Read Me

Simple app to show case simple recommending "system" for an e-shop. App uses MySql DB. All needed docker commands along with DDL and test data is in folder ./docker-db/

App expects DB with this connection string and user/password:

spring.datasource.url=jdbc:mysql://localhost:6033/buy-me-maybe
spring.datasource.username=root
spring.datasource.password=root
spring.jooq.sql-dialect=MYSQL

There 4 rest enpoints

• GET http://localhost:8080/customer/recommend/{customer-id}[/{limit}]

  Returns list of recommended items for a given customer id. By default limited to 10 items, but optionally path parameter {limit} could by supplied (number of items is also limited by setting of recommanders in DB)

• GET http://localhost:8080/customer/fav/{customer-id}

  Returns list of items that customer has selected as favourites

• POST http://localhost:8080/customer/fav/{customer-id}/{item-id}

  Adds given item to the given customer's list of favourites

• DELETE http://localhost:8080/customer/fav/{customer-id}/{item-id}

  Removes given item from favourites list of given customer

4 weighted recommending systems

• DefaultRecommender
• BestSellingRecommender
• NeighbouringRecommender
• OthersFavsRecommender

Test data (all ids sequential):

• 500 customers
• 3943 customer_fav_item - randomly distributed favourite customer items
• 500 items
• 300 orders - randomly distributed among customers
• 9972 order_items - randomly distributed