|
| 1 | +import { ICommonObject, INode, INodeData, INodeOutputsValue, INodeParams } from '../../../src/Interface' |
| 2 | +import { ZepVectorStore, IZepConfig } from 'langchain/vectorstores/zep' |
| 3 | +import { Embeddings } from 'langchain/embeddings/base' |
| 4 | +import { Document } from 'langchain/document' |
| 5 | +import { getBaseClasses, getCredentialData, getCredentialParam } from '../../../src/utils' |
| 6 | +import { IDocument, ZepClient } from '@getzep/zep-js' |
| 7 | + |
| 8 | +class Zep_Existing_VectorStores implements INode { |
| 9 | + label: string |
| 10 | + name: string |
| 11 | + version: number |
| 12 | + description: string |
| 13 | + type: string |
| 14 | + icon: string |
| 15 | + category: string |
| 16 | + baseClasses: string[] |
| 17 | + inputs: INodeParams[] |
| 18 | + credential: INodeParams |
| 19 | + outputs: INodeOutputsValue[] |
| 20 | + |
| 21 | + constructor() { |
| 22 | + this.label = 'Zep Load Existing Index' |
| 23 | + this.name = 'zepExistingIndex' |
| 24 | + this.version = 1.0 |
| 25 | + this.type = 'Zep' |
| 26 | + this.icon = 'zep.png' |
| 27 | + this.category = 'Vector Stores' |
| 28 | + this.description = 'Load existing index from Zep (i.e: Document has been upserted)' |
| 29 | + this.baseClasses = [this.type, 'VectorStoreRetriever', 'BaseRetriever'] |
| 30 | + this.credential = { |
| 31 | + label: 'Connect Credential', |
| 32 | + name: 'credential', |
| 33 | + type: 'credential', |
| 34 | + optional: true, |
| 35 | + description: 'Configure JWT authentication on your Zep instance (Optional)', |
| 36 | + credentialNames: ['zepMemoryApi'] |
| 37 | + } |
| 38 | + this.inputs = [ |
| 39 | + { |
| 40 | + label: 'Embeddings', |
| 41 | + name: 'embeddings', |
| 42 | + type: 'Embeddings' |
| 43 | + }, |
| 44 | + { |
| 45 | + label: 'Base URL', |
| 46 | + name: 'baseURL', |
| 47 | + type: 'string', |
| 48 | + default: 'http://127.0.0.1:8000' |
| 49 | + }, |
| 50 | + { |
| 51 | + label: 'Zep Collection', |
| 52 | + name: 'zepCollection', |
| 53 | + type: 'string', |
| 54 | + placeholder: 'my-first-collection' |
| 55 | + }, |
| 56 | + { |
| 57 | + label: 'Zep Metadata Filter', |
| 58 | + name: 'zepMetadataFilter', |
| 59 | + type: 'json', |
| 60 | + optional: true, |
| 61 | + additionalParams: true |
| 62 | + }, |
| 63 | + { |
| 64 | + label: 'Embedding Dimension', |
| 65 | + name: 'dimension', |
| 66 | + type: 'number', |
| 67 | + default: 1536, |
| 68 | + additionalParams: true |
| 69 | + }, |
| 70 | + { |
| 71 | + label: 'Top K', |
| 72 | + name: 'topK', |
| 73 | + description: 'Number of top results to fetch. Default to 4', |
| 74 | + placeholder: '4', |
| 75 | + type: 'number', |
| 76 | + additionalParams: true, |
| 77 | + optional: true |
| 78 | + } |
| 79 | + ] |
| 80 | + this.outputs = [ |
| 81 | + { |
| 82 | + label: 'Pinecone Retriever', |
| 83 | + name: 'retriever', |
| 84 | + baseClasses: this.baseClasses |
| 85 | + }, |
| 86 | + { |
| 87 | + label: 'Pinecone Vector Store', |
| 88 | + name: 'vectorStore', |
| 89 | + baseClasses: [this.type, ...getBaseClasses(ZepVectorStore)] |
| 90 | + } |
| 91 | + ] |
| 92 | + } |
| 93 | + |
| 94 | + async init(nodeData: INodeData, _: string, options: ICommonObject): Promise<any> { |
| 95 | + const baseURL = nodeData.inputs?.baseURL as string |
| 96 | + const zepCollection = nodeData.inputs?.zepCollection as string |
| 97 | + const zepMetadataFilter = nodeData.inputs?.zepMetadataFilter |
| 98 | + const dimension = nodeData.inputs?.dimension as number |
| 99 | + const embeddings = nodeData.inputs?.embeddings as Embeddings |
| 100 | + const output = nodeData.outputs?.output as string |
| 101 | + const topK = nodeData.inputs?.topK as string |
| 102 | + const k = topK ? parseFloat(topK) : 4 |
| 103 | + |
| 104 | + const credentialData = await getCredentialData(nodeData.credential ?? '', options) |
| 105 | + const apiKey = getCredentialParam('apiKey', credentialData, nodeData) |
| 106 | + |
| 107 | + const zepConfig: IZepConfig & Partial<ZepFilter> = { |
| 108 | + apiUrl: baseURL, |
| 109 | + collectionName: zepCollection, |
| 110 | + embeddingDimensions: dimension, |
| 111 | + isAutoEmbedded: false |
| 112 | + } |
| 113 | + if (apiKey) zepConfig.apiKey = apiKey |
| 114 | + if (zepMetadataFilter) { |
| 115 | + const metadatafilter = typeof zepMetadataFilter === 'object' ? zepMetadataFilter : JSON.parse(zepMetadataFilter) |
| 116 | + zepConfig.filter = metadatafilter |
| 117 | + } |
| 118 | + |
