenSearch Upsert Document'
|
|
this.name = 'openSearchUpsertDocument'
|
|
this.version = 1.0
|
|
this.type = 'OpenSearch'
|
|
this.icon = 'opensearch.png'
|
|
this.category = 'Vector Stores'
|
|
this.description = 'Upsert documents to OpenSearch'
|
|
this.baseClasses = [this.type, 'VectorStoreRetriever', 'BaseRetriever']
|
|
this.inputs = [
|
|
{
|
|
label: 'Document',
|
|
name: 'document',
|
|
type: 'Document',
|
|
list: true
|
|
},
|
|
{
|
|
label: 'Embeddings',
|
|
name: 'embeddings',
|
|
type: 'Embeddings'
|
|
},
|
|
{
|
|
label: 'OpenSearch URL',
|
|
name: 'opensearchURL',
|
|
type: 'string',
|
|
placeholder: 'http://127.0.0.1:9200'
|
|
},
|
|
{
|
|
label: 'Index Name',
|
|
name: 'indexName',
|
|
type: 'string'
|
|
},
|
|
{
|
|
label: 'Top K',
|
|
name: 'topK',
|
|
description: 'Number of top results to fetch. Default to 4',
|
|
placeholder: '4',
|
|
type: 'number',
|
|
additionalParams: true,
|
|
optional: true
|
|
}
|
|
]
|
|
this.outputs = [
|
|
{
|
|
label: 'OpenSearch Retriever',
|
|
name: 'retriever',
|
|
baseClasses: this.baseClasses
|
|
},
|
|
{
|
|
label: 'OpenSearch Vector Store',
|
|
name: 'vectorStore',
|
|
baseClasses: [this.type, ...getBaseClasses(OpenSearchVectorStore)]
|
|
}
|
|
]
|
|
}
|
|
|
|
async init(nodeData: INodeData): Promise<any> {
|
|
const docs = nodeData.inputs?.document as Document[]
|
|
const embeddings = nodeData.inputs?.embeddings as Embeddings
|
|
const opensearchURL = nodeData.inputs?.opensearchURL as string
|
|
const indexName = nodeData.inputs?.indexName as string
|
|
const output = nodeData.outputs?.output as string
|
|
const topK = nodeData.inputs?.topK as string
|
|
const k = topK ? parseFloat(topK) : 4
|
|
|
|
const flattenDocs = docs && docs.length ? flatten(docs) : []
|
|
const finalDocs = []
|
|
for (let i = 0; i < flattenDocs.length; i += 1) {
|
|
finalDocs.push(new Document(flattenDocs[i]))
|
|
}
|
|
|
|
const client = new Client({
|
|
nodes: [opensearchURL]
|
|
})
|
|
|
|
const vectorStore = await OpenSearchVectorStore.fromDocuments(finalDocs, embeddings, {
|
|
client,
|
|
indexName
|
|
})
|
|
|
|
vectorStore.similaritySearchVectorWithScore = async (
|
|
query: number[],
|
|
k: number,
|
|
filter?: object | undefined
|
|
): Promise<[Document, number][]> => {
|
|
const search: RequestParams.Search = {
|
|
index: indexName,
|
|
body: {
|
|
query: {
|
|
bool: {
|
|
filter: { bool: { must: buildMetadataTerms(filter) } },
|
|
must: [
|
|
{
|
|
knn: {
|
|
embedding: { vector: query, k }
|
|
}
|
|
}
|
|
]
|
|
}
|
|
},
|
|
size: k
|
|
}
|
|
}
|
|
|
|
const { body } = await client.search(search)
|
|
|
|
return body.hits.hits.map((hit: any) => [
|
|
new Document({
|
|
pageContent: hit._source.text,
|
|
metadata: hit._source.metadata
|
|
}),
|
|
hit._score
|
|
])
|
|
}
|
|
|
|
if (output === 'retriever') {
|
|
const retriever = vectorStore.asRetriever(k)
|
|
return retriever
|
|
} else if (output === 'vectorStore') {
|
|
;(vectorStore as any).k = k
|
|
return vectorStore
|
|
}
|
|
return vectorStore
|
|
}
|
|
}
|
|
|
|
module.exports = { nodeClass: OpenSearchUpsert_VectorStores }
|