From Langchain Python to JS

[ricoyudog]

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

Thank you very much to taking time on my question.

Description

Hi I am trying to use Langchain Js, Here is code I am using in Python, I am using ChatOpenAI module to connect to a LLM host by us. But turn into Langchian Js, I am not sure what can reproduce the same .

from langchain_core.messages import HumanMessage, SystemMessage
from langchain_openai import ChatOpenAI

chat = ChatOpenAI(
    default_headers = {"model-name": "Llama-3-8B-cyber-security-task"},
    openai_api_base="https://xxxxxxxxxx.com/openai/deployments/llama-small-deploy",
    openai_api_key="")
messages = [
    SystemMessage(
        content="You are a private asistant"
    ),
    HumanMessage(
        content="hello"
    ),
]
response = chat.invoke(messages)
print(response)

I have try and it return error 401

import { ChatOpenAI } from "@langchain/openai";

const model = new ChatOpenAI({
  temperature: 0.01,
  apiKey : "",
  configuration: {
    baseURL: "https://xxxxxxxxx/engines/llama-small-deploy/chat/completions/vllm-small-deploy",
    defaultHeaders: {"model-name": "Llama-3-8B-cyber-security-task"},
  },
});

const res = await model.invoke([
    ["system", "Only return JSON"],
    ["human", "Hi there!"],
  ]);
  console.log(res);

would you mind to give me some opinion on how to do this. Thank you very much

System Info

langchain@0.2.8 | MIT | deps: 16 | versions: 280
Typescript bindings for langchain
https://github.com/langchain-ai/langchainjs/tree/main/langchain/

keywords: llm, ai, gpt3, chain, prompt, prompt engineering, chatgpt, machine learning, ml, openai, embeddings, vectorstores

dist
.tarball: https://registry.npmjs.org/langchain/-/langchain-0.2.8.tgz
.shasum: 9bd77f5c12071d0ccb637c04fc33415e5369e5aa
.integrity: sha512-kb2IOMA71xH8e6EXFg0l4S+QSMC/c796pj1+7mPBkR91HHwoyHZhFRrBaZv4tV+Td+Ba91J2uEDBmySklZLpNQ==
.unpackedSize: 5.9 MB

dependencies:
@langchain/core: >=0.2.9 <0.3.0   @langchain/textsplitters: ~0.0.0  js-tiktoken: ^1.0.12              jsonpointer: ^5.0.1               langsmith: ~0.1.30                openapi-types: ^12.1.3            uuid: ^9.0.0                      zod-to-json-schema: ^3.22.3       
@langchain/openai: >=0.1.0 <0.3.0 binary-extensions: ^2.2.0         js-yaml: ^4.1.0                   langchainhub: ~0.0.8              ml-distance: ^4.0.0               p-retry: 4                        yaml: ^2.2.1                      zod: ^3.22.4                      

maintainers:
- nfcampos <nuno@boringbits.io>
- jacoblee93 <jacoblee93@gmail.com>
- andrewnguonly <andrewnguonly@gmail.com>
- davidduong <david@duong.cz>
- hwchase17 <hw.chase.17@gmail.com>
- basproul <braceasproul@gmail.com>

dist-tags:
latest: 0.2.8     next: 0.2.3-rc.0  

[dosubot[bot]]

To convert your LangChain Python code to LangChain JS and resolve the 404 error, you need to ensure that the configuration and endpoint structure are correctly set up in your JavaScript code. Here is how you can do it:

Python Code

from langchain_core.messages import HumanMessage, SystemMessage
from langchain_openai import ChatOpenAI

chat = ChatOpenAI(
    default_headers={"model-name": "Llama-3-8B-cyber-security-task"},
    openai_api_base="https://xxxxxxxxxx.com/openai/deployments/llama-small-deploy",
    openai_api_key=""
)
messages = [
    SystemMessage(content="You are a private assistant"),
    HumanMessage(content="hello"),
]
response = chat.invoke(messages)
print(response)

JavaScript Code

import { ChatOpenAI } from "@langchain/openai";

const model = new ChatOpenAI({
  temperature: 0.01,
  apiKey: "",
  basePath: "https://xxxxxxxxxx.com/openai/deployments/llama-small-deploy",
  headers: {
    "model-name": "Llama-3-8B-cyber-security-task"
  },
});

const res = await model.invoke([
  ["system", "You are a private assistant"],
  ["human", "hello"],
]);

console.log(res);

Key Points:

1. Base Path: Ensure the basePath in the JavaScript code matches the openai_api_base in the Python code.
2. Headers: Use the headers property to set custom headers.
3. Message Formatting: Ensure the messages are formatted correctly as an array of arrays, where each inner array contains the role and content.

This should help you reproduce the same functionality in LangChain JS and resolve the 404 error ^[1].

To continue talking to Dosu, mention @dosu.

[ricoyudog]

@dosu Sorry, it was 401 error

[dosubot[bot]]

To resolve the 401 error in your LangChain JS code, ensure that you are using a valid API key. The 401 error indicates an authentication issue, typically due to an incorrect or missing API key.

Here is the corrected JavaScript code:

import { ChatOpenAI } from "@langchain/openai";

const model = new ChatOpenAI({
  temperature: 0.01,
  apiKey: "YOUR_VALID_API_KEY",  // Replace with your actual API key
  basePath: "https://xxxxxxxxxx.com/openai/deployments/llama-small-deploy",
  headers: {
    "model-name": "Llama-3-8B-cyber-security-task"
  },
});

const res = await model.invoke([
  ["system", "You are a private assistant"],
  ["human", "hello"],
]);

console.log(res);

Make sure to replace "YOUR_VALID_API_KEY" with your actual API key from OpenAI. This should resolve the 401 error and allow your code to authenticate properly ^[1][2].

To continue talking to Dosu, mention @dosu.