From 731a1a3956d7db6226c97df829f2ad71116bf961 Mon Sep 17 00:00:00 2001
From: Eliot Jones <elioty@hotmail.co.uk>
Date: Sat, 14 Apr 2018 19:18:19 +0100
Subject: [PATCH] add xfinium data validations for non latin characters

---
 .../Integration/AssertablePositionData.cs     |  6 +-
 ...SinglePageNonLatinAcrobatDistillerTests.cs | 60 +++++++++++++++++++
 2 files changed, 65 insertions(+), 1 deletion(-)

diff --git a/src/UglyToad.PdfPig.Tests/Integration/AssertablePositionData.cs b/src/UglyToad.PdfPig.Tests/Integration/AssertablePositionData.cs
index 0a520e79..b38a3170 100644
--- a/src/UglyToad.PdfPig.Tests/Integration/AssertablePositionData.cs
+++ b/src/UglyToad.PdfPig.Tests/Integration/AssertablePositionData.cs
@@ -16,15 +16,19 @@
 
         public string FontName { get; set; }
 
+        public decimal Height { get; set; }
+
         public static AssertablePositionData Parse(string line)
         {
             var parts = line.Split('\t', StringSplitOptions.None);
 
-            if (parts.Length != 6)
+            if (parts.Length < 6)
             {
                 throw new ArgumentException($"Expected 6 parts to the line, instead got {parts.Length}");
             }
 
+            var height = parts.Length < 7 ? 0 : decimal.Parse(parts[6]);
+
             return new AssertablePositionData
             {
                 X = decimal.Parse(parts[0]),
diff --git a/src/UglyToad.PdfPig.Tests/Integration/SinglePageNonLatinAcrobatDistillerTests.cs b/src/UglyToad.PdfPig.Tests/Integration/SinglePageNonLatinAcrobatDistillerTests.cs
index 84aea6d3..81e8c3a2 100644
--- a/src/UglyToad.PdfPig.Tests/Integration/SinglePageNonLatinAcrobatDistillerTests.cs
+++ b/src/UglyToad.PdfPig.Tests/Integration/SinglePageNonLatinAcrobatDistillerTests.cs
@@ -84,6 +84,46 @@
                     Assert.Equal(theirLetter, myLetter);
                     Assert.Equal(theirX, myX, 2);
 
+                    Assert.Equal(pdfBoxData[index].Width, pageLetter.Width, 1);
+
+                    index++;
+                }
+            }
+        }
+
+        [Fact]
+        public void LetterPositionsAreCorrectXfinium()
+        {
+            using (var document = PdfDocument.Open(GetFilename()))
+            {
+                var page = document.GetPage(1);
+
+                var positions = GetXfiniumPositionData();
+
+                var index = 0;
+                foreach (var pageLetter in page.Letters)
+                {
+                    if (index >= positions.Count)
+                    {
+                        break;
+                    }
+
+                    var myX = pageLetter.Location.X;
+                    var theirX = positions[index].X;
+
+                    var myLetter = pageLetter.Value;
+                    var theirLetter = positions[index].Text;
+
+                    if (myLetter == " " && theirLetter != " ")
+                    {
+                        continue;
+                    }
+
+                    Assert.Equal(theirLetter, myLetter);
+                    Assert.Equal(theirX, myX, 2);
+
+                    Assert.Equal(positions[index].Width, pageLetter.Width, 1);
+
                     index++;
                 }
             }
@@ -113,5 +153,25 @@
 
             return result;
         }
+
+        private static IReadOnlyList<AssertablePositionData> GetXfiniumPositionData()
+        {
+            const string data = @"90	90.66	14.439546	H	19	FFJICI+TimesNewRomanPSMT	17.802
+104.4395	90.66	8.885106	e	19	FFJICI+TimesNewRomanPSMT	17.80218
+113.3247	90.66	5.568426	l	19	FFJICI+TimesNewRomanPSMT	17.80218
+118.8931	90.66	5.568426	l	19	FFJICI+TimesNewRomanPSMT	17.80218
+124.4615	90.66	10.003986	o	19	FFJICI+TimesNewRomanPSMT	17.80218
+139.4505	90.66	6.727266	ﺪ	19	FFJIAH+TimesNewRomanPSMT	17.80218
+146.1778	90.66	7.866126	ﻤ	19	FFJIAH+TimesNewRomanPSMT	17.80218
+154.0439	90.66	10.583406	ﺤ	19	FFJIAH+TimesNewRomanPSMT	17.80218
+164.6273	90.66	7.866126	ﻣ	19	FFJIAH+TimesNewRomanPSMT	17.80218
+177.4964	90.66	18.86112	W	19	FFJICI+TimesNewRomanPSMT	17.80218";
+
+            var result = data.Split(new[] { "\r", "\n", "\r\n" }, StringSplitOptions.RemoveEmptyEntries)
+                .Select(AssertablePositionData.Parse)
+                .ToList();
+
+            return result;
+        }
     }
 }