Squelette de base contenant uniquement un point de terminaison /_healthcheck
et comprenant :
/docker
et le README.md
qu'il contient).bin/phpunit
pour réussir les tests unitaires)./vendor/bin/phpcs
, exécutez bin/phpcs
pour exécuter un renifleur de code basé sur la norme PSR2). Si vous souhaitez inclure Redis dans votre projet, exécutez docker-compose -f docker-compose-with-redis.yml up -d
Docker
docker-compose up -d
et attendez que tous les packages requis soient installés.docker exec -ti mylocalproject-php-fpm bash
) et exécutez une composer install
depuis l'intérieur{
"status" : " i am ok "
}
Maintenant, c'est à vous de décider comment vous voulez que votre projet se développe :)
Si vous considérez l’environnement dockerisé comme suffisant pour la production, n’oubliez pas ces deux étapes :
composer install --no-dev --optimize-autoloader --apcu-autoloader
APP_ENV
en pro
dans le fichier .env
. Vous pouvez constater des performances raisonnablement bonnes en exécutant l'outil Apache Bench, avec environ 100 s de requêtes simultanées en 10 s : ab -n 100 -c 10 http://localhost:8000/_healthcheck
Ce projet est architecturé selon un pattern DDD + CQRS, qui contiendra trois parties principales :
Domain
.Application
et Domain
. Extension de l'architecture précédente, il s'agit de la structure de dossiers principale du projet.
Domain
└─── Exception: Contains `AppException`, responsible for any exception about the application which
| will contain an internal error code, plus a collection of `ErrorCodes` along the app
| (important, taking into account that e.g. HTTP error codes sometimes are not informative enough)
|
└─── Model: Core business logic within the aplication, which will also contain validations for correctness.
| In case that we use events through the system (e.g. creation, updates...) they would be placed here as well.
| It will contain behavior and data, following[DomainModel](https://martinfowler.com/eaaCatalog/domainModel.html) Martin Fowler's pattern
|
└─── Repository: Interfaces to access the data model that shall be viewed as a Collection, with a composition as follows:
└─── `findOfId`: Find a specific Model from a given unique ID. Ideally returns a Domain exception when not found
└─── `findOfXxx`: Find the Model through an unique ID
└─── `save`: responsible of create/update the model
└─── `delete`: responsible of delete the model from the collection
Application
└─── Command: Executes an use case by orchestrating domain objects, and ideally produces an output in the shape of an event
| └─── Request: Value objects representing a request per command
|
└─── Query: Implementation of specialized queries (anything different of a `findOfId`) to the model.
| Note: those are part of the Application because the Domain should not be aware at all about the expected Responses
└─── Request: Value objects representing a request per query
└─── Response: Set of Value Objects which we expect to obtain as a result of the query, ideally implementing a `Serializable`
Infrastructure
└─── EventListener: Even though those could be part of the Application layer (as actors regarding different events),
| right now they are an infrastructure concern as the infra too is implemented as an Event schema.
| This could be different depending on your framework of choice (e.g. in Zend they would be middlewares)
|
└─── Repository: Contains folders with the different sources of implementation per repository (e.g, `MySqlUsersRepository`)
|
└─── Service: Third-party specific implementations of services and connectors (REST, MongoDB, Stripe...)
|
└─── Ui: Contains the user interface communication, mainly Console commands (cron/daemons) and Web commands (controllers)