Simple recommendation system with Scala

Main.scala

// Oto Brglez - July 2025

// Our entity
final case class Book(title: String)

// Sample data. Books. yey!
val books @ List(programmingInScala, jsGoodParts, _*) = List(
  Book("Programming in Scala"),
  Book("JavaScript: The Good Parts"),
  Book("Designing Data-Intensive Applications"),
  Book("Grokking Functional Programming"),
  Book("Programming Scala"),
  Book("Clean Architecture"),
  Book("Hands-on Scala Programming"),
  Book("Functional Programming in Scala"),
  Book("The Clean Coder"),
  Book("Clean Code")
): @unchecked

// Generic recommend function that takes item and other items.
// Items are sorted by the result of the "distance function". Nothing fancy.
def recommend[I](
  item: I,
  items: Seq[I],
  distanceF: (I, I) => Double
): Seq[(I, Double)] =
  items
    .filterNot(_ == item)
    .map(other => other -> distanceF(item, other))
    .sortBy(_._2)

// Set of unique boot title words, longer than 3 characters.
def words(book: Book): Set[String] =
  "[a-zA-Z]{3,}".r
    .findAllMatchIn(book.title)
    .map(_.toString.toLowerCase)
    .toSet

// Compute the Jaccard Index/Distance between word sets.
def jaccardDistance(a: Book, b: Book): Double =
  val intersection = words(a) intersect words(b)
  val union        = words(a) union words(b)
  1L - intersection.size.toDouble / union.size.toDouble

// Euclidian distance with the help of diffs between word sets.
def euclideanDistance(a: Book, b: Book): Double =
  val (words1, words2) = words(a) -> words(b)
  val symDiffSize      =
    ((words1 diff words2) union (words2 diff words1)).size
  math.sqrt(symDiffSize.toDouble)

// Sample usage with two different distance functions
@main def main(): Unit =
  println(s"Books related to $programmingInScala (jaccard distance):")
  recommend(
    programmingInScala,
    books,
    jaccardDistance
  ).take(5).foreach(println)

  println(s"\nBooks related to $programmingInScala (euclidean distance):")
  recommend(
    programmingInScala,
    books,
    euclideanDistance
  ).take(5).foreach(println)