From 07fae6da38ff655044896a9358f384d23f8e789b Mon Sep 17 00:00:00 2001
From: Olivier Nizet <onizet@hotmail.com>
Date: Mon, 11 Nov 2024 15:32:23 +0100
Subject: [PATCH] Extend support of nested list for non-W3C compliant html #173

---
 CHANGELOG.md                                  |  1 +
 .../Expressions/Numbering/ListExpression.cs   | 13 +++++-
 test/HtmlToOpenXml.Tests/NumberingTests.cs    | 42 ++++++++++++++++++-
 3 files changed, 54 insertions(+), 2 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index b4edec4..a6003e6 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,7 @@
 - Support deprecrated align attribute for block #171
 - Fix parsing of style attribute with a key with no value
 - Improve parsing of style attribute to avoid an extra call to HtmlDecode
+- Extend support of nested list for non-W3C compliant html #173
 
 ## 3.2.1
 
diff --git a/src/Html2OpenXml/Expressions/Numbering/ListExpression.cs b/src/Html2OpenXml/Expressions/Numbering/ListExpression.cs
index 811976c..ef7fa22 100644
--- a/src/Html2OpenXml/Expressions/Numbering/ListExpression.cs
+++ b/src/Html2OpenXml/Expressions/Numbering/ListExpression.cs
@@ -50,9 +50,20 @@ readonly struct ListContext(string listName, int absNumId, int instanceId, int l
 
     public override IEnumerable<OpenXmlElement> Interpret(ParsingContext context)
     {
-        var liNodes = node.Children.Where(n => n.LocalName == "li");
+        var liNodes = node.Children.Where(n => n.LocalName.Equals("li", StringComparison.OrdinalIgnoreCase));
         if (!liNodes.Any()) yield break;
 
+        // W3C requires that nested list stands below a `li` element but some editors
+        // don't care to respect the standard. Let's reparent those lists
+        var nestedList = node.Children.Where(n => 
+            n.LocalName.Equals("ol", StringComparison.OrdinalIgnoreCase) || 
+            n.LocalName.Equals("ul", StringComparison.OrdinalIgnoreCase));
+        if (nestedList.Any())
+        {
+            foreach (var list in nestedList)
+                list.PreviousElementSibling?.AppendChild(list);
+        }
+
         var listContext = context.Properties<ListContext>("listContext");
         var parentContext = listContext;
         var listStyle = GetListName(node, listContext.Name);
diff --git a/test/HtmlToOpenXml.Tests/NumberingTests.cs b/test/HtmlToOpenXml.Tests/NumberingTests.cs
index a7b914e..17b75de 100644
--- a/test/HtmlToOpenXml.Tests/NumberingTests.cs
+++ b/test/HtmlToOpenXml.Tests/NumberingTests.cs
@@ -514,7 +514,7 @@ public void WithRtl_ReturnsBidi(string dir, bool? expectedValue)
         }
 
         [Test]
-        public void NestedNumberList_ReturnsIncrementalIdentation()
+        public void NestedNumberList_ReturnsIncrementalIndentation()
         {
             const int maxLevel = 8;
             var sb = new System.Text.StringBuilder();
@@ -541,5 +541,45 @@ public void NestedNumberList_ReturnsIncrementalIdentation()
                 TestContext.Out.WriteLine($"{i}. {ident?.Left?.Value}");
             }
         }
+
+        [Test(Description = "Nested list must be a children of a `li` tag but some editor are not respecting the W3C standard (issue #173)")]
+        public async Task NestedNumberList_NonCompliant_ReturnsIncrementalIndentation()
+        {
+            await converter.ParseBody(@"<ol>
+                <li>Item1</li>
+                <li>Item2</li>
+                    <ol><li>Item 2.1</li></ol>
+                </ol>");
+
+            var absNum = mainPart.NumberingDefinitionsPart?.Numbering
+                .Elements<AbstractNum>()
+                .SingleOrDefault();
+            Assert.That(absNum, Is.Not.Null);
+
+            var inst = mainPart.NumberingDefinitionsPart?.Numbering
+                .Elements<NumberingInstance>().Where(i => i.AbstractNumId?.Val == absNum.AbstractNumberId)
+                .SingleOrDefault();
+            Assert.That(inst, Is.Not.Null);
+            Assert.That(inst.NumberID?.Value, Is.Not.Null);
+
+            var elements = mainPart.Document.Body!.ChildElements;
+            Assert.Multiple(() => {
+                Assert.That(elements, Has.Count.EqualTo(3));
+                Assert.That(elements, Is.All.TypeOf<Paragraph>());
+                Assert.That(mainPart.NumberingDefinitionsPart?.Numbering, Is.Not.Null);
+            });
+
+            // assert paragraphs linked to numbering instance
+            Assert.Multiple(() =>
+            {
+                Assert.That(elements.Cast<Paragraph>().Select(e => 
+                    e.ParagraphProperties?.NumberingProperties?.NumberingId?.Val?.Value),
+                    Has.All.EqualTo(inst.NumberID.Value),
+                    "All paragraphs are linked to the same list instance");
+                Assert.That(elements.Take(2).Select(p => p.GetFirstChild<ParagraphProperties>()?.NumberingProperties?.NumberingLevelReference?.Val?.Value), Has.All.EqualTo(0));
+                Assert.That(elements.Last().GetFirstChild<ParagraphProperties>()?.NumberingProperties?.NumberingLevelReference?.Val?.Value, Is.EqualTo(1));
+            });
+            AssertThatOpenXmlDocumentIsValid();
+        }
     }
 }
\ No newline at end of file