I tried to implement the vector stores redis demo and got an error? Error: Expected a Runnable, function or object.

[lycpan233]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

'use strict';
require('dotenv').config()
const { ChatOllama } = require("@langchain/community/chat_models/ollama");
const { RecursiveCharacterTextSplitter } = require("langchain/text_splitter");
const { OllamaEmbeddings } = require("@langchain/community/embeddings/ollama");
const { createStuffDocumentsChain } = require("langchain/chains/combine_documents");
const { ChatPromptTemplate } = require("@langchain/core/prompts");
const { createRetrievalChain } = require("langchain/chains/retrieval");
const { TextLoader } = require("langchain/document_loaders/fs/text");
const { StringOutputParser } = require("@langchain/core/output_parsers");

const { createClient } = require("redis");
const { RedisVectorStore } = require("@langchain/redis");


async function main() {
  console.time('---> start');
  const client = createClient({
    url: "redis://localhost:6379",
  });
  await client.connect();

  const redisAddOptions = {
    redisClient: client,
    indexName: "docs",
  }

  const embeddings = new OllamaEmbeddings({
    model: "qwen",
    maxConcurrency: 5,
  });

  const vectorStore = new RedisVectorStore(embeddings, redisAddOptions);
  await vectorStore.dropIndex(true)
  let hasIndex = await vectorStore.checkIndexExists();
  console.log("hasIndex", hasIndex);

  if (!hasIndex) {
    const loader = new TextLoader("docs/a.txt");
    const docs = await loader.load();

    const splitter = new RecursiveCharacterTextSplitter({
      chunkSize: 256,
      chunkOverlap: 0,
    });
    const splitDocs = await splitter.splitDocuments(docs);

    await vectorStore.addDocuments(splitDocs);
  }
  console.timeEnd('---> start');

  const chatModel = new ChatOllama({
    baseUrl: "http://localhost:11434", // Default value
    model: "qwen",
    max_token: 80000,
    topP: 0.9,
    // temperature: 0
  });

  const prompt =
    ChatPromptTemplate.fromTemplate(`You are an assistant for question-answering tasks. Use the following pieces of retrieved context to answer the question. If you don't know the answer, just say that you don't know. Use three sentences maximum and keep the answer concise.

Question: {input} 

Context: {context} 

Answer:`);

  const documentChain = await createStuffDocumentsChain({
    llm: chatModel,
    outputParser: new StringOutputParser(),
    prompt,
  });

  const retriever = vectorStore.asRetriever()
  const chain = await createRetrievalChain({
    retriever,
    documentChain
  });

  const chainRes = await chain.invoke({ input: "Hot Are You？" });
  console.log(chainRes);

  await client.disconnect();
}

main()

Description

error msg

        throw new Error(`Expected a Runnable, function or object.\nInstead got an unsupported type.`);
              ^

Error: Expected a Runnable, function or object.
Instead got an unsupported type.

I found this, but I don't know what to do with it!

System Info

Mac M1
Node v18.19.0

[dosubot[bot]]

Hey there, @lycpan233! Great to see you diving into some adventurous coding again. How's everything on your end?

Based on the error message you're encountering, it seems like there might be an issue with the type of argument being passed to a function or method within the LangChain JS framework. The error message "Expected a Runnable, function or object. Instead got an unsupported type." suggests that a function, method, or constructor is receiving an argument of a type it doesn't support.

From the code snippet you provided, it appears that you're correctly using the LangChain JS framework's API for creating a RedisVectorStore from a set of documents. However, the issue might not be directly visible in the provided code snippet but could stem from one of the following:

1. Misconfiguration or incompatibility within the LangChain JS framework versions: Ensure that the versions of @langchain/core, @langchain/redis, and @langchain/openai you are using are compatible with each other. An update or a mismatch between these packages could lead to unexpected behavior.

2. Unsupported argument type in RedisVectorStore.fromDocuments or another method: The error message suggests that a function, method, or constructor is receiving an argument of a type it doesn't support. While the provided code snippet doesn't show an obvious misuse, the error could be occurring within the RedisVectorStore.fromDocuments method or the OpenAIEmbeddings constructor, depending on how these are implemented in the version of the LangChain JS framework you're using. It's possible that an update or a change in the framework has introduced new requirements or altered the expected types for arguments.

3. External factors not visible in the snippet: The issue might also be related to environment variables, Redis configuration, or other external factors not visible in the provided code snippet. For example, if the REDIS_URL environment variable is not correctly set, it could potentially lead to issues when connecting to Redis, although this would more likely result in a connection error rather than the type error you're seeing.

Given the information provided, the best course of action would be to:

• Verify that all LangChain JS framework packages are up to date and compatible with each other.
• Review the documentation for RedisVectorStore.fromDocuments and OpenAIEmbeddings to ensure that the arguments being passed are expected and supported by the current version of the framework.
• Check for any recent changes or updates in the LangChain JS framework that might have introduced new requirements or breaking changes related to the creation of a RedisVectorStore.

If the issue persists after these checks, please provide more information about the specific circumstances under which the error occurs, such as the exact line of code that triggers the error, the versions of the LangChain JS packages you're using, and any relevant configuration details.

Regarding the similar issue you found in the LangChain JS repository, it seems to be related to a breaking change introduced in version 0.0.144. The last version that didn't introduce breaking changes in vector stores was LangChain 0.0.143. You can find more details about this issue here. However, it's not clear if this issue is directly related to the error you're encountering. It might be worth trying to downgrade to version 0.0.143 to see if that resolves the issue.

I hope this helps! If you have any further questions, feel free to ask.

Sources

• Breaking change between 0.0.112 and 0.0.144 "TypeError: this._vectorstoreType is not a function"
• examples/src/indexes/vector_stores/redis/redis_index_options.ts
• libs/langchain-community/src/vectorstores/redis.ts

About Dosu

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.

[nikolay17538583]

mehh same thing :(

@lycpan233 have you solved this?

[lycpan233]

It's been so long, I forget the details. Maybe I did it the other way.

https://github.com/lycpan233/langchain-rag-node-js/blob/master/index.js#L85