From 30247ba774555f20bb70709b34a12c18b59f23cc Mon Sep 17 00:00:00 2001
From: Zhiguan Hu <huzhiguan@gmail.com>
Date: Tue, 10 Dec 2019 15:03:05 -0600
Subject: [PATCH 1/2] Fix the bug that happens when all the words in the
 current leaf for VerticalCut/HorizontalCut are all white spaces.

---
 .../DocumentLayoutAnalysis/RecursiveXYCut.cs  | 43 +++++++++++++++----
 1 file changed, 35 insertions(+), 8 deletions(-)

diff --git a/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs b/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs
index 8b3cbaab..8a8d05c4 100644
--- a/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs
+++ b/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs
@@ -70,11 +70,18 @@ namespace UglyToad.PdfPig.DocumentLayoutAnalysis
             XYLeaf root = new XYLeaf(pageWords); // Create a root node.
             XYNode node = VerticalCut(root, minimumWidth, dominantFontWidthFunc, dominantFontHeightFunc);
 
-            var leafs = node.GetLeafs();
-
-            if (leafs.Count > 0)
+            if (node.IsLeaf)
             {
-                return leafs.Select(l => new TextBlock(l.GetLines())).ToList();
+                return new List<TextBlock>{ new TextBlock((node as XYLeaf).GetLines())};
+            }
+            else
+            {
+                var leafs = node.GetLeafs();
+
+                if (leafs.Count > 0)
+                {
+                    return leafs.Select(l => new TextBlock(l.GetLines())).ToList();
+                }
             }
 
             return new List<TextBlock>();
@@ -84,6 +91,18 @@ namespace UglyToad.PdfPig.DocumentLayoutAnalysis
             Func<IEnumerable<decimal>, decimal> dominantFontWidthFunc,
             Func<IEnumerable<decimal>, decimal> dominantFontHeightFunc, int level = 0)
         {
+            // order words left to right
+            var words = leaf.Words.Where(w => !string.IsNullOrWhiteSpace(w.Text)).OrderBy(w => w.BoundingBox.Left).ToArray();
+
+            if(!words.Any())
+            {
+                return new XYNode(null);
+            }
+            else
+            {
+                //Create new leaf with non-whitespace words.
+                leaf = new XYLeaf(words);
+            }
             if (leaf.CountWords() <= 1 || leaf.BoundingBox.Width <= minimumWidth)
             {
                 // we stop cutting if 
@@ -92,9 +111,6 @@ namespace UglyToad.PdfPig.DocumentLayoutAnalysis
                 return leaf;
             }
 
-            // order words left to right
-            var words = leaf.Words.Where(w => !string.IsNullOrWhiteSpace(w.Text)).OrderBy(w => w.BoundingBox.Left).ToArray();
-
             // determine dominantFontWidth and dominantFontHeight
             decimal domFontWidth = dominantFontWidthFunc(words.SelectMany(x => x.Letters)
                 .Select(x => Math.Abs(x.GlyphRectangle.Width)));
@@ -177,6 +193,18 @@ namespace UglyToad.PdfPig.DocumentLayoutAnalysis
             Func<IEnumerable<decimal>, decimal> dominantFontWidthFunc,
             Func<IEnumerable<decimal>, decimal> dominantFontHeightFunc, int level = 0)
         {
+            var words = leaf.Words.Where(w => !string.IsNullOrWhiteSpace(w.Text)).OrderBy(w => w.BoundingBox.Bottom).ToArray(); // order bottom to top
+
+            if (!words.Any())
+            {
+                return new XYNode(null);
+            }
+            else
+            {
+                //Create new leaf with non-whitespace words.
+                leaf = new XYLeaf(words);
+            }
+
             if (leaf.CountWords() <= 1)
             {
                 // we stop cutting if 
@@ -184,7 +212,6 @@ namespace UglyToad.PdfPig.DocumentLayoutAnalysis
                 return leaf;
             }
 
-            var words = leaf.Words.Where(w => !string.IsNullOrWhiteSpace(w.Text)).OrderBy(w => w.BoundingBox.Bottom).ToArray(); // order bottom to top
 
             // determine dominantFontWidth and dominantFontHeight
             decimal domFontWidth = dominantFontWidthFunc(words.SelectMany(x => x.Letters)

From 9baa8c3ca0f433c40a43a8735e066ddfef53980f Mon Sep 17 00:00:00 2001
From: Zhiguan Hu <huzhiguan@gmail.com>
Date: Wed, 11 Dec 2019 10:05:17 -0600
Subject: [PATCH 2/2] Fix format as suggested.

---
 src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs b/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs
index 8a8d05c4..1476452a 100644
--- a/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs
+++ b/src/UglyToad.PdfPig/DocumentLayoutAnalysis/RecursiveXYCut.cs
@@ -103,6 +103,7 @@ namespace UglyToad.PdfPig.DocumentLayoutAnalysis
                 //Create new leaf with non-whitespace words.
                 leaf = new XYLeaf(words);
             }
+
             if (leaf.CountWords() <= 1 || leaf.BoundingBox.Width <= minimumWidth)
             {
                 // we stop cutting if 
@@ -212,7 +213,6 @@ namespace UglyToad.PdfPig.DocumentLayoutAnalysis
                 return leaf;
             }
 
-
             // determine dominantFontWidth and dominantFontHeight
             decimal domFontWidth = dominantFontWidthFunc(words.SelectMany(x => x.Letters)
                 .Select(x => Math.Abs(x.GlyphRectangle.Width)));