#18 Namespace support in XPath.

Author: Denis Ivanov

src/AngleSharp.XPath.Tests/HtmlDocumentNavigatorTests.cs

@@ -65,7 +65,23 @@ public void SelectPrecedingNodeInDocumentWithDoctype_ShouldReturnNode()
         }
 
         [Test]
-        public void SelectNodeWithNamespace_ShouldReturnNode()
+        public void SelectSingleNode_IgnoreNamespaces_ShouldReturnNode()
+        {
+            // Arrange
+            var xml = "<?xml version=\"1.0\" encoding=\"utf-8\" ?><urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\" xmlns:xhtml=\"http://www.w3.org/1999/xhtml\"><url><loc>https://www.test.com/de/accounts/profile</loc><xhtml:link rel=\"alternate\" hreflang=\"fr\" href=\"https://www.test.com/fr/accounts/profile\" /><xhtml:link rel=\"alternate\" hreflang=\"en\" href=\"https://www.test.com/en/accounts/profile\" /><xhtml:link rel=\"alternate\" hreflang=\"it\" href=\"https://www.test.com/it/accounts/profile\" /><changefreq>weekly</changefreq><priority>0.4</priority></url></urlset>";
+            var parser = new XmlParser();
+            var doc = parser.ParseDocument(xml);
+
+            // Act
+            var node = doc.DocumentElement.SelectSingleNode("/urlset/url/link");
+
+            // Assert
+            Assert.IsNotNull(node);
+            Assert.That(node.NodeName, Is.EqualTo("xhtml:link"));
+        }
+
+        [Test]
+        public void SelectSingleNode_DontIgnoreNamespaces_ShouldReturnNode()
         {
             // Arrange
             var xml = "<?xml version=\"1.0\" encoding=\"utf-8\" ?><urlset xmlns=\"http://www.sitemaps.org/schemas/sitemap/0.9\" xmlns:xhtml=\"http://www.w3.org/1999/xhtml\"><url><loc>https://www.test.com/de/accounts/profile</loc><xhtml:link rel=\"alternate\" hreflang=\"fr\" href=\"https://www.test.com/fr/accounts/profile\" /><xhtml:link rel=\"alternate\" hreflang=\"en\" href=\"https://www.test.com/en/accounts/profile\" /><xhtml:link rel=\"alternate\" hreflang=\"it\" href=\"https://www.test.com/it/accounts/profile\" /><changefreq>weekly</changefreq><priority>0.4</priority></url></urlset>";
@@ -77,7 +93,7 @@ public void SelectNodeWithNamespace_ShouldReturnNode()
             namespaceManager.AddNamespace("d", "http://www.sitemaps.org/schemas/sitemap/0.9");
 
             // Act
-            var node = doc.DocumentElement.SelectSingleNode("/urlset/url/link");
+            var node = doc.DocumentElement.SelectSingleNode("/d:urlset/d:url/xhtml:link", namespaceManager, false);
 
             // Assert
             Assert.IsNotNull(node);

src/AngleSharp.XPath/Extensions.cs

@@ -16,11 +16,12 @@ public static class Extensions
         /// Creates a new navigator for the given document.
         /// </summary>
         /// <param name="document">The document to extend.</param>
+        /// <param name="ignoreNamespaces"></param>
         /// <returns>The navigator for XPath expressions.</returns>
-        public static XPathNavigator CreateNavigator(this IDocument document)
+        public static XPathNavigator CreateNavigator(this IDocument document, bool ignoreNamespaces = true)
         {
             var doc = document ?? throw new ArgumentNullException(nameof(document));
-            return new HtmlDocumentNavigator(doc, doc.DocumentElement);
+            return new HtmlDocumentNavigator(doc, doc.DocumentElement, ignoreNamespaces);
         }
 
         [DebuggerStepThrough]
@@ -35,14 +36,29 @@ internal static string GetOrAdd(this XmlNameTable table, string array)
         /// </summary>
         /// <param name="element">The element to start looking from.</param>
         /// <param name="xpath">The XPath expression.</param>
+        /// <param name="ignoreNamespaces"></param>
         /// <returns>The node matching <paramref name="xpath"/> query, if any.</returns>
         /// <exception cref="ArgumentNullException">Throws if <paramref name="element"/> or <paramref name="xpath"/> is <c>null</c></exception>
-        public static INode SelectSingleNode(this IElement element, string xpath)
+        public static INode SelectSingleNode(this IElement element, string xpath, bool ignoreNamespaces = true)
+        {
+            return element.SelectSingleNode(xpath, new XmlNamespaceManager(new NameTable()), ignoreNamespaces);
+        }
+
+        /// <summary>
+        /// Selects a single node (or returns null) matching the <see cref="XPath"/> expression.
+        /// </summary>
+        /// <param name="element">The element to start looking from.</param>
+        /// <param name="xpath">The XPath expression.</param>
+        /// <param name="resolver"></param>
+        /// <param name="ignoreNamespaces"></param>
+        /// <returns>The node matching <paramref name="xpath"/> query, if any.</returns>
+        /// <exception cref="ArgumentNullException">Throws if <paramref name="element"/> or <paramref name="xpath"/> is <c>null</c></exception>
+        public static INode SelectSingleNode(this IElement element, string xpath, IXmlNamespaceResolver resolver, bool ignoreNamespaces = true)
         {
             var el = element ?? throw new ArgumentNullException(nameof(element));
             var xp = xpath ?? throw new ArgumentNullException(nameof(xpath));
-            var nav = new HtmlDocumentNavigator(el.Owner, el);
-            var it = nav.SelectSingleNode(xp);
+            var nav = new HtmlDocumentNavigator(el.Owner, el, ignoreNamespaces);
+            var it = nav.SelectSingleNode(xp, resolver ?? new XmlNamespaceManager(new NameTable()));
             return ((HtmlDocumentNavigator) it)?.CurrentNode;
         }
 
