"Cannot read properties of undefined (reading 'callKeys')"

[Elindo586]

Why am I getting this error and how can I fix it?

"what day is today?"
[chain/start] [1:chain:SimpleSequentialChain] Entering Chain run with input: {
  "input": {
    "input": "\"what day is today?\""
  }
}
[chain/start] [1:chain:SimpleSequentialChain > 2:chain:LLMChain] Entering Chain run with input: {
  "input": {
    "input": "\"what day is today?\""
  }
}
[chain/error] [1:chain:SimpleSequentialChain > 2:chain:LLMChain] [1ms] Chain run errored with error: "Cannot read properties of undefined (reading 'callKeys')" 
[chain/error] [1:chain:SimpleSequentialChain] [2ms] Chain run errored with error: "Cannot read properties of undefined (reading 'callKeys')"
 ⨯ TypeError: Cannot read properties of undefined (reading 'callKeys')
    at LLMChain._call (file:///C:/Users/Edgar/Desktop/Stuff/Acode/nextprojectd/node_modules/langchain/dist/chains/llm_chain.js:114:36)
    at LLMChain.call (file:///C:/Users/Edgar/Desktop/Stuff/Acode/nextprojectd/node_modules/langchain/dist/chains/base.js:110:24)
    at async SimpleSequentialChain._call (file:///C:/Users/Edgar/Desktop/Stuff/Acode/nextprojectd/node_modules/langchain/dist/chains/sequential_chain.js:240:22)
    at async SimpleSequentialChain.call (file:///C:/Users/Edgar/Desktop/Stuff/Acode/nextprojectd/node_modules/langchain/dist/chains/base.js:101:28)
    at async SimpleSequentialChain.run (file:///C:/Users/Edgar/Desktop/Stuff/Acode/nextprojectd/node_modules/langchain/dist/chains/base.js:66:30)
    at async useStore (webpack-internal:///(api)/./pages/api/usestore.js:93:24) {
  page: '/api/usestore'
}
 ✓ Compiled /_error in 148ms (514 modules)

The code I have is:

import { TextLoader } from "langchain/document_loaders/fs/text";
import { CharacterTextSplitter } from "langchain/text_splitter";
import { OpenAIEmbeddings } from "langchain/embeddings/openai";
import { FaissStore } from "langchain/vectorstores/faiss";

import { OpenAI } from "langchain/llms/openai";
import { PromptTemplate } from "langchain/prompts";
import {
  SimpleSequentialChain,
  LLMChain,
  RetrievalQAChain,
  loadQAStuffChain,
} from "langchain/chains";

let loader;
let docs;
let splitter;
let documents;
let embeddings1;
let vectorstore;

let template1;
let reviewPromptTemplate1;
let reviewChain1;

let model1;
let model2;
let embeddings;
let vectoreStore;
let vectorChain;
let overallChain;

const useStore = async (req, res) => {
  if (req.method === "POST") {
    loader = new TextLoader("./documents/restaurant.txt");
    docs = await loader.load();
    splitter = new CharacterTextSplitter({
      chunkSize: 200,
      chunkOverlap: 50,
    });
    documents = await splitter.splitDocuments(docs);
    embeddings1 = new OpenAIEmbeddings();
    vectorstore = await FaissStore.fromDocuments(documents, embeddings1);
    await vectorstore.save("./");

    const message = req.body.newText;
    const firstOne = req.body.firstMsg;

    console.log(message);
    if (firstOne) {
      model1 = new OpenAI({
        openAIApiKey: process.env.OPENAI_API_KEY,
        temperature: 0,
      });
      template1 = `
    Be very funny when answering questions. You are a helpful bot that creates a 'thank you' response text.
    If customers are unsatisfied, offer them a real world assistant to talk to.
    You will get a sentiment and subject as input and evaluate.
    
    text: {input}
    `;
      reviewPromptTemplate1 = new PromptTemplate({
        template: template1,
        inputVariables: ["input"],
      });

      reviewChain1 = new LLMChain({
        model1,
        prompt: reviewPromptTemplate1,
      });

      model2 = new OpenAI({
        openAIApiKey: process.env.OPENAI_API_KEY,
        temperature: 0,
      });

      embeddings = new OpenAIEmbeddings();
      vectoreStore = await FaissStore.load("./", embeddings);

      vectorChain = new RetrievalQAChain({
        combineDocumentsChain: loadQAStuffChain(model2),
        retriever: vectoreStore.asRetriever(),
        returnSourceDocuments: false,
      });
    }
    overallChain = new SimpleSequentialChain({
      chains: [reviewChain1, vectorChain],
      verbose: true,
    });

    const result = await overallChain.run({
      input: message,
    });

    console.log(result.text);

    return res.status(200).json({ output: result });
  } else {
    res.status(405).json({ say: "Only POST is allowed" });
  }
};
export default useStore;

[dosubot[bot]]

🤖

Hello,

The error message "Cannot read properties of undefined (reading 'callKeys')" typically indicates that you're trying to access a property or method on an object that is undefined. In this case, it seems like the error is originating from the LLMChain class.

Looking at the stack trace, the error is thrown at this line:

at LLMChain._call (file:///C:/Users/Edgar/Desktop/Stuff/Acode/nextprojectd/node_modules/langchain/dist/chains/llm_chain.js:114:36)

This suggests that the LLMChain instance is trying to access a property or method named 'callKeys' which doesn't exist or is undefined.

Looking at your code, it seems like the LLMChain instances reviewChain1 and vectorChain are being used in the SimpleSequentialChain:

overallChain = new SimpleSequentialChain({
  chains: [reviewChain1, vectorChain],
  verbose: true,
});

However, reviewChain1 and vectorChain are only defined inside the if (firstOne) block. This means that if firstOne is false or undefined, reviewChain1 and vectorChain will also be undefined, which could be causing the error.

To fix this, you could initialize reviewChain1 and vectorChain outside of the if (firstOne) block, or ensure that firstOne is always true when you're running this code.

Here's how you could initialize reviewChain1 and vectorChain outside of the if (firstOne) block:

let reviewChain1 = null;
let vectorChain = null;

// ...

if (firstOne) {
  // ...
  reviewChain1 = new LLMChain({
    model1,
    prompt: reviewPromptTemplate1,
  });

  // ...
  vectorChain = new RetrievalQAChain({
    combineDocumentsChain: loadQAStuffChain(model2),
    retriever: vectoreStore.asRetriever(),
    returnSourceDocuments: false,
  });
}

// ...

Then, in the SimpleSequentialChain, you could check if reviewChain1 and vectorChain are not null before adding them to the chains array:

overallChain = new SimpleSequentialChain({
  chains: [reviewChain1, vectorChain].filter(chain => chain !== null),
  verbose: true,
});

I hope this helps! Let me know if you have any other questions.

Sources

• langchain/src/load/map_keys.ts

---

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.