langchain-py-js-tutorial

A small repo for a tutorial I wrote on using LangChain's python library from a JS app.

Overview

This repo was created for a blog post I wrote on using the python langchain library from a Javascript app. The code is organized in two folders. The flask server code lives in langchain-flask-server and the JS test file and data lives in js-test. The flask server has one POST endpoint that accepts json and a query, and then uses LangChain's pandas_dataframe_agent to answer the query.

🔗 Blog Post URL

Prerequisites

You must have the following installed on your computer:

python
pip
virtualenv
node version 17.5 or greater

Installation

Clone the Repo
Create the virtual environment

python3 -m venv .venv

Activate the virtual environment

. .venv/bin/activate

Install the required packages:

pip install Flask
pip install pandas
pip install langchain
pip install python-dotenv

Running the flask server

Open a new terminal window for the langchain-flask-server
Run the server

flask run

You'll know it's working if you navigate to http://127.0.0.1:5000/ and see "I'm building something cool today!"

In order for the API to execute the langchain POST request, you'll have to create a .env file and add your OpenAI API key.

(more on python-dotenv)

Using the API

Open a terminal window for the js test file
Run the code

node api-test