@@ -51,14 +67,29 @@ public static INode SelectSingleNode(this IElement element, string xpath)
         /// </summary>
         /// <param name="element">The element to start looking from.</param>
         /// <param name="xpath">The XPath expression.</param>
+        /// <param name="ignoreNamespaces"></param>
+        /// <returns>List of nodes matching <paramref name="xpath"/> query.</returns>
+        /// <exception cref="ArgumentNullException">Throws if <paramref name="element"/> or <paramref name="xpath"/> is <c>null</c></exception>
+        public static List<INode> SelectNodes(this IElement element, string xpath, bool ignoreNamespaces = true)
+        {
+            return element.SelectNodes(xpath, new XmlNamespaceManager(new NameTable()), ignoreNamespaces);
+        }
+
+        /// <summary>
+        /// Selects a list of nodes matching the <see cref="XPath"/> expression.
+        /// </summary>
+        /// <param name="element">The element to start looking from.</param>
+        /// <param name="xpath">The XPath expression.</param>
+        /// <param name="resolver"></param>
+        /// <param name="ignoreNamespaces"></param>
         /// <returns>List of nodes matching <paramref name="xpath"/> query.</returns>
         /// <exception cref="ArgumentNullException">Throws if <paramref name="element"/> or <paramref name="xpath"/> is <c>null</c></exception>
-        public static List<INode> SelectNodes(this IElement element, string xpath)
+        public static List<INode> SelectNodes(this IElement element, string xpath, IXmlNamespaceResolver resolver, bool ignoreNamespaces = true)
         {
             var el = element ?? throw new ArgumentNullException(nameof(element));
             var xp = xpath ?? throw new ArgumentNullException(nameof(xpath));
-            var nav = new HtmlDocumentNavigator(el.Owner, el);
-            var it = nav.Select(xp);
+            var nav = new HtmlDocumentNavigator(el.Owner, el, ignoreNamespaces);
+            var it = nav.Select(xp, resolver ?? new XmlNamespaceManager(new NameTable()));
             var result = new List<INode>();
 
             while (it.MoveNext())

src/AngleSharp.XPath/HtmlDocumentNavigator.cs

@@ -11,18 +11,21 @@ public class HtmlDocumentNavigator : XPathNavigator
 		private readonly IDocument _document;
         private INode _currentNode;
 		private int _attrIndex;
+        private readonly bool _ignoreNamespaces;
 
         /// <summary>
         /// Creates a new XPath navigator for the given document using the provided root node.
         /// </summary>
         /// <param name="document">The document to navigate.</param>
         /// <param name="currentNode">The node to start navigation.</param>
-        public HtmlDocumentNavigator(IDocument document, INode currentNode)
+        /// <param name="ignoreNamespaces"></param>
+        public HtmlDocumentNavigator(IDocument document, INode currentNode, bool ignoreNamespaces)
 		{
 			_document = document ?? throw new ArgumentNullException(nameof(document));
             NameTable = new NameTable();
             _currentNode = currentNode ?? throw new ArgumentNullException(nameof(currentNode));
             _attrIndex = -1;
+            _ignoreNamespaces = ignoreNamespaces;
         }
 
         /// <inheritdoc />
@@ -57,10 +60,20 @@ public HtmlDocumentNavigator(IDocument document, INode currentNode)
                 : NameTable.GetOrAdd(_currentNode.NodeName);
 
         /// <inheritdoc />
-        public override string NamespaceURI => string.Empty
-            /*_attrIndex != -1
-                ? NameTable.GetOrAdd(CurrentElement.Attributes[_attrIndex].NamespaceUri ?? string.Empty)
-                : NameTable.GetOrAdd(CurrentElement?.NamespaceUri ?? string.Empty)*/;
+        public override string NamespaceURI
+        {
+            get
+            {
+                if (_ignoreNamespaces)
+                {
+                    return string.Empty;
+                }
+
+                return _attrIndex != -1
+                    ? NameTable.GetOrAdd(CurrentElement.Attributes[_attrIndex].NamespaceUri ?? string.Empty)
+                    : NameTable.GetOrAdd(CurrentElement?.NamespaceUri ?? string.Empty);
+            }
+        }
 
         /// <inheritdoc />
         public override string Prefix =>
@@ -169,7 +182,7 @@ public override string Value
         /// <inheritdoc />
         public override XPathNavigator Clone()
 		{
-			return new HtmlDocumentNavigator(_document, _currentNode);
+			return new HtmlDocumentNavigator(_document, _currentNode, _ignoreNamespaces);
 		}
 
         /// <inheritdoc />