Chain of responsability with Scala partial functions

Partial functions

In this article, we will explore a fundamental yet highly useful feature of Scala: partial functions.

A partial function (as opposed to a total function) is a function that is not defined for all possible inputs. A partial function g: A => B is a function for which there exist some values a in the domain A such that g(a) is not defined. Scala has a good support of partial functions.

Chain of responsability

At Devoxx France 2023, Edson Yanaga gave a talk in which he revisits the GoF design patterns using the new features of Java. Here is his implementation of the Chain of Responsability pattern. During the talk, I thought to myself that it’s even easier in Scala.

For illustration purposes, let’s take the example of the FizzBuzz kata.

Here is the implementation in Scala:

object FizzBuzz extends App {
  val multipleOfThree: PartialFunction[Int, String] = {
    case i if i % 3 == 0 => "Fizz"
  }
  val multipleOfFive: PartialFunction[Int, String] = {
    case i if i % 5 == 0 => "Buzz"
  }
  val multipleOfBoth: PartialFunction[Int, String] = {
    case i if i % 3 == 0 && i % 5 == 0 => "FizzBuzz"
  }
  val default: PartialFunction[Int, String] = {
    case i => i.toString
  }

  val fizzBuzz = List(multipleOfBoth, multipleOfThree, multipleOfFive, default)
    .reduce(_ orElse _)

  (1 to 100).foreach(i => println(fizzBuzz(i)))
}

The Chain of Responsibility is implemented using the orElse method to chain the calls.

I also frequently use this technique to implement my heuristics in bot programming competitions on codingame. This allows me to easily change the priorities of actions just by changing the order in the list.

Partial functions

Chain of responsability

Links