Storing Robot Configuration

Robot code must be highly configurable. Components change, values need to be tweaked, and sensors get relocated. A common approach to robot configuration is to define a few static final classes throughout the codebase, and fill them with settings variables. Lib5K provides a much cleaner way to handle robot configuration. This is via the use of a JSON configuration file stored on the robot.

How it works

First, you define a Java class to store your data. This is basically just a class filled with empty variables. You can then pass this class, and a filename into SingleInstanceJSONConfig<T>. SingleInstanceJSONConfig<T> will parse the provided JSON file, and return an object of your class, but with all the variables filled out. This process uses Google's GSON library in the backend, so anything that works in GSON will work here too.

By default, SingleInstanceJSONConfig<T> will try to load a file called robotconfig.json. This file should be placed in src/main/deploy in your code. When you deploy code to the robot, the file is copied to /home/lvuser/deploy on the RoboRIO.

Pre-made datatypes

Some config types are very commonly used. The same data will always be needed for defining a motor, for example. Due to this, Lib5K includes datatypes for many common config settings, which you can just drop in your java file:

• JSONMotor
• JSONEncoder
• JSONPIDGains

The JavaDoc for each of these classes also contains an example of how to represent the data in JSON.

Example

The following is an example Java object, and the JSON file to go along with it:

class DriveTrainConfig {
    JSONMotor leftMotor;
    JSONEncoder leftEncoder;
    JSONMotor rightMotor;
    JSONEncoder rightEncoder;
}

class ElevatorConfig {
    JSONMotor motor;
    JSONEncoder encoder;
}

class MyConfig {
    DriveTrainConfig drivetrain;
    ElevatorConfig elevator;
}

{
    "drivetrain": {
        "leftMotor": {
            "id": 0
        },
        "rightMotor": {
            "id": 1,
            "inverted": true
        },
        "leftEncoder": {
            "cpr":1440
        },
        "rightEncoder": {
            "cpr":1440
        }
    },
    "elevator": {
        "motor": {
            "id": 2,
            "inverted": true
        },
        "encoder": {
            "cpr":4096
        }
    }
}

Loading the example

Load the example above would be as simple as:

static SingleInstanceJSONConfig<MyConfig> configLoader = new SingleInstanceJSONConfig<>(MyConfig.class);
static MyConfig config = configLoader.getConfig();

Making Gradle pre-check your config before deploying to a robot

Lib5K provides the ConfigValidator tool to automatically check your config file for you during build time. Just add a new unit test to the robot that looks something like:

@Test public void validateConfig() {
    new ConfigValidator(myRobotConfig).validate();
}

Where myRobotConfig is a static config object defined in your actual robot code. See Lib5K's SingleInstanceJSONConfigTest class for an example.