Implementation of a video rental store using DDD inspired by the book Implementing domain-driven design by Vaughn Vernon
Application description:
We wamt to build an online video rental store with three main features:
1.- Have an inventory of all the films
List all films available and existing in the online store
2.- Calculate the price for rentals
The price of rentals is based on the type of film and how many days the film is rented. The customers say when renting for how many days they want to rent and pay upfront. If the film is returned late, then money for the extra days is charged when returning.
The online store has 3 types of films:
1.- new film: price is times number of days rented
2.- regular film: price is for the first 3 days and then times the number of days over 3
3.- old film: price is for the first 5 days and then times the number of days over 5
3.- Keep track of the costumers bonus points Customers get bonus points when renting films. A new release gives 2 points and other films give one point per rental
Let's start with a diagram out of a event storming session:
There has been identified 4 different bounded contexts i.e. customer lifecycle (customer), film lifecycle (purchases), filmtransactions (rentals) and filmpayments (payments). There could be a fifth, gamification, but for the time being bonuscard has been modelled as a value object for customer.
We used package by feature, i.e. screaming architecture, giving us the following benefits:
- We can see immediately, just by opening the project, what the application does
- Allows promoting bounded contexts to microservices/different projects easily.
Film transactions have an event store associated and use the in memory read model for checking the state of the movie. The rental itself is the core of the domain and this allows for more exploitations of the data should it be needed.
A return is considered late for one day rental if one day (3600 seconds * 24 hours) or more has passed.
-
The different 'bounded contexts' are all glued up in the use case layer and share a common database. Future improvements could be to separate the data stores and use the events identified in the event storming phase to decouple the modules if complexity starts growing.
-
The filmtransaction repository works with the assumption that the events are in order. The column createdOn is used also as rentedOn. A different field rentedOn could have been added to allow this separation of concern.
This application uses maven as build tool. To run the tests execute:
mvnw test
To build the application type the following command:
mvnw spring-boot:run
This application includes swagger to facilitate using the exposed API. Go to the following URL to see the available operations:
http://localhost:8080/swagger-ui.html