Project description:

• This is a Java Spring Boot project taken as a home assignment while being in a Wallapop backender application process. It uses Maven as the dependency Manager. The project is based on moving a Robot around a 2D field. To achieve that, we'll run a server that receives instructions through an API, and stores the result on a file database to be able to persist between runs.

  The project has an API-Service-Repository architecture. As an example of design patterns, Factory and Strategy are used on the robot direction.

---

How to use:

• In order to get the server running, we need to run the file run.sh, this will execute a maven plugin to install the dependencies and run the application. The server will be using port 3000. Once we have the app running, we can communicate with it with this two POST endpoints:

POST /initialize_configuration

This endpoint is required first to initialize the app. It will create and store the field configuration.

Body params:

"fieldXSize": integer
"fieldYSize": integer
"obstacles": List of list of integers
"roverInitialPositionX": integer
"roverInitialPositionY": integer
"roverInitialDirection": string (n,w,e,s)

Example of use and response:

POST /move_rover

With this endpoint we can send movement instructions to the server.

Body params:

"commands": string (sequence of f,b,l,r)

Example of use and response:

In case of finding an obstacle:

---

Problems encountered:

• Indecision to separate the files as package-per-feature or package-per-layer. At the end I decided package-per-layer because is more clear on projects of this small size. In case of adding extra features we could study the possibility of separate the files package-per-feature, because this way we can increase the cohesion and the autonomy of related files on bigger projects.

• First of all thought about implementing the field using a matrix representing field obstacles, but using a HashMap to store obstacles is more memory efficient.

---

Things to improve:

• I´ve created a small custom repository that reads and writes on a file, to have persistance and at the same time don't need to set up Hibernate and MySQL. A good next step in case we want to extend this project could be adding an ORM like Hibernate as the persistance framework.

• For this exercise I´m assuming the initial configuration POST body will be correct, but a next step will be adding validation.

• Adding token authorization provided on the requests headers to protect endpoints from external users.

• Return a DTO instead of the whole entity from the repository. Then we´ll manage to:

  • Just fetch required data from the database, avoiding possible extra queries.
  • Hide unnecessary entity attributes for the repository user, we´ll only send the needed values.