From 16b66965559195f80cf2741e8a7dbed0cdd4d0a4 Mon Sep 17 00:00:00 2001
From: Kali Live user <charles.beauville@epfl.ch>
Date: Sun, 15 Mar 2020 12:02:25 +0000
Subject: [PATCH] Test 2

---
 src/main/scala/kmeans/KMeans.scala | 22 +++++++++++-----------
 1 file changed, 11 insertions(+), 11 deletions(-)

diff --git a/src/main/scala/kmeans/KMeans.scala b/src/main/scala/kmeans/KMeans.scala
index 1f65829..6e26635 100644
--- a/src/main/scala/kmeans/KMeans.scala
+++ b/src/main/scala/kmeans/KMeans.scala
@@ -45,10 +45,9 @@ class KMeans extends KMeansInterface {
   }
 
   def classify(points: ParSeq[Point], means: ParSeq[Point]): ParMap[Point, ParSeq[Point]] = {
-    val meansWithPoints = points.groupBy( p => findClosest(p, means) )
-    means
-      .map(m => (m, if (meansWithPoints.contains(m)) List() ++ meansWithPoints(m) else List()))
-      .toMap
+    val pointsMeanMap = points.par.groupBy(findClosest(_, means))
+    // So iterate over means get (empty) list and return map
+    means.par.map(mean => mean -> pointsMeanMap.getOrElse(mean, ParSeq())).toMap
   }
 
   def findAverage(oldMean: Point, points: ParSeq[Point]): Point = if (points.isEmpty) oldMean else {
@@ -64,20 +63,21 @@ class KMeans extends KMeansInterface {
   }
 
   def update(classified: ParMap[Point, ParSeq[Point]], oldMeans: ParSeq[Point]): ParSeq[Point] = {
-    oldMeans.map( mean => findAverage(mean, classified(mean)) )
+    oldMeans.par.map(oldMean => findAverage(oldMean, classified(oldMean)))
   }
 
   def converged(eta: Double, oldMeans: ParSeq[Point], newMeans: ParSeq[Point]): Boolean = {
-    oldMeans
-      .zip(newMeans)
-      .map(p => p._1.squareDistance(p._2) <= eta)
-      .forall( t => t )
+    (oldMeans zip newMeans).forall{
+      case (oldMean, newMean) => oldMean.squareDistance(newMean) <=  eta
+    }
   }
 
   @tailrec
   final def kMeans(points: ParSeq[Point], means: ParSeq[Point], eta: Double): ParSeq[Point] = {
-    val freshMints = update(classify(points, means), means)
-    if (!converged(eta)(means, freshMints)) kMeans(points, freshMints, eta) else freshMints // your implementation need to be tail recursive
+    val classified = classify(points, means)
+    val newMeans = update(classified, means)
+
+    if (!converged(eta, means, newMeans)) kMeans(points, newMeans, eta) else newMeans
   }
 }