Add benchmark project, some small optimizations (#39)

* Add benchmark project, some small optimizations * Use .NET 8 SDK in build pipeline * Update unit test project to .NET 8 * Make GetProfile protected internal
feature23 · Aug 12, 2024 · 248ef9e · 248ef9e
1 parent f2b0321
commit 248ef9e
Show file tree

Hide file tree

Showing 10 changed files with 161 additions and 21 deletions.
diff --git a/F23.StringSimilarity.Benchmarks/Benchmarks.cs b/F23.StringSimilarity.Benchmarks/Benchmarks.cs
@@ -0,0 +1,117 @@
+using BenchmarkDotNet.Attributes;
+
+namespace F23.StringSimilarity.Benchmarks;
+
+[MemoryDiagnoser]
+public class Benchmarks
+{
+    [Benchmark]
+    public void Cosine()
+    {
+        var cosine = new Cosine();
+        _ = cosine.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void Damerau()
+    {
+        var damerau = new Damerau();
+        _ = damerau.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void Jaccard()
+    {
+        var jaccard = new Jaccard();
+        _ = jaccard.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void JaroWinkler()
+    {
+        var jaro = new JaroWinkler();
+        _ = jaro.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void Levenshtein()
+    {
+        var levenshtein = new Levenshtein();
+        _ = levenshtein.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void LongestCommonSubsequence()
+    {
+        var lcs = new LongestCommonSubsequence();
+        _ = lcs.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void MetricLCS()
+    {
+        var metricLcs = new MetricLCS();
+        _ = metricLcs.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void NGram()
+    {
+        var ngram = new NGram();
+        _ = ngram.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void NormalizedLevenshtein()
+    {
+        var normalizedLevenshtein = new NormalizedLevenshtein();
+        _ = normalizedLevenshtein.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void OptimalStringAlignment()
+    {
+        var osa = new OptimalStringAlignment();
+        _ = osa.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void QGram()
+    {
+        var qGram = new QGram();
+        _ = qGram.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void RatcliffObershelp()
+    {
+        var ratcliffObershelp = new RatcliffObershelp();
+        _ = ratcliffObershelp.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void SorensenDice()
+    {
+        var sorensenDice = new SorensenDice();
+        _ = sorensenDice.Distance("hello", "world");
+    }
+
+    [Benchmark]
+    public void WeightedLevenshtein()
+    {
+        var weightedLevenshtein = new WeightedLevenshtein(new ExampleCharSub());
+        _ = weightedLevenshtein.Distance("hello", "world");
+    }
+
+    private class ExampleCharSub : ICharacterSubstitution
+    {
+        public double Cost(char c1, char c2)
+        {
+            // The cost for substituting 't' and 'r' is considered smaller as these 2 are located next to each other on a keyboard
+            if (c1 == 't' && c2 == 'r') return 0.5;
+
+            // For most cases, the cost of substituting 2 characters is 1.0
+            return 1.0;
+        }
+    }
+}
diff --git a/F23.StringSimilarity.Benchmarks/F23.StringSimilarity.Benchmarks.csproj b/F23.StringSimilarity.Benchmarks/F23.StringSimilarity.Benchmarks.csproj
@@ -0,0 +1,18 @@
+<Project Sdk="Microsoft.NET.Sdk">
+
+    <PropertyGroup>
+        <OutputType>Exe</OutputType>
+        <TargetFramework>net8.0</TargetFramework>
+        <ImplicitUsings>enable</ImplicitUsings>
+        <Nullable>enable</Nullable>
+    </PropertyGroup>
+
+    <ItemGroup>
+      <PackageReference Include="BenchmarkDotNet" Version="0.13.12" />
+    </ItemGroup>
+
+    <ItemGroup>
+      <ProjectReference Include="..\src\F23.StringSimilarity\F23.StringSimilarity.csproj" />
+    </ItemGroup>
+
+</Project>
diff --git a/F23.StringSimilarity.Benchmarks/Program.cs b/F23.StringSimilarity.Benchmarks/Program.cs
@@ -0,0 +1,4 @@
+using BenchmarkDotNet.Running;
+using F23.StringSimilarity.Benchmarks;
+
+BenchmarkRunner.Run<Benchmarks>();
diff --git a/F23.StringSimilarity.sln b/F23.StringSimilarity.sln
@@ -7,6 +7,8 @@ Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "F23.StringSimilarity", "src
 EndProject
 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "F23.StringSimilarity.Tests", "test\F23.StringSimilarity.Tests\F23.StringSimilarity.Tests.csproj", "{68F339E6-278F-4B04-A6ED-422AAD30591F}"
 EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "F23.StringSimilarity.Benchmarks", "F23.StringSimilarity.Benchmarks\F23.StringSimilarity.Benchmarks.csproj", "{3A9605B1-820C-43C2-8F9B-72BCA5F5543B}"
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU
@@ -21,6 +23,10 @@ Global
 		{68F339E6-278F-4B04-A6ED-422AAD30591F}.Debug|Any CPU.Build.0 = Debug|Any CPU
 		{68F339E6-278F-4B04-A6ED-422AAD30591F}.Release|Any CPU.ActiveCfg = Release|Any CPU
 		{68F339E6-278F-4B04-A6ED-422AAD30591F}.Release|Any CPU.Build.0 = Release|Any CPU
+		{3A9605B1-820C-43C2-8F9B-72BCA5F5543B}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{3A9605B1-820C-43C2-8F9B-72BCA5F5543B}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{3A9605B1-820C-43C2-8F9B-72BCA5F5543B}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{3A9605B1-820C-43C2-8F9B-72BCA5F5543B}.Release|Any CPU.Build.0 = Release|Any CPU
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE

diff --git a/src/F23.StringSimilarity/Jaccard.cs b/src/F23.StringSimilarity/Jaccard.cs
@@ -24,6 +24,7 @@
 
 using System;
 using System.Collections.Generic;
+using System.Linq;
 using F23.StringSimilarity.Interfaces;
 
 // ReSharper disable LoopCanBeConvertedToQuery
@@ -83,14 +84,13 @@ public double Similarity(string s1, string s2)
             var profile1 = GetProfile(s1);
             var profile2 = GetProfile(s2);
 
-            var union = new HashSet<string>();
-            union.UnionWith(profile1.Keys);
-            union.UnionWith(profile2.Keys);
+            // SSNET Specific: use LINQ for more optimal distinct count
+            var unionCount = profile1.Keys.Concat(profile2.Keys).Distinct().Count();
 
             int inter = profile1.Keys.Count + profile2.Keys.Count
-                        - union.Count;
+                        - unionCount;
 
-            return 1.0 * inter / union.Count;
+            return 1.0 * inter / unionCount;
         }
 
 

diff --git a/src/F23.StringSimilarity/Levenshtein.cs b/src/F23.StringSimilarity/Levenshtein.cs
@@ -109,7 +109,7 @@ public double Distance<T>(ReadOnlySpan<T> s1, ReadOnlySpan<T> s2, int limit)
             // create two work vectors of integer distances
             int[] v0 = new int[s2.Length + 1];
             int[] v1 = new int[s2.Length + 1];
-            int[] vtemp;
+            // SSNET: removed unneeded int[] vtemp;
 
             // initialize v0 (the previous row of distances)
             // this row is A[0][i]: edit distance for an empty s
@@ -155,9 +155,7 @@ public double Distance<T>(ReadOnlySpan<T> s1, ReadOnlySpan<T> s2, int limit)
                 // System.arraycopy(v1, 0, v0, 0, v0.length);
 
