Fix usage of List.Contains

Author: theolivenbaum

src/UglyToad.PdfPig.DocumentLayoutAnalysis/DecorationTextBlockClassifier.cs

@@ -128,13 +128,13 @@ public static IReadOnlyList<IReadOnlyList<TextBlock>> Get(IReadOnlyList<IReadOnl
                 throw new ArgumentException("The algorithm cannot be used with a document of less than 2 pages.", nameof(pagesTextBlocks));
             }
 
-            ConcurrentDictionary<int, List<TextBlock>> pageDecorations = new ConcurrentDictionary<int, List<TextBlock>>();
+            ConcurrentDictionary<int, OrderedSet<TextBlock>> pageDecorations = new ConcurrentDictionary<int, OrderedSet<TextBlock>>();
 
             ParallelOptions parallelOptions = new ParallelOptions() { MaxDegreeOfParallelism = maxDegreeOfParallelism };
 
             Parallel.For(0, pagesTextBlocks.Count, parallelOptions, p =>
             {
-                if (!pageDecorations.TryAdd(p, new List<TextBlock>()))
+                if (!pageDecorations.TryAdd(p, new OrderedSet<TextBlock>()))
                 {
                     throw new ArgumentException("Cannot add element with index " + p + " in ConcurrentDictionary.");
                 }
@@ -165,7 +165,7 @@ public static IReadOnlyList<IReadOnlyList<TextBlock>> Get(IReadOnlyList<IReadOnl
                     var score = Score(current, previousPage, nextPage, minimumEditDistanceNormalised, similarityThreshold, n);
                     if (score >= similarityThreshold)
                     {
-                        if (!pageDecorations[p].Contains(current)) pageDecorations[p].Add(current);
+                        pageDecorations[p].TryAdd(current);
                     }
                 }
 
@@ -180,7 +180,7 @@ public static IReadOnlyList<IReadOnlyList<TextBlock>> Get(IReadOnlyList<IReadOnl
                     var score = Score(current, previousPage, nextPage, minimumEditDistanceNormalised, similarityThreshold, n);
                     if (score >= similarityThreshold)
                     {
-                        if (!pageDecorations[p].Contains(current)) pageDecorations[p].Add(current);
+                        pageDecorations[p].TryAdd(current);
                     }
                 }
 
@@ -195,7 +195,7 @@ public static IReadOnlyList<IReadOnlyList<TextBlock>> Get(IReadOnlyList<IReadOnl
                     var score = Score(current, previousPage, nextPage, minimumEditDistanceNormalised, similarityThreshold, n);
                     if (score >= similarityThreshold)
                     {
-                        if (!pageDecorations[p].Contains(current)) pageDecorations[p].Add(current);
+                        pageDecorations[p].TryAdd(current);
                     }
                 }
 
@@ -210,12 +210,12 @@ public static IReadOnlyList<IReadOnlyList<TextBlock>> Get(IReadOnlyList<IReadOnl
                     var score = Score(current, previousPage, nextPage, minimumEditDistanceNormalised, similarityThreshold, n);
                     if (score >= similarityThreshold)
                     {
-                        if (!pageDecorations[p].Contains(current)) pageDecorations[p].Add(current);
+                        pageDecorations[p].TryAdd(current);
                     }
                 }
             });
 
-            return pageDecorations.OrderBy(x => x.Key).Select(x => x.Value).ToList();
+            return pageDecorations.OrderBy(x => x.Key).Select(x => x.Value.GetList()).ToList();
         }
 
         /// <summary>

src/UglyToad.PdfPig.DocumentLayoutAnalysis/OrderedSet.cs

@@ -0,0 +1,56 @@
+namespace UglyToad.PdfPig.DocumentLayoutAnalysis
+{
+    using System;
+    using System.Collections;
+    using System.Collections.Generic;
+    using System.Text;
+
+    internal class OrderedSet<T>
+    {
+        private readonly HashSet<T> _set;
+        private readonly List<T> _list;
+
+        public OrderedSet() : this(EqualityComparer<T>.Default)
+        {
+
+        }
+
+        public OrderedSet(IEqualityComparer<T> comparer)
+        {
+            _set = new HashSet<T>(comparer);
+            _list = new List<T>();
+        }
+
+        public int Count => _set.Count;
+
+        public bool TryAdd(T item)
+        {
+            if (_set.Contains(item)) return false;
+
+            _list.Add(item);
+            _set.Add(item);
+
+            return true;
+        }
+
+        public void Clear()
+        {
+            _list.Clear();
+            _set.Clear();
+        }
+
+        public bool Contains(T item)
+        {
+            return item is not null && _set.Contains(item);
+        }
+
+        public void CopyTo(T[] array, int arrayIndex)
+        {
+            _list.CopyTo(array, arrayIndex);
+        }
+        public List<T> GetList()
+        {
+            return _list;
+        }
+    }
+}

src/UglyToad.PdfPig/PdfExtensions.cs

@@ -122,12 +122,12 @@ private static double GetEstimatedSizeMultiplier(IFilter filter)
         /// Returns an equivalent token where any indirect references of child objects are
         /// recursively traversed and resolved.
         /// </summary>
-        internal static T? Resolve<T>(this T? token, IPdfTokenScanner scanner, List<IndirectReference>? visited = null) where T : IToken
+        internal static T? Resolve<T>(this T? token, IPdfTokenScanner scanner, HashSet<IndirectReference>? visited = null) where T : IToken
         {
             return (T?)ResolveInternal(token, scanner, visited ?? []);
         }
 
-        private static IToken? ResolveInternal(this IToken? token, IPdfTokenScanner scanner, List<IndirectReference> visited)
+        private static IToken? ResolveInternal(this IToken? token, IPdfTokenScanner scanner, HashSet<IndirectReference> visited)
         {
             if (token is StreamToken stream)
             {