Integrating external applications with Sylius

Installation

Clone this repository by running the following command:

$ git clone https://github.com/SyliusLabs/IntegratingExternalApplicationsWorkshop.git

Install the dependencies by running composer install in both RecommendationEngine and Sylius directories.

We recommend to load these directories as different PhpStorm projects in order to receive the best autocompletion support.

Timeline

1. Sylius: dispatching an AMQP message based on orders’ post complete ResourceControllerEvent.

   Post complete event is named sylius.order.post_complete.

   We will use OldSound RabbitMqBundle to handle these messages.

   AMQP message contract:

   {
       "type": "customer_placed_order",
       "payload": {
           "customerEmail": "customer@example.com",
           "orderToken": "aa61ce8bb0efeef6153fccc047df4550",
           "productsCodes": ["tv", "router", "computer"]
       }
   }

   You can test if your implementation is working correctly by running the PHPUnit integration test:

   $ bin/phpunit tests/RecommendationEngineIntegrationTest.php

2. RecommendationEngine: denormalise the AMQP message to an event.

   We will use SyliusLabs/RabbitMqSimpleBusBundle to denormalise the message and publish it on the event bus.

3. RecommendationEngine: handle the event and update Neo4j database (projector) based on provided models.

   In order to use Neo4j you need to change the neo4j user password by logging in at http://localhost:7474/browser/ with neo4j username and neo4j password.

   We will use GraphAware's Neo4j OGM to persist models to Neo4j and SimpleBus' EventBus to handle the event dispatched before.

4. RecommendationEngine: write an API endpoint for getting product recommendations.

   This API endpoint should be bound to /recommented-products/{productCode} and return an array of recommended products' codes in JSON format like the following:

   GET /recommented-products/product_1
   
   ["recommended_product_1", "recommended_product_2", "recommended_product_3"]
   

   If limit GET parameter is passed, it should restrict this endpoint to return maximum limit recommended products:

   GET /recommented-products/product_1?limit=1
   
   ["recommended_product_1"]
   

   Repository method providing recommended products' codes based on product code is already provided.

   You can test if your implementation is working correctly by running the PHPUnit integration test:

   $ bin/phpunit tests/RecommendationEngineIntegrationTest.php

5. Sylius: write a recommendation provider which returns the recommended products.

   The proposed architecture involves ProductRecommendationProvider which returns an array of recommended products entities. It would use ProductRecommendationClient which returns an array of recommended products codes based on the response of recommendation engine API endpoint. This will make caching of these products codes easier and achievable by decorator pattern.

6. Sylius: write a partial rendering the recommended products.