Enabled ClipPaths option to filter letters. (#905)

* Enabled ClipPaths option to filter letters.

* parsingOptions is not passed as  an argument now.

src/UglyToad.PdfPig.Tests/Integration/LetterFilterTests.cs

@@ -0,0 +1,46 @@
+namespace UglyToad.PdfPig.Tests.Integration
+{
+    using System.Linq;
+
+    public class LetterFilterTests
+    {
+        [Fact]
+        public void CanFilterClippedLetters()
+        {
+            var one = IntegrationHelpers.GetDocumentPath("ClipPathLetterFilter-Test1.pdf");
+
+            using (var doc1 = PdfDocument.Open(one, new ParsingOptions { ClipPaths = true }))
+            using (var doc2 = PdfDocument.Open(one, new ParsingOptions { ClipPaths = false }))
+            {
+                var allLetters = doc2.GetPage(5).Letters.Count;
+                var filteredLetters = doc1.GetPage(5).Letters.Count;
+
+                Assert.True(filteredLetters < allLetters,
+                    "Expected filtered letter count to be lower than non-filtered"); // Filtered: 3158 letters, Non-filtered: 3184 letters
+            }
+        }
+
+        [Fact]
+        public void CanFilterClippedLetters_CheckBleedInSpecificWord()
+        {
+            var one = IntegrationHelpers.GetDocumentPath("ClipPathLetterFilter-Test2.pdf");
+
+            using (var doc1 = PdfDocument.Open(one, new ParsingOptions { ClipPaths = true }))
+            using (var doc2 = PdfDocument.Open(one, new ParsingOptions { ClipPaths = false }))
+            {
+                var allWords = doc2.GetPage(1).GetWords().ToList();
+                var filteredWords = doc1.GetPage(1).GetWords().ToList();
+
+                // The table has hidden columns at the left end. Letters from these columns get merged in words 
+                // which is incorrect. Filtering letters based on clip path should fix that...
+                const string wordToSearchAfterWhichTheActualTableStarts = "ARISER";
+
+                var indexOfCheckedWordInAllWords = allWords.FindIndex(x => x.Text.Equals(wordToSearchAfterWhichTheActualTableStarts)) + 1;
+                Assert.True(allWords[indexOfCheckedWordInAllWords].Text == "MLIA0U01CP00O0I3N6G2");
+
+                var indexOfCheckedWordInFilteredWords = filteredWords.FindIndex(x => x.Text.Equals(wordToSearchAfterWhichTheActualTableStarts)) + 1;
+                Assert.True(filteredWords[indexOfCheckedWordInFilteredWords].Text == "ACOGUT");
+            }
+        }
+    }
+}

src/UglyToad.PdfPig.Tests/Integration/StreamProcessorTests.cs

@@ -140,9 +140,7 @@
             }
 
             public override void RenderGlyph(IFont font,
-                IColor strokingColor,
-                IColor nonStrokingColor,
-                TextRenderingMode textRenderingMode,
+                CurrentGraphicsState currentState,
                 double fontSize,
                 double pointSize,
                 int code,

src/UglyToad.PdfPig.Tests/Integration/VisualVerification/SkiaHelpers/SkiaGlyphStreamProcessor.cs

@@ -61,9 +61,7 @@
         }
 
         public override void RenderGlyph(IFont font,
-            IColor strokingColor,
-            IColor nonStrokingColor,
-            TextRenderingMode textRenderingMode,
+            CurrentGraphicsState currentState,
             double fontSize,
             double pointSize,
             int code,
@@ -74,6 +72,7 @@
             in TransformationMatrix transformationMatrix,
             CharacterBoundingBox characterBoundingBox)
         {
+            var textRenderingMode = currentState.FontState.TextRenderingMode;
             if (textRenderingMode == TextRenderingMode.Neither)
             {
                 return;
@@ -85,7 +84,7 @@
             if (font.TryGetNormalisedPath(code, out var path))
             {
                 var skPath = path.ToSKPath();
-                ShowVectorFontGlyph(skPath, strokingColor, nonStrokingColor, textRenderingMode, renderingMatrix,
+                ShowVectorFontGlyph(skPath,  currentState.CurrentStrokingColor!,  currentState.CurrentNonStrokingColor!, textRenderingMode, renderingMatrix,
                     textMatrix, transformationMatrix);
             }
             else

