LangChain Tutorial (JS) #1: Intro & Setup

LangChain JS TutorialAging📅 2024-01-10

Tutorial Overview

┌─────────────────────────────────────────────────────────────────────┐
│                         SETUP WORKFLOW                              │
├─────────────────────────────────────────────────────────────────────┤
│                                                                     │
│   1. Prerequisites                                                  │
│      │                                                              │
│      ├──► Install required software                                 │
│      │    • Node.js / Python / etc.                                 │
│      │    • Code editor (VS Code, Cursor)                           │
│      │                                                              │
│      ▼                                                              │
│   2. Configuration                                                  │
│      │                                                              │
│      ├──► Get API keys / credentials                                │
│      ├──► Set environment variables                                 │
│      │                                                              │
│      ▼                                                              │
│   3. Initialize Other Tools                                              │
│      │                                                              │
│      ├──► Run setup commands                                        │
│      ├──► Verify installation                                       │
│      │                                                              │
│      ▼                                                              │
│   4. Ready to Build!                                                │
│                                                                     │
└─────────────────────────────────────────────────────────────────────┘

What You'll Learn

Environment configuration
Dependencies installation
Initial setup steps

Transcript

[00:00] what is Lang chain and why should you learn this framework in 2024 Lang chain is a framework that allows developers to create AI powered applications in both JavaScript and python one of the most important features of Lang chain is its ability to connect our AI applications to external data sources like databases API endpoints text files and pretty much anything you can imagine this series is beginner friendly and you are going to have a lot of fun following along L

[00:32] chain is surprisingly easy to use and within the next few minutes you will have your very first AI chatbot up and running in this series we will first focus on the fundamentals of using L chain and thereafter we will work on a few exciting projects to practice our skills you can find all the code from the series on this gith up repository simply change the branch to a specific lesson to view the source code we will be looking at the JavaScript script version of Lang chain in this series but

[01:02] I am working on a python version of the series as well and you should see a link to that series showing up right now in order to follow along I do recommend that you have a basic understanding of JavaScript you also need nodejs installed which you can download from nodejs.org and you will also need a code editor like vs code open a folder in vs code and then open up the integrated terminal like so let's initiate a note project by running npm init Dy this will

[01:35] create a package.json file in this folder open the package.json file and let's make one small change under main let's add a new property called type and for the value change it to module and that will allow us to use the import syntax when importing packages we can now save this file and close it in order to use Lang chain in our project we need to install a few dependencies IES in the terminal enter npm install Lang chain

[02:05] and since we'll be using the open AI models in this video Let's also install that dependency by typing at Lang chain slop aai and let's run this now that we have our dependencies installed let's move on to the fun part in our project let's create a new file and call it lm. JS since we'll be using models from open AI we need to import the following import chat open AI from at Lang chain slop AI like so we can now instantiate

[02:38] our model by typing const model equals new chat open AI this chat openi class takes in an object as input and one of the properties that we need to populate is the open AI API key so let's go ahead and get our API key from open AI you can get your open I API Key by visiting this URL then click on create new secret key give it a name I'll just call mine

[03:08] something like L chain JS tutorial and then click on create secret key go ahead and copy this value and then add it to this property we will move this API key to an environment variable in a second but from experience you will have use cases where the key needs to be passed into the class directly for instance on one of the projects I worked on for a client we had to dynamically populate this key based on the logged in user now all we have to do to run this model is

[03:39] to call model. invoke and then pass it any text like Hello this invoke method is a promise which we need to await and let's also assign this response to a variable as well and lastly let's console log the response we can run this code in the terminal by typing node and the file name like lmjs let's run this and we can see the response coming back from the AI and we can see in the response that we get this

[04:11] AI message object and in this object we have a property called content which contains the response from the model let's also have a look at a few different ways that we can call this model we already had a look at invoke but we can also change this to batch and batch will allow us to pass multiple proms to the model and get several responses like let's pass in an array with hello and how are you and let's run this by calling node atlm and in the

[04:41] terminal we can see that we now get an array of values back and we received two AI messages one that contains the text hello how can I assist you today and the second message with the content as an AI I do not have feelings which is related to this how are you another way to call this this is to stream the response so let's change it from batch to stream and let's change this prompt back to a single string and let's ask it something like write a poem about AI now what

[05:13] stream will do is return the response in chunks as the response is generated by the AI so I'm just going to comment out this console log for now and then instead let's type 408 and in Brackets const chunk or of response and then let's console log chunk. content and let's run node llm and when I hit enter you will see the response streaming in the terminal and there we go there is one more method

[05:44] that I do want to show you and that is the stream log method what stream log will do is return each of the steps that is being executed by the model so let's simply return the chunk only and let's see what we get and in in the logs we can see a lot of information being returned and this is basically each of those words that are being streamed back to us as an AI message chunk and at the end of the list we can see this value object which contains the final response

[06:15] from the model for the majority of this series we'll simply use the invoke method so I'm going to change this back to invoke and let's add back this comment and let's remove this code let's have a look at improving this code first let's remove move this API key and store it in an environment variable instead so I'm going to cut this value is create a new file called EnV in this file create a property called open aior

[06:45] apore key and then paste in that value the name of this property is important so be sure to use the same name then back in our code we can simply remove this property and what we'll do instead is pull in that value from our environment variables to do this we do need to install another package so in the terminal enter npm install. EnV at the top of our code let's import all

[07:15] ASV from EnV then type Dov doc config this will now make the environment variables available to this code but we don't have to do anything else Lang chain will automatically look in the environment variable file for a property called open AI API key and then assign that to this class and we can test that in the terminal by running node llm and we still get our response back this chat open AI class takes in a

[07:48] few important parameters which will give us full control over the behavior of this chatbot but before we continue please like this video And subscribe to my channel for more cont content on Lang chain so in this object we can pass in a few important properties like the model name and here we can specify any of the available open AI models like GPT 3.5 turbo or the more advanced GPT 4 but let's stick with GPT 3.5 turbo for now

[08:20] secondly we can also set the temperature and this will control how creative or how factual the model will be one being fully creative and zero being strict and factual let's set the value of 0.7 to make the model slightly creative we can also specify the maximum amount of tokens that we are willing to spend on a generation for instance like a value like 1,000 and another important property that you'll find yourself using

[08:52] all the time is verbose verbose will allow us to debug the model so let's set verbose to true and let's run our script again and this time we can see a lot more information being returned from the chain like the amount of tokens that were used and some additional information about the steps that were executed this will come in handy when we start working on complex chains and agents congratulations you have just created your very first Lang chain application and chatbot next let's have

[09:23] a look at how we can use prompt templates to control the conversation

LangChain Tutorial (JS) #1: Intro & Setup ​

Tutorial Overview ​

What You'll Learn ​

Transcript ​

LangChain Tutorial (JS) #1: Intro & Setup

Tutorial Overview

What You'll Learn

Transcript