langchain-js-restorable-memory-vectorstore

a restorable memory vector store for LangChainJS that works in browsers

Overview

This simple class adds a fromJSON method to the MemoryVectorStore from LangChainJS in order to be able to preserve and restore any vector store contents while still be browser-compatible (as the typical save and load methods of vector stores assume to run within a Node.js environment).

If you replace your MemoryVectorStore with a RestorableMemoryVectorStore, you can now use the original toJSON to serialize your vector store into a JSON object (which may then be persisted using a browser's IndexedDB or similar) and later restore it using fromJSON.

Installation

langchain-js-restorable-memory-vectorstore comes as an ECMAScript module (ESM). You may either install the module using npm (or similar) if you plan to use a bundler:

npm install langchain-js-restorable-memory-vectorstore

Or you may dynamically import it using an import expression

const { RestorableMemoryVectorStore } = await import "https://rozek.github.io/langchain-js-restorable-memory-vectorstore/dist/RestorableMemoryVectorStore.js"

Usage in Node.js or Browser Environments

Assuming that you have installed the module, the RestorableMemoryVectorStore can be used as follows

  import { RestorableMemoryVectorStore } from 'langchain-js-restorable-memory-vectorstore'
  import { OpenAIEmbeddings } from '@langchain/openai'
  
/**** create a vector store with documents ****/

  const Embedder = new OpenAIEmbeddings()
  const Store    = new RestorableMemoryVectorStore(Embedder)
  
/**** add some documents ****/

  await Store.addDocuments([
    { pageContent: 'Hello world', metadata: { source:'greeting' } },
    { pageContent: 'Bye world',   metadata: { source:'farewell' } },
  ])
  
/**** serialize the vector store to JSON ****/

  const Serialization = JSON.stringify(Store) // invokes "toJSON"
    
/**** later, restore the vector store ****/

  const restoredStore = await RestorableMemoryVectorStore.fromJSON(
    Serialization, Embedder
  )
      
/**** the restored store is now ready to be used ****/

  const Results = await restoredStore.similaritySearch('hello', 1)
  console.log(Results)

Usage within Svelte

For Svelte, it is recommended to import the package in a module context:

<script context="module">
  import { RestorableMemoryVectorStore } from 'langchain-js-restorable-memory-vectorstore'
  import { OpenAIEmbeddings } from '@langchain/openai'
</script>

<script>
  const Embedder = new OpenAIEmbeddings()
  const Store    = new RestorableMemoryVectorStore(Embedder)
  
/**** add some documents ****/

  await Store.addDocuments([
    { pageContent: 'Hello world', metadata: { source:'greeting' } },
    { pageContent: 'Bye world',   metadata: { source:'farewell' } },
  ])
  
/**** serialize the vector store to JSON ****/

  const Serialization = JSON.stringify(Store) // invokes "toJSON"
    
/**** later, restore the vector store ****/

  const restoredStore = await RestorableMemoryVectorStore.fromJSON(
    Serialization, Embedder
  )
      
/**** the restored store is now ready to be used ****/

  const Results = await restoredStore.similaritySearch('hello', 1)
  console.log(Results)
</script>

API Reference

RestorableMemoryVectorStore

extends MemoryVectorStore with a fromJSON method

Constructor

new RestorableMemoryVectorStore(Embedder:EmbeddingsInterface)

• Embedder: An implementation of the LangChain EmbeddingsInterface to use for creating embeddings

Static Methods

fromJSON

static async fromJSON(
  JSONString:string, Embedder:EmbeddingsInterface
): Promise<RestorableMemoryVectorStore>

restores a vector store from a serialized JSON string.

• JSONString: the serialized vector store in JSON format.
• Embedder: an implementation of the LangChain EmbeddingsInterface to use for creating embeddings
• returns: a new RestorableMemoryVectorStore instance populated with the serialized data

Build Instructions

You may easily build this package yourself.

Just install NPM according to the instructions for your platform and follow these steps:

1. either clone this repository using git or download a ZIP archive with its contents to your disk and unpack it there
2. open a shell and navigate to the root directory of this repository
3. run npm install in order to install the complete build environment
4. execute npm run build to create a new build

You may also look into the author's build-configuration-study for a general description of his build environment.

Test Instructions

langchain-js-restorable-memory-vectorstore comes with a few tests. Just use

npm run test

to run them and get a report on the console.

License

MIT License