I am getting back to Laravel after several years and trying to understand how an already existing REST API coded in Laravel works. I can't understand how a particular controller method with multiple Form Request parameters works (or if it actually does).
The REST API was coded in Laravel 5.1. I've looked at the official documentation (both 5.1 and the latest) and tried to search the web and SO for related topics (e.g., "laravel controller multiple form requests", "laravel controller multiple type-hint requests", etc.), but I can't seem to find a clear explanation. Maybe I'm looking at it from a wrong angle.
public function store(ProductRequest $productRequest, PromoRequest $promoRequest)
{
// Validate product
$product = new Product($productRequest->all());
// Validate promo
if ($promoRequest->get('promo')) {
$promo = new Promo($promoRequest->get('promo'));
}
...
}
In most documentation, the controller would accept only one Request object. I did actually see some examples that have multiple Form Request parameters, but often they were recommended to use only one Form Request. But best practice aside, how does this code work? When this method is called, how does Laravel know how to split the Request into two separate Form Request classes?
Please feel free to let me know if and how I can explain my question more clearly.
you can write like this:
function example(Request $request) {
$productRequest = new ProductRequest($request->all());
$promoRequest = new PromoRequest($request->all());
}
but on validate you should change the way to this:
$data = $productRequest->validate($productRequest->rules());
another way that i test in laravel 8 is to add both form request object in methd parameters like :
function example(ProductRequest $productRequest,PromoRequest $promoRequest) {
$productRequest->validated();
$promoRequest->validated();
}
The Laravel service container is a powerful tool for managing class
dependencies and performing dependency injection. Dependency injection
is a fancy phrase that essentially means this: class dependencies are
"injected" into the class via the constructor or, in some cases,
"setter" methods.
You can read more about it here: https://laravel.com/docs/5.8/container
Edit: Additional question: Still, I can't understand how the HTTP request (which is only one) can be split into two different Request classes?
The HTTP request isn't split; it is merely sent through both classes.
Ex:
function example(Request $request) {
$productRequest = new ProductRequest($request);
$promoRequest = new PromoRequest($request);
}
Would be the same.
Related
First important information: I’m new to laravel, so your patience is appreciated.
I’m currently migrating a framework of mine to laravel, so I’m in the early stages.
Currently, I’m trying to set up an API endpoint to make small changes on some records. I’ve already managed to set up a API for inserting records and works perfectly. However, for setting up an API for small changes (patch), I’m having difficulties, probably because I’m not fully familiar with laravel’s Request class.
My successful insert set up looks like this:
\routes\api.php
Route::post('/categories/',[ApiCategoriesInsertController::class, 'insertCategories'], function($insertCategoriesResults) {
return response()->json($insertCategoriesResults);
})->name('api.categories.insert');
\app\Http\Controllers\ApiCategoriesInsertController.php
// some code
public function insertCategories(Request $req): array
{
$this->arrCategoriesInsertParameters['_tblCategoriesIdParent'] = $req->post('id_parent');
// some code
}
With this set up, I’m able to retrieve “id_parent” data set through POST.
So, I tried to do exactly the same architecture for patch, but doesn’t seem to work:
\routes\api.php
Route::patch('/records/',[ApiRecordsPatchController::class, 'patchRecords'], function($patchRecordsResults) {
return response()->json($patchRecordsResults);
})->name('api.records.patch');
\app\Http\Controllers\ApiRecordsPatchController.php
// some code
public function patchRecords(Request $req): array
{
$this->arrRecordsPatchParameters['_strTable'] = $req->post('strTable');
// some code
}
In this case, I´m using postman (PATCH request), testing the data in the "Body tab" with key "strTable" and value "123xxx" and I´m receiving “strTable” as null.
Any idea of why this is happening or if I should use another method in the Request class?
Thanks!
You can access parameters on the Request object using one of the following methods:
$req->strTable;
// or
$req->input('strTable');
The input method also accepts a second parameter which will be used as the default return value if the key is not present in the Request.
If you want to check whether or not the Request contains a value before you attempt to access it, you can use filled:
if ($req->filled('strTable')) {
// The request contains a value
}
Turns out that the way I had set up was in fact working and retrieving data:
$req->post('strTable');
The problem was in how I was testing it. In postman, there are several options to configure:
form-data
x-www-form-urlencoded
raw
binary
I had already switched to x-www-form-urlencoded to test it, but I forgot to fill the “key” and “value” information again. I didn’t realize that the fields blank as we switch between them.
Summing it up: It works when x-www-form-urlencoded selected but doesn’t work with form-data selected. Don’t know what the difference between them yet, but I’ll research it further.
By the way, it worked also with the suggestion from Rube Hart:
I have a Laravel Application and another APP making calls via API to Laravel. These 2 projects are separated.
Laravel and App have their own multilanguage system. They work independently but uses the same key translations.
So my idea was that all Laravel responses must be translations key, like: 'messages.success'.
With this response, the App can translate it.
