Fixed issue with :::image::: handling.

Not properly closing the tag when a long caption was provided.

Author: markjulmar

.claude/settings.local.json

@@ -0,0 +1,8 @@
+{
+  "permissions": {
+    "allow": [
+      "Bash(git -C /Users/mark/git/ConvertLearnToDoc log --oneline -20)",
+      "Bash(git -C /Users/mark/git/ConvertLearnToDoc remote -v)"
+    ]
+  }
+}

lib/DocsToMarkdown/Converters/DivSpan.cs

@@ -18,7 +18,8 @@ internal class DivSpan() : BaseConverter("div", "span")
         { "has-pivot", ZonePivotSection},
         { "xp-tag-hexagon", IgnoreBlock },
         { "embeddedvideo", EmbeddedVideo },
-        { "tabGroup", TabGroup }
+        { "tabGroup", TabGroup },
+        { "visually-hidden", IgnoreBlock}
     };
 
     private static string TabGroup(string className, HtmlConverter converter, HtmlNode htmlNode)

lib/DocsToMarkdown/Converters/Image.cs

@@ -44,7 +44,18 @@ public override string Convert(HtmlConverter converter, HtmlNode htmlInput)
                             "type=\"content\"";
             var border = hasBorder ? "border=\"true\" " : "";
             var lightbox = lightBoxUrl != null ? $" lightbox=\"{lightBoxUrl}\"" : "";
-            imageText = $":::image {type} {border}source=\"{source}\" alt-text=\"{altText}\"{lightbox}:::";
+
+            var captionText = LookAheadForImageCaptionText(htmlInput);
+
+            var sb = new StringBuilder(
+                $":::image {type} {border}source=\"{source}\" alt-text=\"{altText}\"{lightbox}:::");
+            if (captionText.Length > 0)
+            {
+                sb.AppendLine(captionText);
+                sb.Append(":::image-end:::");
+            }
+
+            imageText = sb.ToString();
         }
 
         else if (lightBoxUrl != null)
@@ -64,4 +75,15 @@ public override string Convert(HtmlConverter converter, HtmlNode htmlInput)
 
         return converter.ParentPrefix != "" ? Environment.NewLine + imageText : imageText;
     }
+
+    private static string LookAheadForImageCaptionText(HtmlNode htmlInput)
+    {
+        // <p><img src="test.png" alt="test"></p>
+        // <div class="visually-hidden"><p>
+        // Some Graphic.</p></div>
+        var testNode = htmlInput.ParentNode.NextSibling;
+        return testNode?.HasClass("visually-hidden") == true
+            ? testNode.InnerText.Replace("\n", Environment.NewLine)
+            : string.Empty;
+    }
 }

lib/DocsToMarkdown/Converters/Text.cs

@@ -8,6 +8,7 @@ internal class Text() : BaseConverter("#text")
     public override string Convert(HtmlConverter converter, HtmlNode htmlInput)
     {
         Debug.Assert(CanConvert(htmlInput));
-        return System.Net.WebUtility.HtmlDecode(htmlInput.InnerText);
+        return System.Net.WebUtility.HtmlDecode(htmlInput.InnerText)
+            .Replace(0xa0.ToString(), " ");
     }
 }

lib/Markdig.Renderer.Docx/DocxRenderer.cs

@@ -1,6 +1,7 @@
 using System.Diagnostics;
 using System.Drawing;
 using System.Reflection;
+using System.Text;
 using Markdig.Extensions.Yaml;
 using Markdig.Renderer.Docx.Blocks;
 using Markdig.Renderer.Docx.Inlines;
@@ -245,8 +246,9 @@ public Drawing InsertImage(Paragraph currentParagraph, MarkdownObject owner, str
             picture.Description = altText;
             currentParagraph.Add(picture);
 
-            if (!string.IsNullOrEmpty(title))
-                picture.Drawing.AddCaption(": " + title);
+            var captionText = SanitizeCaptionText(title);
+            if (!string.IsNullOrEmpty(captionText))
+                picture.Drawing.AddCaption(": " + captionText);
 
             return picture.Drawing;
         }
@@ -292,6 +294,21 @@ private string DetermineContentTypeFromUrl(string imageUrl)
 
     private static bool IsInternetUrl(string path) => path?.ToLower().StartsWith("http") == true;
 
