Note

If this file is rendered in TEXT mode in VS Code, click the preview button.
(upper-right corner, has a 📖+🔍 symbol)

LLM Basics in the Developer Sandbox

Gain familiarity with LLM tools by learning how to easily create them with Apache Camel. In this lab, you'll create an agent that uses live feeds to assist clients.

Get started with the tutorial

Note

If using the Developer Sandbox for the first time, you'll be required to register for a Red Hat account.

To launch this tutorial, you only need 2 steps:

1. If you haven't already done so, click the link below to provision the tutorial:

   _{(right-click & open in new tab)}
   Provision Tutorial in the Developer Sandbox

2. _{(In VS Code, open Readme.md in preview mode, if not showing automatically)}

   Once provisioned, from VS Code in DevSpaces, unfold the ENDPOINTS group in the left panel, open the Tutorial documentation and make sure you choose the LLM tools tile to get started, as per the actions in the picture below:

   Mouse actions in VS Code to open the tutorial.

How to recover from failed workspace

Sometimes the provisioning process goes wrong.
Try following the steps below to delete the failed workspace and restart.

1. Click Workspaces at the top of the screen as illustrated below.
   Dev Spaces's dashboard will open.

2. Find the ⋮ button (failed workspace).

3. Select Delete Workspace

4. In the confirmation panel, tick the box, and click Delete

Mouse actions in VS Code to delete failed workspace.

After the actions above, retry provisioning your lab.
Jump back to:

• Get started with the tutorial

Introduction

This repository contains a collection of labs runnable in the Developer Sandbox, a free to use OpenShift environment, where users can follow guided instructions to fully cover the material in a fully remote development platform.

The following article in Red Hat Developers guides readers on how to get started to enjoy the learning material in the Developer Sandbox.

• Tutorial: Tool Up your LLM with Apache Camel on OpenShift

Running locally

This tutorial is intended to run in the Developer Sandbox.

However, if you're familiar with the content and you just want to run the solution locally, you can.
You'll need to make sure you meet the pre-requisites and follow the steps below.

Pre-requisites

Your environment must have:

• Java 17 installed
• Camel JBang installed (tested with 4.8.1)
• Connectivity to an LLM with Tools support

Run with Camel JBang

1. Change to the directory:

   cd camel/tools/solution/local

   
   

2. Edit the application.properties file and configure the following setting:

   • llm.url
   
   

3. Run the application using Camel JBang with:

   camel run *

Export and Deploy

If you'd like to deploy on OpenShift, export the solution to a Camel Quarkus application using the following script:

../../scripts/xx

When the export complets, you can run locally your Camel Quarkus project with:

./mvnw quarkus:dev

If you'd like to deploy on OpenShift

1. Configure the llm.url setting as necessary.
2. Make sure your oc client points to your target namespace.
3. Run the command:

   ./mvnw package -Dquarkus.kubernetes.deploy=true