All of these are working fine.
The problem appeared when I started working with Laravel Requests for validating forms.
In this case, the validation errors are automatically translated so the App receives the response translated with the default language of the Laravel application.
So what can I do?
I thought with 2 ideas but I don't know if they can work.
1: Passing the language into params. Don't know if it can work, how can I set the language before Laravel validates the Request?
2: Override the functionality of Requests to return messages without translate, so instead of "Felicidades" return "messages.success". I really like this approach. But how can I do it for all the rules? Overriding the messages function like this:
public function messages()
{
return [
'unique' => 'validation.unique'
];
}
For every rule works... but I feel bad.
Another approaches?
What is the best way to fix this problem?
I would suggest that you use this hacky solution in 2 lines of code. Go to /resources/lang/{code}/validation.php. You can see that it returns an array of messages by default. Modify it like so:
// Replace return in the first line
$ret = [
/* all the translations go here as normal */
];
// Add this as the last line. This will replace all values with their keys.
return array_combine(array_keys($ret), array_keys($ret));
After that you can use validation as per usual and you'll get validation message keys instead of messages. Cheers and hope this helps.
I am new to Laravel coming from CakePHP where the form and save method for a form is one and the same function name. I saw in many Laravel tutorials that the from method (that displays the form) is different than the method to save form (that actually saves data). Why using 2 different method names?
For example what's wrong with:
pub function xyz(Request $request)
{
if($results->isMethod('post')){
... then save and return redirect
}
... the code for showing the form in case there is no POST.
then having 2 routes one for GET and one for POST on the same url?
It is because people like to filter out things at route level not in controller, Also it helps developer to apply middleware grouping for each route separately. so that they can apply roles and permission etc. easily at route level.
It will looks horrible if mix all things in controller.
Think about middleware and groups in your code.
It is because you don't wanna mix a lot of logic in the same method . The case you have simple is the simple scenario . But there will be case where you wanna pass initial data in the create form . You have to write logic for that also in the same method and while you store the data you need to do the validation and calculate other business logic . If you combine all those things in one method it will mix all the things in one method and code difficult to read
I am trying to understand how POST routing will work. I have a method defined, signup(), and I want to use the same method to detect if the user wants to sign up (so load the signup view) or if the user already in the signup view (form) and posting his details to register.
Can this be done in one function in laravel? if yes, then how? Is this controlled by Routes and if yes, can someone please clarify this with an example?
Laravel documentation is really confusing for beginners.
Thanks in advance,
While this is possible but it's not recommended way to do that, you should keep your routes separated from each other (using GET and POST) and should use different methods as well. Basically any form submission should use POST request (using POST HTTP method) and to show the form just use a GET method but anyways, you can do it (what you have asked for) like this way:
// Declare the route
Route::post('signup', 'UserController#signup');
Now in your signup check for the submit button to make sure that, the form is submitted, so if the input submit is available in the $_POST array then the form is submitted otherwise, it's not submitted but an empty form was presented to the user or a failed validation redirect happened. Maybe something like this:
public function signup()
{
if(Input::has('submit')) {
// It's a submission, so Validate submitted Form data
// if invalid then redirect back with inputs and errors
// Otherwise save it
}
else {
// show the form
return View::make('user.signup');
}
}
Don't do it
This is just an idea but, it's a bad idea, just think about what happens if you have errors on your form and you want to redirect back then the whole thing would become messy, the controller method will become totally unmanageable after a while because it does many things while it should have only one specific responsibility.
I have this practical experience, because, I used to think that, if I can use one function for loading and saving and even also updating then it would be smart but to be honest it was stupid and obviously it's an anti-pattern, not the best practice, against KISS (Keep It Simple Stupid) principle. This kind of coding is a bad idea and you'll suffer for it in future and you would not dare to touch the code thinking that if you brake anything because you'll be confused by your own code.
Just use separate methods to show a form and save submitted data, Also check this on slideshare.
Yes, you can use one route to do it:
Route::any('signup', 'SignupController#signup');
Or two routes pointing to the same url:
Route::get('signup', 'SignupController#getSignup');
Route::post('signup', 'SignupController#postSignup');
In both cases you'll need a controller:
Here it is with all related methods:
class SignupController extends Controller {
// This one is for Route::any()
public function signup()
{
if (Input::has('email'))
{
// create your user
}
return View::make('signup');
}
// those two are for the second option
public function getSignup()
{
return View::make('signup');
}
public function postSignup()
{
// create your user
}
}
Here is what I'm tyrnig to do:
Make ajax request to retrieve JSON data from a PHP script
Insert that information into DataStore Models
Store those models within a controller
Display the information using {{#each}} with a handlebars template
Does ember-data have a built in way of retrieving data? If not, where
should the AJAX request be implemented?
What is the best way to insert the JSON data into the DS model?
What is the best way to then sync the models up with a Controller?
Any examples that implement all of the 4 steps would also be very helpful, since I can't seem to find any.
