ogrre/laravel-chatgpt

Library to add chat management for openai client

1.0.6 2024-01-04 10:58 UTC

Requires

• php: ^8.1
• illuminate/database: ^7.0|^8.0|^9.0|^10.0
• illuminate/support: ^7.0|^8.0|^9.0|^10.0
• openai-php/client: ^0.6.2

MIT

• Ogrre

README

Documentation

This libray need openai php client, so don't forget to add in your .env this variables:

# .env

OPENAI_API_KEY=
OPENAI_API_ORGANIZATION=
OPENAI_MODEL=gpt-4

Installation:

To install the Laravel Chat for ChatGPT library, run the following command:

composer require ogrre/laravel-chatgpt

After the installation, publish the vendor files by executing the command:

php artisan vendor:publish --provider="Ogrre\ChatGPT\ChatServiceProvider"

By default, the service provider will be automatically registered in the app.php file. However, if needed, you can manually add the service provider in the config/app.php file:

# config/app.php

'providers' => [
    // ...
    Ogrre\ChatGPT\ChatServiceProvider,
];

Finally, run the migration command to create the necessary database tables:

php artisan migrate

Usage instructions

To associate a chat with a model, the model must use the hasChat trait. For example, in the User model:

class User extends Authenticatable
{
    use HasFactory, HasChat;
    // ...
}

Once the model is set up with the hasChat trait, you can create a new chat using the following command:

$optional_title = "Translate this word";
$optional_role = "You're a helpful assistant";

$user_chat = $user->newChat($optional_title, $optional_role);

After creating a new chat or retrieving an existing one, you have two options to interact with it:

Using the chat object directly:

$chat->gpt("say hello in French");

Using the model directly (in this example, the User model):

$user->chatgpt("say hello in French", $chat);

The Chat model is a standard Laravel Model, so you can use all the usual functions, including delete and others.

Additionally, the library provides a ChatResource for displaying chats in an API, allowing for easy integration with API endpoints.

Contributing

Contributions are welcome and will be fully credited. I accept contributions via Pull Requests on Github.

Support me