Flowise/packages/components/nodes/vectorstores/Weaviate/Weaviate.ts

311 lines
12 KiB
TypeScript

import { flatten } from 'lodash'
import weaviate, { WeaviateClient, ApiKey } from 'weaviate-ts-client'
import { WeaviateLibArgs, WeaviateStore } from '@langchain/weaviate'
import { Document } from '@langchain/core/documents'
import { Embeddings } from '@langchain/core/embeddings'
import { ICommonObject, INode, INodeData, INodeOutputsValue, INodeParams, IndexingResult } from '../../../src/Interface'
import { getBaseClasses, getCredentialData, getCredentialParam, normalizeKeysRecursively, parseJsonBody } from '../../../src/utils'
import { addMMRInputParams, resolveVectorStoreOrRetriever } from '../VectorStoreUtils'
import { index } from '../../../src/indexing'
import { VectorStore } from '@langchain/core/vectorstores'
class Weaviate_VectorStores implements INode {
label: string
name: string
version: number
description: string
type: string
icon: string
category: string
badge: string
baseClasses: string[]
inputs: INodeParams[]
credential: INodeParams
outputs: INodeOutputsValue[]
constructor() {
this.label = 'Weaviate'
this.name = 'weaviate'
this.version = 4.0
this.type = 'Weaviate'
this.icon = 'weaviate.png'
this.category = 'Vector Stores'
this.description =
'Upsert embedded data and perform similarity or mmr search using Weaviate, a scalable open-source vector database'
this.baseClasses = [this.type, 'VectorStoreRetriever', 'BaseRetriever']
this.credential = {
label: 'Connect Credential',
name: 'credential',
type: 'credential',
description: 'Only needed when using Weaviate cloud hosted',
optional: true,
credentialNames: ['weaviateApi']
}
this.inputs = [
{
label: 'Document',
name: 'document',
type: 'Document',
list: true,
optional: true
},
{
label: 'Embeddings',
name: 'embeddings',
type: 'Embeddings'
},
{
label: 'Record Manager',
name: 'recordManager',
type: 'RecordManager',
description: 'Keep track of the record to prevent duplication',
optional: true
},
{
label: 'Weaviate Scheme',
name: 'weaviateScheme',
type: 'options',
default: 'https',
options: [
{
label: 'https',
name: 'https'
},
{
label: 'http',
name: 'http'
}
]
},
{
label: 'Weaviate Host',
name: 'weaviateHost',
type: 'string',
placeholder: 'localhost:8080'
},
{
label: 'Weaviate Index',
name: 'weaviateIndex',
type: 'string',
placeholder: 'Test'
},
{
label: 'Weaviate Text Key',
name: 'weaviateTextKey',
type: 'string',
placeholder: 'text',
optional: true,
additionalParams: true
},
{
label: 'Weaviate Metadata Keys',
name: 'weaviateMetadataKeys',
type: 'string',
rows: 4,
placeholder: `["foo"]`,
optional: true,
additionalParams: true
},
{
label: 'Top K',
name: 'topK',
description: 'Number of top results to fetch. Default to 4',
placeholder: '4',
type: 'number',
additionalParams: true,
optional: true
},
{
label: 'Weaviate Search Filter',
name: 'weaviateFilter',
type: 'json',
additionalParams: true,
optional: true,
acceptVariable: true
}
]
addMMRInputParams(this.inputs)
this.inputs.push({
label: 'Alpha (for Hybrid Search)',
name: 'alpha',
description:
'Number between 0 and 1 that determines the weighting of keyword (BM25) portion of the hybrid search. A value of 1 is a pure vector search, while 0 is a pure keyword search.',
placeholder: '1',
type: 'number',
additionalParams: true,
optional: true
})
this.outputs = [
{
label: 'Weaviate Retriever',
name: 'retriever',
baseClasses: this.baseClasses
},
{
label: 'Weaviate Vector Store',
name: 'vectorStore',
baseClasses: [this.type, ...getBaseClasses(WeaviateStore)]
}
]
}
//@ts-ignore
vectorStoreMethods = {
async upsert(nodeData: INodeData, options: ICommonObject): Promise<Partial<IndexingResult>> {
const weaviateScheme = nodeData.inputs?.weaviateScheme as string
const weaviateHost = nodeData.inputs?.weaviateHost as string
const weaviateIndex = nodeData.inputs?.weaviateIndex as string
const weaviateTextKey = nodeData.inputs?.weaviateTextKey as string
const weaviateMetadataKeys = nodeData.inputs?.weaviateMetadataKeys as string
const docs = nodeData.inputs?.document as Document[]
const embeddings = nodeData.inputs?.embeddings as Embeddings
const recordManager = nodeData.inputs?.recordManager
const credentialData = await getCredentialData(nodeData.credential ?? '', options)
const weaviateApiKey = getCredentialParam('weaviateApiKey', credentialData, nodeData)
const clientConfig: any = {
scheme: weaviateScheme,
host: weaviateHost
}
if (weaviateApiKey) clientConfig.apiKey = new ApiKey(weaviateApiKey)
