Vectors Speecoding

Scaffold for the vectors speecoding project built with Laravel Frawework and OpenAI.

Goal

Build a simple application to find livecoders based on a text search using Vector Embeddings.

Requirements

List of technologies used in this project:

• PHP 8.2^
• Laravel 11.x

Also, you will need an account at Qdrant Cloud and a OpenAI API Key since we'll use the text-embedding-ada-002 for this example.

[TIP] You can replace OpenAI for any other Vector Embedding of your choice. Feel free to make it useful for you as well!

Packages Used

• Qdrant-php - PHP Client for Qdrant Vector Database
• OpenAI Laravel - Laravel Integration for OpenAI Client

Build by yourself

Here's a quick to-do list on how you can build the whole project

• Create a Laravel Command php artisan make:command SetupVectorDatabaseCommand
  • Inside this command you should use the Qdrant Client to :
    • Create a Collection called twitch-streamers;
    • Create a Vector called streamers with a size 1536 (ada-002 embedding size)
• Create a Laravel Command php artisan make:command StreamersImportCommand
  • Inside this command, retrieve the data from streamers.json and load it to your Qdrant Collection
• Using a Laravel Livewire Component which is already provided on the app, you should:
  • Embed your text input value using OpenAI text-embedding-ada-002
  • Search in Qdrant the closest results and display it into the view.