src/UglyToad.PdfPig/Geometry/GeometryExtensions.cs

@@ -517,6 +517,26 @@
             }
         }
 
+        /// <summary>
+        /// Whether the one of rectangle corners is located inside the path.
+        /// </summary>
+        /// <param name="path">The path that should contain the rectangle corner.</param>
+        /// <param name="rectangle">The rectangle that should be intersected within the path.</param>
+        /// <param name="includeBorder">If set to false, will return false if the rectangle is on the path's border.</param>
+        public static bool IntersectsWith(this PdfPath path, PdfRectangle rectangle, bool includeBorder = false)
+        {
+            // NB, For later dev: Might not work for concave outer path, as it can contain all the points of the inner rectangle, but have overlapping edges.
+            var clipperPaths = path.Select(sp => sp.ToClipperPolygon().ToList()).ToList();
+            var fillType = path.FillingRule == FillingRule.NonZeroWinding ? ClipperPolyFillType.NonZero : ClipperPolyFillType.EvenOdd;
+            foreach (var point in rectangle.ToClipperPolygon())
+            {
+                if (PointInPaths(point, clipperPaths, fillType, includeBorder))
+                    return true;
+            }
+
+            return false;
+        }
+        
         /// <summary>
         /// Gets the <see cref="PdfRectangle"/> that is the intersection of two rectangles.
         /// <para>Only works for axis-aligned rectangles.</para>

src/UglyToad.PdfPig/Graphics/BaseStreamProcessor.cs

@@ -282,9 +282,7 @@
                 var boundingBox = font.GetBoundingBox(code);
 
                 RenderGlyph(font,
-                    currentState.CurrentStrokingColor!,
-                    currentState.CurrentNonStrokingColor!,
-                    currentState.FontState.TextRenderingMode,
+                    currentState,
                     fontSize,
                     pointSize,
                     code,
@@ -317,9 +315,7 @@
         /// Render glyph implement.
         /// </summary>
         public abstract void RenderGlyph(IFont font,
-            IColor strokingColor,
-            IColor nonStrokingColor,
-            TextRenderingMode textRenderingMode,
+            CurrentGraphicsState currentState, 
             double fontSize,
             double pointSize,
             int code,

src/UglyToad.PdfPig/Graphics/ContentStreamProcessor.cs

@@ -87,9 +87,7 @@ namespace UglyToad.PdfPig.Graphics
         }
 
         public override void RenderGlyph(IFont font,
-            IColor strokingColor,
-            IColor nonStrokingColor,
-            TextRenderingMode textRenderingMode,
+            CurrentGraphicsState currentState,
             double fontSize,
             double pointSize,
             int code,
@@ -109,6 +107,13 @@ namespace UglyToad.PdfPig.Graphics
                     transformationMatrix,
                     new PdfRectangle(0, 0, characterBoundingBox.Width, 0));
 
+            if (ParsingOptions.ClipPaths)
+            {
+                var currentClipping = currentState.CurrentClippingPath;
+                if (currentClipping?.IntersectsWith(transformedGlyphBounds) == false)
+                    return;
+            }
+
             Letter letter = null;
             if (Diacritics.IsInCombiningDiacriticRange(unicode) && currentOffset > 0 && letters.Count > 0)
             {
@@ -147,9 +152,9 @@ namespace UglyToad.PdfPig.Graphics
                     transformedPdfBounds.Width,
                     fontSize,
                     font.Details,
-                    textRenderingMode,
-                    strokingColor,
-                    nonStrokingColor,
+                     currentState.FontState.TextRenderingMode,
+                     currentState.CurrentStrokingColor!,
+                     currentState.CurrentNonStrokingColor!,
                     pointSize,
                     TextSequence);
             }