- Support new docs page structure with multiple content elements.

- Add support for Tab Groups to the HTML > Markdown conversion.

Author: markjulmar

lib/DocsToMarkdown/Converters/Anchor.cs

@@ -1,5 +1,4 @@
 using System.Diagnostics;
-using System.Text;
 using HtmlAgilityPack;
 
 namespace Julmar.DocsToMarkdown.Converters;
@@ -9,12 +8,18 @@ internal class Anchor() : BaseConverter("a")
     public override string Convert(HtmlConverter converter, HtmlNode htmlInput)
     {
         Debug.Assert(CanConvert(htmlInput));
-        // TODO -- look to data tags
-        
+
         var url = htmlInput.GetAttributeValue("href", "");
         if (string.IsNullOrEmpty(url)) // emit the tag directly if no URL.
             return htmlInput.OuterHtml;
 
+        var role = htmlInput.GetAttributeValue("role", "");
+        if (role == "tab")
+        {
+            Debug.Assert(false, "Tab role should be handled in TabConverter");
+            return string.Empty;
+        }
+        
         // Handle lightbox images.
         if (url.EndsWith("#lightbox", StringComparison.InvariantCultureIgnoreCase)
             && htmlInput.ChildNodes.Any(cn => cn.Name.Equals("img", StringComparison.CurrentCultureIgnoreCase)))

lib/DocsToMarkdown/Converters/DivSpan.cs

@@ -18,8 +18,55 @@ internal class DivSpan() : BaseConverter("div", "span")
         { "has-pivot", ZonePivotSection},
         { "xp-tag-hexagon", IgnoreBlock },
         { "embeddedvideo", EmbeddedVideo },
+        { "tabGroup", TabGroup }
     };
 
+    private static string TabGroup(string className, HtmlConverter converter, HtmlNode htmlNode)
+    {
+        var list = htmlNode.ChildNodes.FirstOrDefault(n => n.NodeType == HtmlNodeType.Element && n.Name == "ul");
+        if (list == null)
+            return IgnoreBlock(className, converter, htmlNode);
+        
+        var sb = new StringBuilder();
+        foreach (var li in list.ChildNodes.Where(n => n.Name == "li"))
+        {
+            var anchor = li.ChildNodes.FirstOrDefault(n => n.NodeType == HtmlNodeType.Element && n.Name == "a");
+            if (anchor?.Name != "a" || anchor.GetAttributeValue("role","") != "tab" )
+                return IgnoreBlock(className, converter, htmlNode);
+
+            sb.Append(TabPanel(converter, htmlNode, anchor));
+        }
+
+        return sb.AppendLine("---").ToString();
+    }
+
+    private static string TabPanel(HtmlConverter converter, HtmlNode parentNode, HtmlNode anchor)
+    {
+        var data = anchor.GetAttributeValue("data-tab", "");
+        if (string.IsNullOrEmpty(data))
+            return string.Empty;
+
+        var section = parentNode.ChildNodes.FirstOrDefault(n =>
+            n.Name == "section" &&
+            n.GetAttributeValue("role", "").Equals("tabpanel", StringComparison.OrdinalIgnoreCase) &&
+            n.GetAttributeValue("data-tab", "") == data);
+        if (section == null)
+            return anchor.OuterHtml;
+        
+        var id = "#tab/" + data;
+
+        var sb = new StringBuilder();
+        var anchorText = System.Net.WebUtility.HtmlDecode(anchor.InnerText);
+        if (string.IsNullOrWhiteSpace(anchorText))
+            anchorText = id;
+
+        sb.AppendLine($"# [{anchorText}]({id})")
+            .AppendLine()
+            .Append(converter.ConvertChildren(section));
+        
+        return sb.ToString();
+    }
+    
     private static string EmbeddedVideo(string className, HtmlConverter converter, HtmlNode htmlNode)
     {
         var frame = htmlNode.ChildNodes.FirstOrDefault(cn => cn.Name == "iframe");

lib/DocsToMarkdown/HtmlConverter.cs

@@ -59,12 +59,12 @@ public IEnumerable<string> GetMarkdown()
         if (Metadata.PageKind == "module")
             yield break;
 
-        var root = htmlDocument.DocumentNode.SelectSingleNode("//div[@class='content']") 
-                   ?? htmlDocument.DocumentNode.SelectSingleNode("//div[@class='content ']");
-        if (root == null)
+        var roots = htmlDocument.DocumentNode.SelectNodes("//div[@class='content']") 
+                   ?? htmlDocument.DocumentNode.SelectNodes("//div[@class='content ']");
+        if (roots == null)
             throw new ArgumentException("Cannot find content div to parse.");
 
-        foreach (var node in root.ChildNodes)
+        foreach (var node in roots.SelectMany(n => n.ChildNodes))
         {
             var result = Convert(node);
             if (!string.IsNullOrWhiteSpace(result))

lib/Markdig.Renderer.Docx/DocxRenderer.cs

@@ -272,13 +272,13 @@ private string DetermineContentTypeFromUrl(string imageUrl)
             ".bmp" => "image/bmp",
             ".ico" => "image/x-icon",
             ".cur" => "image/x-icon",
-            ".png" => "image/png",
-            ".jpg" => "image/jpeg",
-            ".jpeg" => "image/jpeg",
-            ".gif" => "image/gif",
-            ".svg" => "image/svg+xml",
-            ".tif" => "image/tiff",
-            ".tiff" => "image/tiff",
+            ".png" => ImageContentType.Png,
+            ".jpg" => ImageContentType.Jpg,
+            ".jpeg" => ImageContentType.Jpeg,
+            ".gif" => ImageContentType.Gif,
+            ".svg" => ImageContentType.Svg,
+            ".tif" => ImageContentType.Tiff,
+            ".tiff" => ImageContentType.Tiff,
             _ => null
         };
 

src/ConvertDocx/Properties/launchSettings.json

@@ -2,7 +2,7 @@
   "profiles": {
     "ConvertDocToDocx": {
       "commandName": "Project",
-      "commandLineArgs": "-d -f Docx \"https://learn.microsoft.com/en-us/training/modules/empower-educators-it-leaders-work-together-reference-groups/\" \"/users/mark/Desktop/test.docx\""
+      "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\""
     }
   }
 }