Skip to content

Latest commit

 

History

History
98 lines (67 loc) · 3.9 KB

0-prerequisites.md

File metadata and controls

98 lines (67 loc) · 3.9 KB

Lab 0: Setup

So you want to build a RAG app with React? Brilliant, as this is exactly what this workshop is for!

There are a couple of things you'll need before we start:

0: Assumed installations

Please ensure you have the following tools installed:

  1. Node.js
  2. npm

To check you have Node.js and npm installed, run the following commands:

node -v
npm -v
  1. tsx

If you don't have tsx installed please make sure you have a global install configured by running the below command:

npm install -g tsx

If you receive an error, download and install Node.js and npm using these instructions.

2: Ollama installation

There are many open source and proprietary machine learning models out there that can be used when building RAG applications. To make local development easy, we shall be using Ollama to run our models locally.

brew install ollama

We'll cover how to pull and run models locally in the workshop.

3. Initialization of the starting point application:

To help you get started, a very simple web application is included in this repository under the movie-rag folder. To initialize the application, please follow the below commands in a terminal to start the application.

cd movie-rag
npm install
npm run start

Please ensure that you are present in the top-level folder for this project when you start. These steps should be the same for Windows and Mac.

4: direnv (or alternative) [OPTIONAL]

This lab makes use of environment variables for some attributes that are loaded using process.env within our application. For ease in the workshop I recommend using a shell environment loading tool such as direnv configured to support env files.

If using direnv also make sure that you have configured your profile to accept .env files:

cat $HOME/.config/direnv/direnv.toml

Example config:

[global]
load_dotenv=true

5. An Elastic cluster to store your documents and perform searches against [OPTIONAL]

If you're attending the workshop at React Day Berlin, check out the cluster credentials .env file for your assigned group ont he day. You will see configuration that looks a bit like this:

ELASTIC_DEPLOYMENT=https://my-random-elastic-deployment:123
ELASTIC_API_KEY=ARandomKey!
INDEX_NAME="movies"

For those following along outside this time, or who want their own cluster, please follow either the start-local steps or register for an Elastic Cloud free trial via the below steps:

  1. Create a trial account at https://cloud.elastic.co/ using the Start free trial button.
  2. Add the basic settings for your new cluster:
    • Name your deployment something interesting.
    • Choose your preferred cloud provider (any is fine).
    • Choose your region, ideally one close to your physical location.
    • Keeping the Hardware profile as the default of Storage optimized is fine.
    • Keep Version as the latest.
  3. Configure the advanced cluster settings:
    • Set zone availability on the Hot tier of Elasticsearch to 1 (you don't need to worry about data loss for our toy project).
    • Make sure you have autoscaling enabled on the Machine Learning instances.
    • Keep the default settings for Kibana (the UI and data visualization layer).
    • Remove the Integrations Server instances. These are used for application monitoring which is out of scope of today's workshop.
  4. Hit the Create deployment button.
  5. Take a note of your deployment credentials somewhere safe. Especially the password!
  6. Navigate to your deployment once ready with the Continue button.

Happy workshopping!