/be-riesenie

Primary LanguagePHP

BE riesenie

ER diagram

Endpointy

  • [GET] /api/products liast of products
    • pagination with page query parameter
  • [POST] /api/products create new product
    • Request body: { "name": "AMD 1235", "price": 450000, "category": { "name": "New GPU" },"images": [ {"name": "image1","path": "/files/image1.png"}, {"name": "image2","path": "/files/image2.png"}, {"name": "image3","path": "/files/image3.png"} ] }
  • [GET] /api/products/{id} product detail
  • [PATCH] /api/products/{id} product update
    • Request body contains data you want to edit
  • [DELETE] /api/products/{id} delete product
    • partially implemented

CACHING

1.

  • During GET /api/products/{id} request
    • When someone accesses product detail data is cached, and retrieved if it's not older than 1800 seconds
  • Making sure we don't have incorrect data, cache for product should be invalidated during **PATCH ** /api/products/{id} request
  • After data is deleted from database with DELETE request, remove data from redis cache and Elasticsearch
  • Product data can also be cached in POST request when new product is created, Also add to Elasticsearch

2.

  • Another option to cache products is to create console command and then run it in cron job
  • This way we can prepare products and don't have to wait for user to access it, and having worse experience
  • If we have too many products caching all of them might be inefficient and costly.
    • But caching products that are accessed the most might be good idea. (Of we will need to track how many times are product accessed)

SEARCHING

  • Query parameter name will look up products based on name
    • GET /api/products?page=1&name=nvidia
  • Query parameter category will return products based on their category
    • GET /api/products?category=gpu
  • Query parameter price with gt, lt, lte, gte
    • GET /api/products?price[gt]=100&price[lt]=200

Elasticsearch

  • For using ElasticsSearch in our api, we need to install Elasticsearch instance on our machine, and it's best to use already existing package like friendsofsymfony/elastica-bundle making implementation simple

  • After that configure bundle and port (default 9200)

  • Define indices on entity product

  • After proper setup every new or modified entity will be updated automatically in Elasticsearch