                 // Flip references to current and previous row
-                vtemp = v0;
-                v0 = v1;
-                v1 = vtemp;
+                (v0, v1) = (v1, v0); // SSNET specific: Swap v0 and v1 using tuples
             }
 
             return v0[s2.Length];

diff --git a/src/F23.StringSimilarity/NGram.cs b/src/F23.StringSimilarity/NGram.cs
@@ -103,7 +103,7 @@ public double Distance(string s0, string s1)
             char[] sa = new char[sl + n - 1];
             float[] p; // 'previous' cost array, horizontally
             float[] d; // Cost array, horizontally
-            float[] d2; // Placeholder to assist in swapping p and d
+            // SSNET removed unneeded: float[] d2; // Placeholder to assist in swapping p and d
 
             // Construct sa with prefix
             for (int i1 = 0; i1 < sa.Length; i1++)
@@ -172,9 +172,7 @@ public double Distance(string s0, string s1)
                     d[i] = Math.Min(Math.Min(d[i - 1] + 1, p[i] + 1), p[i - 1] + ec);
                 }
                 // Copy current distance counts to 'previous row' distance counts
-                d2 = p;
-                p = d;
-                d = d2;
+                (p, d) = (d, p); // SSNET specific: swap p and d using tuples
             }
 
             // Our last action in the above loop was to switch d and p, so p now

diff --git a/src/F23.StringSimilarity/ShingleBased.cs b/src/F23.StringSimilarity/ShingleBased.cs
@@ -41,7 +41,7 @@ public abstract class ShingleBased
         /// <summary>
         /// Pattern for finding multiple following spaces
         /// </summary>
-        private static readonly Regex SPACE_REG = new Regex("\\s+");
+        private static readonly Regex SPACE_REG = new Regex("\\s+", RegexOptions.Compiled);
 
         /// <summary> 
         /// </summary>
@@ -59,7 +59,7 @@ protected ShingleBased(int k)
 
         protected ShingleBased() : this(DEFAULT_K) { }
 
-        public IDictionary<string, int> GetProfile(string s)
+        protected internal Dictionary<string, int> GetProfile(string s)
         {
             var shingles = new Dictionary<string, int>();
 
@@ -79,7 +79,7 @@ public IDictionary<string, int> GetProfile(string s)
                 }
             }
 
-            return new ReadOnlyDictionary<string, int>(shingles);
+            return shingles;
         }
     }
 }
diff --git a/src/F23.StringSimilarity/WeightedLevenshtein.cs b/src/F23.StringSimilarity/WeightedLevenshtein.cs
@@ -23,6 +23,7 @@
  */
 
 using System;
+using System.Threading;
 using F23.StringSimilarity.Interfaces;
 // ReSharper disable SuggestVarOrType_Elsewhere
 // ReSharper disable TooWideLocalVariableScope
@@ -112,7 +113,7 @@ public double Distance(string s1, string s2, double limit)
             // create two work vectors of floating point (i.e. weighted) distances
             double[] v0 = new double[s2.Length + 1];
             double[] v1 = new double[s2.Length + 1];
-            double[] vtemp;
+            // SSNET: removed unneeded double[] vtemp;
 
             // initialize v0 (the previous row of distances)
             // this row is A[0][i]: edit distance for an empty s1
@@ -166,9 +167,7 @@ public double Distance(string s1, string s2, double limit)
                 // copy v1 (current row) to v0 (previous row) for next iteration
                 // System.arraycopy(v1, 0, v0, 0, v0.length);
                 // Flip references to current and previous row
-                vtemp = v0;
-                v0 = v1;
-                v1 = vtemp;
+                (v0, v1) = (v1, v0); // SSNET Specific: Swap references using tuples instead of temporary
             }
 
             return v0[s2.Length];

diff --git a/test/F23.StringSimilarity.Tests/F23.StringSimilarity.Tests.csproj b/test/F23.StringSimilarity.Tests/F23.StringSimilarity.Tests.csproj
@@ -30,4 +30,4 @@
     <ProjectReference Include="..\..\src\F23.StringSimilarity\F23.StringSimilarity.csproj" />
   </ItemGroup>
 
-</Project>
+</Project>