| 119 | + const vectorStore = await ZepExistingVS.fromExistingIndex(embeddings, zepConfig) |
| 120 | + |
| 121 | + if (output === 'retriever') { |
| 122 | + const retriever = vectorStore.asRetriever(k) |
| 123 | + return retriever |
| 124 | + } else if (output === 'vectorStore') { |
| 125 | + ;(vectorStore as any).k = k |
| 126 | + return vectorStore |
| 127 | + } |
| 128 | + return vectorStore |
| 129 | + } |
| 130 | +} |
| 131 | + |
| 132 | +interface ZepFilter { |
| 133 | + filter: Record<string, any> |
| 134 | +} |
| 135 | + |
| 136 | +function zepDocsToDocumentsAndScore(results: IDocument[]): [Document, number][] { |
| 137 | + return results.map((d) => [ |
| 138 | + new Document({ |
| 139 | + pageContent: d.content, |
| 140 | + metadata: d.metadata |
| 141 | + }), |
| 142 | + d.score ? d.score : 0 |
| 143 | + ]) |
| 144 | +} |
| 145 | + |
| 146 | +function assignMetadata(value: string | Record<string, unknown> | object | undefined): Record<string, unknown> | undefined { |
| 147 | + if (typeof value === 'object' && value !== null) { |
| 148 | + return value as Record<string, unknown> |
| 149 | + } |
| 150 | + if (value !== undefined) { |
| 151 | + console.warn('Metadata filters must be an object, Record, or undefined.') |
| 152 | + } |
| 153 | + return undefined |
| 154 | +} |
| 155 | + |
| 156 | +class ZepExistingVS extends ZepVectorStore { |
| 157 | + filter?: Record<string, any> |
| 158 | + args?: IZepConfig & Partial<ZepFilter> |
| 159 | + |
| 160 | + constructor(embeddings: Embeddings, args: IZepConfig & Partial<ZepFilter>) { |
| 161 | + super(embeddings, args) |
| 162 | + this.filter = args.filter |
| 163 | + this.args = args |
| 164 | + } |
| 165 | + |
| 166 | + async initalizeCollection(args: IZepConfig & Partial<ZepFilter>) { |
| 167 | + this.client = await ZepClient.init(args.apiUrl, args.apiKey) |
| 168 | + try { |
| 169 | + this.collection = await this.client.document.getCollection(args.collectionName) |
| 170 | + } catch (err) { |
| 171 | + if (err instanceof Error) { |
| 172 | + if (err.name === 'NotFoundError') { |
| 173 | + await this.createNewCollection(args) |
| 174 | + } else { |
| 175 | + throw err |
| 176 | + } |
| 177 | + } |
| 178 | + } |
| 179 | + } |
| 180 | + |
| 181 | + async createNewCollection(args: IZepConfig & Partial<ZepFilter>) { |
| 182 | + if (!args.embeddingDimensions) { |
| 183 | + throw new Error( |
| 184 | + `Collection ${args.collectionName} not found. You can create a new Collection by providing embeddingDimensions.` |
| 185 | + ) |
| 186 | + } |
| 187 | + |
| 188 | + this.collection = await this.client.document.addCollection({ |
| 189 | + name: args.collectionName, |
| 190 | + description: args.description, |
| 191 | + metadata: args.metadata, |
| 192 | + embeddingDimensions: args.embeddingDimensions, |
| 193 | + isAutoEmbedded: false |
| 194 | + }) |
| 195 | + } |
| 196 | + |
| 197 | + async similaritySearchVectorWithScore( |
| 198 | + query: number[], |
| 199 | + k: number, |
| 200 | + filter?: Record<string, unknown> | undefined |
| 201 | + ): Promise<[Document, number][]> { |
| 202 | + if (filter && this.filter) { |
| 203 | + throw new Error('cannot provide both `filter` and `this.filter`') |
| 204 | + } |
| 205 | + const _filters = filter ?? this.filter |
| 206 | + const ANDFilters = [] |
| 207 | + for (const filterKey in _filters) { |
| 208 | + let filterVal = _filters[filterKey] |
| 209 | + if (typeof filterVal === 'string') filterVal = `"${filterVal}"` |
| 210 | + ANDFilters.push({ jsonpath: `$[*] ? (@.${filterKey} == ${filterVal})` }) |
| 211 | + } |
| 212 | + const newfilter = { |
| 213 | + where: { and: ANDFilters } |
| 214 | + } |
| 215 | + await this.initalizeCollection(this.args!).catch((err) => { |
| 216 | + console.error('Error initializing collection:', err) |
| 217 | + throw err |
| 218 | + }) |
| 219 | + const results = await this.collection.search( |
| 220 | + { |
| 221 | + embedding: new Float32Array(query), |
| 222 | + metadata: assignMetadata(newfilter) |
| 223 | + }, |
| 224 | + k |
| 225 | + ) |
| 226 | + return zepDocsToDocumentsAndScore(results) |
| 227 | + } |
| 228 | + |
| 229 | + static async fromExistingIndex(embeddings: Embeddings, dbConfig: IZepConfig & Partial<ZepFilter>): Promise<ZepVectorStore> { |
| 230 | + const instance = new this(embeddings, dbConfig) |
| 231 | + return instance |
| 232 | + } |
| 233 | +} |
| 234 | + |
| 235 | +module.exports = { nodeClass: Zep_Existing_VectorStores } |
0 commit comments