Laravel calling model from controller? Is this the right approach? - laravel

Question

I am trying to split some code and let model handle all database stuff, such as create, retrieve etc.
At the moment all of that code was in the controller and I quickly realized that code is very messy so instead I wanted to do something like this:
public function incompletedEntity(EntityRequestPartial $request)
{
$partial_entity = EntityTrash::saveEntity();
}
And my model:
public static function saveEntity(Request $request)
{
$entity = new EntityTrash();
$entity->lat = $request->input('lat');
$entity->lng = $request->input('lng');
$entity->slug = $request->input('name');
$user_id = Auth::id();
$entity->name = $request->input('name');
$entity->type = $request->input('type');
$entity->email = $request->input('email');
$entity->tags = $request->input('tags');
$entity->slug = $user_id;
$entity->building_name = $request->input('building_name');
$entity->address = $request->input('address');
$entity->town = $request->input('town');
$entity->postcode = $request->input('postcode');
$entity->telephone = $request->input('telephone');
$entity->save();
}
However, I cannot call that function because I am not passing an argument from the controller, what is the right approach for this? Should I do $request->input in controller and assign it to an array maybe? and deal with it in the controller?

Answer 1

If you wish to split out the code so that controllers don't touch models, I would recommend that you look into the repository pattern.
The idea behind this would be that in your controller you have
$this->entityRepository->create($data);
You could either pass in an array, or each individual property
$this->entityRepository->create($lat, $lng, $name, $type..etc);
This way, your controller can retrieve all of the data and validate it.
$data = $request->only(['lat', 'lng', 'name', 'type', 'email', 'tags', ..etc]);
$validator = Validator::make($data, ['name' => ['required']]);
Obviously you can use request validation or whatever you're comfortable with.
This way, your controller is responsible for receiving data and validating it. It can then pass it on blindly to the repository knowing that it is valid, and trusting that the repository will do its job.
As a side note, I highly recommend that you do not have your models interact with the Request object directly. The Illuminate\Http\Request class belongs to the HTTP layer, and the model belongs to the persistence layer. These two layers should never know of the existence of the other.

Answer 2

Generally speaking, your model should never depend on the Request object. Models can be used for any access of a database, or other storage area. Many of those access points may not be HTTP at all.
Your HTTP controllers should always be used to pass information from an HTTP request to a deeper layer of your application (e.g. a Model). So if you are using the Request object anywhere, it should be in your HTTP controllers.
I think it's important for you to review why the controller layer is there and why we don't just have Models and Views.

Answer 3

I would recommend you to take a look at the official Laravel Documentation first: https://laravel.com/docs and the Controller / Eloquent Model chapters. Moreover, Adam Wathans talk Github from the current Laracon is a good read for further refactoring.
First, make use of the CRUD pattern from the Laravel Controller in addition with the Eloquent Model. Then pass the validated data to your model via the create/save method.
<?php
namespace App\Http\Controllers;
use App\EntityTrash;
use App\Http\Controllers\Controller;
class EntityTrashController extends Controller
{
public function create(EntityRequestPartial $request)
{
return EntityTrash::create($request->only(['lat', 'lng', 'name', 'type', 'email', ...]));
}
}
Finally, maybe the repository pattern can improve your code as well.