Hiccup application

This is a simple application that try to simulate player for Dragons of mugloar game. The application is named in honor of the main character from How to train dragon movie.

Getting Started

For development I have used Java 17 and with source compatibility only for this version. It means that if you want to build or run release jar in your PC, you need to have Java 17 version. Project use gradle for build and other repetitive tasks. To build project you need to execute:

./gradlew build

If you want to run application you can either:

./gradlew bootRun --args="-maxTurn 100"

or:

java -jar ./build/libs/hiccup-0.0.2-SNAPSHOT.jar -maxTurn 100

If you want to know what other options exist for this application:

java -jar ./build/libs/hiccup-0.0.2-SNAPSHOT.jar -help

Result of some random game in console:

Configuration

It is possible to change game service url or logging level by using application.yml file. Just put near jar a file with content:

hiccup:
  gameService:
    baseUrlForAPIv2: https://dragonsofmugloar.com

logging:
  level:
    root: INFO
    com:
      github:
        viktornar:
          hiccup: INFO

Depending on log level you can see more or less information in console. So if you specify DEBUG instead of INFO you will be able to see more details about what happening inside application. If you do not wan't to see any infromation related to log, but only want to see end result, you can specify NONE

What was used?

Under the hood I decided to use spring boot since it supports DI, web client, and it is even possible to do native build (experimental feature of spring).

I have added only basic dependencies from spring that needed to work with IoC and web client.

I have decided to use state machine. Most of the existing solution are quite big and would be too much for such kind of task. I decided to implement some kind of state machine in RxJava (wanted to try this awesome library). At the end I have quite flexible state machine with possibility to redefine default flow on transition stage. At the end it seems that solution is also too complicated, but at least it is not difficult to test all flow.

For data mapping from one data (reward, basket, game) class into another (trainer context) I'm using map struct. This library let to simplify such data flow between data classes.

At the end I was able to reach level 15 with 4001 score in best case. On average, I'm always able to reach level 15 with around 3000 score, but I believe that it can be improved, but I need to know what kind of logic is under the hood in game engine.

What was wrong?

State machine? It seems that mys solution is not so good as I thought at start. Api of state machine that was created is complicated and not clear.

It would be interesting to implement decision tree for solving this game, but I have a limited amount of time and will leave this maybe (repeat, maybe :D) in the future.

Development

I'm using unit tests, check style and sonar to ensure code quality and gradle for build and dependency management.

To run tests:

./gradlew test # you can run ./gradlew check as well

To run checkstyle:

./gradlew checkstykeMain

To run sonar:

docker-compose up -d -f ./src/main/docker/sonar.yml
./gradlew sonarqube

To run build:

./gradlew build

Of course, you can just import project in IntelliJ IDEA and use UI to start mentioned gradle task earlier.

Reference Documentation

For further reference, please consider the following sections:

• Official Gradle documentation
• Spring Boot Gradle Plugin Reference Guide
• Create an OCI image
• Spring Configuration Processor

Additional Links

These additional references should also help you:

• Gradle Build Scans – insights for your project's build