https://github.com/redhat-middleware-workshops/devsandbox-category-llm-basics
Get started setting up LLM essentials for your applications
https://github.com/redhat-middleware-workshops/devsandbox-category-llm-basics
Last synced: about 1 year ago
JSON representation
Get started setting up LLM essentials for your applications
- Host: GitHub
- URL: https://github.com/redhat-middleware-workshops/devsandbox-category-llm-basics
- Owner: RedHat-Middleware-Workshops
- Created: 2024-08-28T16:37:52.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2024-11-17T00:09:33.000Z (over 1 year ago)
- Last Synced: 2025-04-12T23:14:22.178Z (about 1 year ago)
- Language: Java
- Size: 6.77 MB
- Stars: 2
- Watchers: 10
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: Readme.md
Awesome Lists containing this project
README
> [!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](https://workspaces.openshift.com/#https://github.com/RedHat-Middleware-Workshops/devsandbox-catalog-demo.git)
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](#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](https://developers.redhat.com/articles/2024/09/26/tutorial-tool-your-llm-apache-camel-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:
```bash
cd camel/tools/solution/local
```
1. Edit the `application.properties` file and configure the following setting:
- `llm.url`
1. Run the application using Camel JBang with:
```bash
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:
```bash
../../scripts/xx
```
When the export complets, you can run locally your Camel Quarkus project with:
```bash
./mvnw quarkus:dev
```
If you'd like to deploy on OpenShift
1. Configure the `llm.url` setting as necessary.
1. Make sure your `oc` client points to your target namespace.
1. Run the command:
```bash
./mvnw package -Dquarkus.kubernetes.deploy=true
```