<edit>
Like I said in the comments, this questions asks a lot at once, so to follow up, here's a work in progress fiddle: http://jsfiddle.net/schawaska/dWcUp/
This is not 100%, but covers some of your questions. It uses the FixtureAdapter.
I'll be updating it as I find time.
</edit>
1 Make ajax request to retrieve JSON data from a PHP script
Ember-Data will take care of that for you. Consider the following:
window.App = Em.Application.create();
App.Store = DS.Store.extend({
revision: 12,
adapter: 'DS.RESTAdapter'
});
App.Product = DS.Model.extend({
name: DS.attr('string'),
imageUrl: DS.attr('string')
})
The code above defines a data store (almost like an ORM) within your client app. Ember uses convention over configuration (heavily), so as per configuration this code expects your backend to have a resource in the same domain as /products which talks to GET, POST, PUT and DELETE.
2 Insert that information into DataStore Models
Considering what I said above, by calling one of the following:
App.store.find(App.Product) or App.Product.find()
EmberData will fire a GET request through AJAX targeting your resource /products, and if you say App.Product.find(1), it will target /products/1.
Your app store will use its adapter and the adapter's serializer to translate the JSON result and materialize its data into your App.Product model.
3 Store those models within a controller
This is done when defining your application router. Regardless of what you do, Ember will run its own workflow, but it provides you several hooks, giving you the control of that specific action. Consider the following:
App.Router.map(function() {
this.resource('products');
});
App.ProductsRoute = Ember.Route.extend({
model: function() {
return App.Product.find();
}
});
The code above populates the model used in the products route (which you would access at http://yourdomain.com/#/products). Internally it will generate the code for your ProductsController, or you can define your own, which should extend ArrayController.
Controllers will have a content property which is an alias to the model or model collection. Again, convention over configuration.
4 Display the information using {{#each}} with a handlebars template
Providing you're following the conventions, in your handlebars template, you should iterate through your collection like this:
{{#each product in controller}}
{{#linkTo 'product' product}}
{{product.name}}
{{/linkTo}}
{{/each}}
Does ember-data have a built in way of retrieving data? If not, where
should the AJAX request be implemented?
Yes, simply call App.Product.find() for a product model and it return you a blank ModelArray while firing the AJAX request to the products resource in your backend, then materialize/populate your data into each model once it receives the data back from the server.
What is the best way to insert the JSON data into the DS model?
You shouldn't be concerned about this if you're using ember-data. The framework does that for you in most cases. That's why we love it. You might, however, have to configure mapping, namespace and plurals depending on your backend.
What is the best way to then sync the models up with a Controller?
Something similar to this:
var product = App.Product.createRecord({
name: 'laptop',
imageUrl: 'path/to/image.png'
});
product.save();
This should fire a POST or PUT request to your backend API.
You should definitely check:
http://emberjs.com/guides/
https://peepcode.com/products/emberjs
http://toranbillups.com/blog/archive/2013/01/03/Intro-to-ember-js-and-the-new-router-api/
Making the AJAX request
// Find all pets.
var pets = App.Pet.find();
// Find pet with ID of 1.
var pet = App.Pet.find(1);
Into DataStore Models
pets above will be a DS.RecordArray containing many App.Pet models, whereas pet is just one App.Pet.
Store in Controller
App.IndexRoute = Ember.Route.extend({
model: function() {
return App.Pet.find(4);
}
});
The router is used to setup the controller, and so we specify here that the IndexController should hold one App.Pet with the ID of 4. This can of course be dynamic. Since your controller represents only one model, it should be of the type ObjectController, but if it was used to store many pets, then it should be of the type ArrayController.
By specify the model, you will have access to it in your IndexController and index view (data-template-name="index"). This is because when you move into your index route, the IndexController is located/instantiated, IndexView is instantiated and placed into the DOM, all after consulting the IndexRoute for setting up the controller.
You can now do something like this in your template (although model. is not necessary): {{model.name}}, which will get you your pet's name.
Display using #each
Find all your pets using a modified version of the above code, but returning all of the pets. Remember, this is done by specifying no arguments to your find method:
App.IndexRoute = Ember.Route.extend({
model: function() {
return App.Pet.find();
}
});
Now we can do loop through all of the pets in the index template. Whilst there are many ways to loop, such as including content./model., excluding .content/model, using this, controller, et cetera..., it's not necessary, but that's for another day. What matters at the moment is that this will work for you, and will be the most self-intuitive:
{{#each pet in controller}}
{{pet.name}}
{{/each}}
I'll put together a jsFiddle for this if you would like. Please let me know.
Questions
Does ember-data have a built in way of retrieving data? If not, where
should the AJAX request be implemented?
Yes, that's Ember Data module which has some good guides on EmberJS.com.
What is the best way to insert the JSON data into the DS model?
Using Ember Data as per the examples up above.
What is the best way to then sync the models up with a Controller?
Using the model hook in the appropriate route to specify which model(s) your controller represents.