const client: WeaviateClient = weaviate.client(clientConfig)
const flattenDocs = docs && docs.length ? flatten(docs) : []
const finalDocs = []
for (let i = 0; i < flattenDocs.length; i += 1) {
if (flattenDocs[i] && flattenDocs[i].pageContent) {
const doc = { ...flattenDocs[i] }
if (doc.metadata) {
doc.metadata = normalizeKeysRecursively(doc.metadata)
}
finalDocs.push(new Document(doc))
}
}
const obj: WeaviateLibArgs = {
//@ts-ignore
client,
indexName: weaviateIndex
}
if (weaviateTextKey) obj.textKey = weaviateTextKey
if (weaviateMetadataKeys) obj.metadataKeys = JSON.parse(weaviateMetadataKeys.replace(/\s/g, ''))
try {
if (recordManager) {
const vectorStore = (await WeaviateStore.fromExistingIndex(embeddings, obj)) as unknown as VectorStore
await recordManager.createSchema()
const res = await index({
docsSource: finalDocs,
recordManager,
vectorStore,
options: {
cleanup: recordManager?.cleanup,
sourceIdKey: recordManager?.sourceIdKey ?? 'source',
vectorStoreName: weaviateTextKey ? weaviateIndex + '_' + weaviateTextKey : weaviateIndex
}
})
return res
} else {
await WeaviateStore.fromDocuments(finalDocs, embeddings, obj)
return { numAdded: finalDocs.length, addedDocs: finalDocs }
}
} catch (e) {
throw new Error(e)
}
},
async delete(nodeData: INodeData, ids: string[], options: ICommonObject): Promise<void> {
const weaviateScheme = nodeData.inputs?.weaviateScheme as string
const weaviateHost = nodeData.inputs?.weaviateHost as string
const weaviateIndex = nodeData.inputs?.weaviateIndex as string
const weaviateTextKey = nodeData.inputs?.weaviateTextKey as string
const weaviateMetadataKeys = nodeData.inputs?.weaviateMetadataKeys as string
const embeddings = nodeData.inputs?.embeddings as Embeddings
const recordManager = nodeData.inputs?.recordManager
const credentialData = await getCredentialData(nodeData.credential ?? '', options)
const weaviateApiKey = getCredentialParam('weaviateApiKey', credentialData, nodeData)
const clientConfig: any = {
scheme: weaviateScheme,
host: weaviateHost
}
if (weaviateApiKey) clientConfig.apiKey = new ApiKey(weaviateApiKey)
const client: WeaviateClient = weaviate.client(clientConfig)
const obj: WeaviateLibArgs = {
//@ts-ignore
client,
indexName: weaviateIndex
}
if (weaviateTextKey) obj.textKey = weaviateTextKey
if (weaviateMetadataKeys) obj.metadataKeys = JSON.parse(weaviateMetadataKeys.replace(/\s/g, ''))
const weaviateStore = new WeaviateStore(embeddings, obj)
try {
if (recordManager) {
const vectorStoreName = weaviateTextKey ? weaviateIndex + '_' + weaviateTextKey : weaviateIndex
await recordManager.createSchema()
;(recordManager as any).namespace = (recordManager as any).namespace + '_' + vectorStoreName
const filterKeys: ICommonObject = {}
if (options.docId) {
filterKeys.docId = options.docId
}
const keys: string[] = await recordManager.listKeys(filterKeys)
await weaviateStore.delete({ ids: keys })
await recordManager.deleteKeys(keys)
} else {
await weaviateStore.delete({ ids })
}
} catch (e) {
throw new Error(e)
}
}
}
async init(nodeData: INodeData, _: string, options: ICommonObject): Promise<any> {
const weaviateScheme = nodeData.inputs?.weaviateScheme as string
const weaviateHost = nodeData.inputs?.weaviateHost as string
const weaviateIndex = nodeData.inputs?.weaviateIndex as string
const weaviateTextKey = nodeData.inputs?.weaviateTextKey as string
const weaviateMetadataKeys = nodeData.inputs?.weaviateMetadataKeys as string
const embeddings = nodeData.inputs?.embeddings as Embeddings
let weaviateFilter = nodeData.inputs?.weaviateFilter
const credentialData = await getCredentialData(nodeData.credential ?? '', options)
const weaviateApiKey = getCredentialParam('weaviateApiKey', credentialData, nodeData)
const clientConfig: any = {
scheme: weaviateScheme,
host: weaviateHost
}
if (weaviateApiKey) clientConfig.apiKey = new ApiKey(weaviateApiKey)
const client: WeaviateClient = weaviate.client(clientConfig)
const obj: WeaviateLibArgs = {
//@ts-ignore
client,
indexName: weaviateIndex
}
if (weaviateTextKey) obj.textKey = weaviateTextKey
if (weaviateMetadataKeys) obj.metadataKeys = JSON.parse(weaviateMetadataKeys.replace(/\s/g, ''))
if (weaviateFilter) {
weaviateFilter = typeof weaviateFilter === 'object' ? weaviateFilter : parseJsonBody(weaviateFilter)
}
const vectorStore = (await WeaviateStore.fromExistingIndex(embeddings, obj)) as unknown as VectorStore
return resolveVectorStoreOrRetriever(nodeData, vectorStore, weaviateFilter)
}
}
module.exports = { nodeClass: Weaviate_VectorStores }