+    private static string SanitizeCaptionText(string text)
+    {
+        if (string.IsNullOrWhiteSpace(text))
+            return null;
+
+        var builder = new StringBuilder(text.Length);
+        foreach (var ch in text)
+        {
+            builder.Append(char.IsControl(ch) ? ' ' : ch);
+        }
+
+        var cleaned = builder.ToString();
+        return string.Join(" ", cleaned.Split(' ', StringSplitOptions.RemoveEmptyEntries));
+    }
+
     /// <summary>
     /// Returns a specific embedded resource by name.
     /// </summary>

lib/Markdig.Renderer.Docx/TripleColonExtensions/TripleColonProcessor.cs

@@ -1,4 +1,4 @@
-using System.Drawing;
+using System.Drawing;
 using IContainer = DXPlus.IContainer;
 
 namespace Markdig.Renderer.Docx.TripleColonExtensions;
@@ -79,8 +79,10 @@ private static void HandleImage(IDocxRenderer owner, IContainer document, Paragr
         if (extension.Container?.Count > 0 && type == "complex")
         {
             // Should be strictly text as this is for screen readers.
-            description = string.Join("\r\n", extension.Container.Select(b => (b as ParagraphBlock)?.Inline)
-                .SelectMany(ic => ic.Select(il => il.ToString())));
+            description = string.Join("\r\n", extension.Container
+                .Select(b => (b as ParagraphBlock)?.Inline)
+                .Where(inline => inline != null)
+                .SelectMany(inline => inline.Select(il => il.ToString())));
         }
 
         var drawing = owner.InsertImage(currentParagraph, extension.Container ?? (MarkdownObject)extension.Inlines, source, title, description, hasBorder?.ToLower()=="true");

src/ConvertDocx/Program.cs

@@ -1,4 +1,4 @@
-using CommandLine;
+using CommandLine;
 using Julmar.DocsToMarkdown;
 using LearnDocUtils;
 using MSLearnRepos;
@@ -146,7 +146,12 @@ private static async Task DownloadMarkdown(CommandLineOptions options)
             var downloader = new DocsConverter(tempFolder, new Uri(options.InputFile));
             var createdFiles = await downloader.ConvertAsync(!options.PreferPlainMarkdown, 
 #if DEBUG
-                tag => Console.Error.WriteLine($"Skipped: {tag.TrimStart().Substring(0, 20)}"));
+                tag =>
+                {
+                    var trimmed = tag.TrimStart();
+                    var prefix = trimmed.Length > 20 ? trimmed[..20] : trimmed;
+                    Console.Error.WriteLine($"Skipped: {prefix}");
+                });
 #else
                 null);
 #endif
@@ -245,7 +250,12 @@ private static async Task<List<string>> DownloadAndConvertAsync(CommandLineOptio
             var downloader = new DocsConverter(tempFolder, new Uri(options.InputFile));
             var createdFiles = await downloader.ConvertAsync(!options.PreferPlainMarkdown, 
 #if DEBUG
-                tag => Console.Error.WriteLine($"Skipped: {tag.TrimStart().Substring(0, 20)}"));
+                tag =>
+                {
+                    var trimmed = tag.TrimStart();
+                    var prefix = trimmed.Length > 20 ? trimmed[..20] : trimmed;
+                    Console.Error.WriteLine($"Skipped: {prefix}");
+                });
 #else
                 null);
 #endif

src/ConvertDocx/Properties/launchSettings.json

@@ -2,7 +2,7 @@
   "profiles": {
     "ConvertDocToDocx": {
       "commandName": "Project",
-      "commandLineArgs": "-d -f Docx \"https://learn.microsoft.com/en-us/azure/dev-box/concept-what-are-dev-box-customizations?tabs=team-customizations/\" \"/users/mark/Desktop/test.docx\""
+      "commandLineArgs": "-f Docx \"https://learn.microsoft.com/training/modules/accessibility-build-foundation-inclusive-learning/\" \"/users/mark/Desktop/test.docx\""
     }
   }
 }