MemoryVectorStore does not support Document id property

[dschenkelman]

Checked other resources

• I added a very descriptive title to this issue.
• I searched the LangChain.js documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain.js rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

The following code:

import { MemoryVectorStore } from "langchain/vectorstores/memory";
import { Document } from "@langchain/core/documents";
import { OpenAIEmbeddings } from "@langchain/openai";
import dotenv from "dotenv"; 

dotenv.config();

const embeddings = new OpenAIEmbeddings();

const docs = [
  new Document({
        metadata: {id: '1'},
        pageContent: 'Document related to test',
        id: '1'
  })
];


const ms =  MemoryVectorStore.fromDocuments(
  docs,
  embeddings
).then(ms => {
    ms.asRetriever().invoke('test').then((docs) => console.log(docs));
});

Prints

[
  Document {
    pageContent: 'Document related to test',
    metadata: { id: '1' },
    id: undefined
  }
]

Note that id is undefined

Error Message and Stack Trace (if applicable)

No response

Description

When storing Documents with id field in MemoryVectorStore and then retrieving, the id is not retrieved.

System Info

All

[dosubot]

@dschenkelman Hello there! I'm here to assist you with any bugs, questions, or contributions you have regarding this repository.

The id field of a Document becomes undefined when storing and retrieving it using MemoryVectorStore in LangChain.js because the idKey used to store and retrieve the document IDs is set to "doc_id" by default. If the documents do not have a metadata field with this key, the IDs will not be correctly stored or retrieved.

To fix this, ensure that the documents have the correct metadata field for the ID, or set the idKey to match the metadata field used in your documents. For example, if your documents use "id" as the metadata key for IDs, initialize the MultiVectorRetriever with idKey: "id":

const retriever = new MultiVectorRetriever({
  vectorstore: new MemoryVectorStore(),
  idKey: "id",
  childK: 20,
  parentK: 5,
});

Additionally, ensure that the metadata field includes the correct ID key when saving context to maintain the association between document IDs and their content ^[1][2][3][4].

To continue talking to Dosu, mention @dosu.