From a23e6458bce37504f5d359a4b6bb881bdad99989 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Holger=20B=C3=B6rchers?= <holger@boerchers.org>
Date: Thu, 9 Dec 2021 17:12:32 +0100
Subject: [PATCH] cleanup code.

---
 src/Helper.cs  | 14 +++++++++-----
 src/Point.cs   | 12 ++++++++++--
 src/Program.cs | 34 +++++++++++++++++++---------------
 3 files changed, 38 insertions(+), 22 deletions(-)

diff --git a/src/Helper.cs b/src/Helper.cs
index 8e02c85..79269db 100644
--- a/src/Helper.cs
+++ b/src/Helper.cs
@@ -1,7 +1,7 @@
-using System.Diagnostics;
+using ScottPlot;
+using System.Diagnostics;
 using System.Drawing;
 using System.Globalization;
-using ScottPlot;
 
 namespace kMeans;
 
@@ -33,11 +33,9 @@ public static class Helper
         return plot;
     }
 
-    public static void ExportAndPreviewPlot(Plot plot, string path)
+    public static void ExportPlot(Plot plot, string path)
     {
         plot.SaveFig(path);
-        using var p = Process.Start(new ProcessStartInfo(path) { UseShellExecute = true });
-        p?.WaitForExit();
     }
 
     public static IEnumerable<Point> ParseCsv(string path)
@@ -51,4 +49,10 @@ public static class Helper
             yield return new Point(x, y, -1);
         }
     }
+
+    public static void PreviewPlot(string path)
+    {
+        using var p = Process.Start(new ProcessStartInfo(path) { UseShellExecute = true });
+        p?.WaitForExit();
+    }
 }
\ No newline at end of file
diff --git a/src/Point.cs b/src/Point.cs
index 6bf034e..565c0c9 100644
--- a/src/Point.cs
+++ b/src/Point.cs
@@ -3,6 +3,14 @@
 public record Point(double X, double Y, int ClusterId)
 {
     public int ClusterId { get; set; } = ClusterId;
-    public double X { get; set; } = X;
-    public double Y { get; set; } = Y;
+    public double X { get; private set; } = X;
+    public double Y { get; private set; } = Y;
+
+    public bool SetCoordinates(double x, double y)
+    {
+        if (!(Math.Abs(X - x) > 1e-5) || !(Math.Abs(Y - y) > 1e-5)) return false;
+        X = x;
+        Y = y;
+        return true;
+    }
 }
\ No newline at end of file
diff --git a/src/Program.cs b/src/Program.cs
index 2538051..4bf48c7 100644
--- a/src/Program.cs
+++ b/src/Program.cs
@@ -6,6 +6,7 @@ public static class Program
 {
     private const string Coordinates = "coordinates.csv";
     private const int K = 3;
+    private const int MaxLoops = 30;
     private const string PlotOut = "plot.png";
 
     public static void Main()
@@ -13,19 +14,14 @@ public static class Program
         Console.WriteLine("k-Means-Algorithm");
         var points = ParseCsv(Coordinates).ToList();
 
-        var centroids = InitializeRandomlyCentroids(points, K).ToList();
-        var update = true;
-        while (update)
-        {
-            AssignPointsToCentroids(points, centroids);
-            update = UpdateCentroids(points, centroids);
-        }
+        KMeansCalculation(points, K, out var centroids);
 
         var plot = CreatePlot(points, centroids);
-        ExportAndPreviewPlot(plot, PlotOut);
+        ExportPlot(plot, PlotOut);
+        PreviewPlot(PlotOut);
     }
 
-    private static void AssignPointsToCentroids(List<Point> points, List<Point> centroids)
+    private static void AssignPointsToCentroids(IEnumerable<Point> points, IReadOnlyCollection<Point> centroids)
     {
         foreach (var point in points)
         {
@@ -49,7 +45,7 @@ public static class Program
         return Math.Sqrt(Math.Pow(centroid.X - point.X, 2) + Math.Pow(centroid.Y - point.Y, 2));
     }
 
-    private static IEnumerable<Point> InitializeRandomlyCentroids(IReadOnlyCollection<Point> points, int k)
+    private static IEnumerable<Point> InitializeRandomCentroids(IReadOnlyCollection<Point> points, int k)
     {
         var minX = points.Min(p => p.X);
         var maxX = points.Max(p => p.X);
@@ -67,7 +63,18 @@ public static class Program
         }
     }
 
-    private static bool UpdateCentroids(IReadOnlyCollection<Point> points, List<Point> centroids)
+    private static void KMeansCalculation(IReadOnlyCollection<Point> points, int k, out IReadOnlyCollection<Point> centroids)
+    {
+        centroids = InitializeRandomCentroids(points, k).ToArray();
+
+        for (var i = 0; i < MaxLoops; i++)
+        {
+            AssignPointsToCentroids(points, centroids);
+            if (!UpdateCentroids(points, centroids)) break;
+        }
+    }
+
+    private static bool UpdateCentroids(IReadOnlyCollection<Point> points, IEnumerable<Point> centroids)
     {
         // calculate mean of all points of one cluster.
         var updated = false;
@@ -87,9 +94,6 @@ public static class Program
         var sumY = pointsOfCluster.Sum(p => p.Y);
         var meanY = sumY / pointsOfCluster.Length;
 
-        if (!(Math.Abs(centroid.X - meanX) > 1e-5) || !(Math.Abs(centroid.Y - meanY) > 1e-5)) return false;
-        centroid.X = meanX;
-        centroid.Y = meanY;
-        return true;
+        return centroid.SetCoordinates(meanX, meanY);
     }
 }
\ No newline at end of file