Modify ReverseFormatter

Create ReverseFormatterTest.cs

Core-Tests/Core-Tests.csproj

@@ -90,6 +90,7 @@
     <Compile Include="PagesStorageProviderTests.cs" />
     <Compile Include="Properties\AssemblyInfo.cs" />
     <Compile Include="ProviderLoaderTests.cs" />
+    <Compile Include="ReverseFormatterTests.cs" />
     <Compile Include="SettingsStorageProviderTests.cs" />
     <Compile Include="TestSettingsStorageProvider.cs" />
     <Compile Include="UsersStorageProviderTests.cs" />

Core-Tests/ReverseFormatterTests.cs

@@ -0,0 +1,36 @@
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+using NUnit.Framework;
+
+namespace ScrewTurn.Wiki.Tests {
+	
+	[TestFixture]
+	public class ReverseFormatterTests {
+
+		[Test]
+		[TestCase("<b>text</b>", "'''text'''")]
+		[TestCase("<strong>text</strong>", "'''text'''")]
+		[TestCase("<i>text</i>", "''text''")]
+		[TestCase("<em>text</em>", "''text''")]
+		[TestCase("<u>text</u>", "__text__")]
+		[TestCase("<s>text</s>", "--text--")]
+		[TestCase("<h1>text</h1>", "==text==")]
+		[TestCase("<h2>text</h2>", "===text===")]
+		[TestCase("<h3>text</h3>", "====text====")]
+		[TestCase("<h4>text</s>", "=====text=====")]
+		[TestCase("<sup>text</sup>", "<sup>text</sup>")]
+		[TestCase("<sub>text</sub>", "<sub>text</sub>")]
+		[TestCase("<pre>text</pre>", "(((text)))")]
+		[TestCase("<code>text</code>", "@@text@@")]
+		[TestCase("<div>text</div>", "text{br}")]
+		[TestCase("<ol><li>1</li><li>2</li></ol>","{br}# 1{br}# 2{br}{br}")]
+		[TestCase("<ul><li>1</li><li>2</li></ul>", "{br}* 1{br}* 2{br}{br}")]
+		[TestCase("<a id=\"Init\" />I'm an anchor", "[anchor|#init]I'm an anchor")]
+		[TestCase("<a class=\"internallink\" href=\"#init\" title=\"This recall an anchor\"></a>", "[#init|This recall an anchor]")]
+		public void PlainTest(string input, string output) {
+			Assert.AreEqual(output, ReverseFormatter.ReverseFormat(input));
+		}
+	}
+}

Core/Core.csproj

@@ -61,6 +61,9 @@
     <CodeAnalysisRuleSet>AllRules.ruleset</CodeAnalysisRuleSet>
   </PropertyGroup>
   <ItemGroup>
+    <Reference Include="SgmlReaderDll">
+      <HintPath>..\References\Lib\MindTouch SGML Reader\SgmlReaderDll.dll</HintPath>
+    </Reference>
     <Reference Include="System" />
     <Reference Include="System.Core">
       <RequiredTargetFramework>3.5</RequiredTargetFramework>

Core/ReverseFormatter.cs

@@ -5,6 +5,8 @@ using System.Linq;
 using System.Text;
 using System.Text.RegularExpressions;
 using System.Web;
+using System.Xml;
+using System.IO;
 
 namespace ScrewTurn.Wiki {
 
@@ -73,14 +75,288 @@ namespace ScrewTurn.Wiki {
 		// A=1 - ATitle=2 - AHref=3 - ATarget=4 - ImageAlt=5 - ImageSrc=6 --- Href/Src=http://www.server.com/Blah.ashx/GetFile.aspx...
 		private static readonly Regex ImageInlineRegexIE = new Regex(@"(<a title=\""?(.*?)\"" href=\""(.*?)\""?( target=_blank)?>)?<img alt=\""?(.*?)\""? src=\""(.*?)\"">(</a>)?", RegexOptions.Compiled | RegexOptions.IgnoreCase | RegexOptions.CultureInvariant);
 
+
+		private static List<string> listText= new List<string>();
+		//private static string result = "";
+
+		private static string processImage(XmlNode node) {
+			string result = "";
+			if(node.Attributes.Count != 0) {
+				foreach(XmlAttribute attName in node.Attributes) {
+					if(attName.Name.ToString() == "src") {
+						string[] path = attName.Value.ToString().Split('=');
+					//result += "|" + processChild(node.ChildNodes);
+					result += "{" + "UP(" + path[1].Split('&')[0] + ")}" + path[2];
+					}
+				}
+			}
+			return result;
+		}
+
+		/// <summary>
+		/// Processes the child Image.
+		/// </summary>
+		/// <param name="nodes">The nodes.</param>
+		/// <returns></returns>
+		private static string processChildImage(XmlNodeList nodes) {
+			string image ="";
+			string p ="";
+			string url = "";
+			string result ="";
+			foreach(XmlNode node in nodes) {
+				if(node.Name.ToLowerInvariant() == "img")
+					image += processImage(node);
+				else if(node.Name.ToLowerInvariant() == "p") {
+					p += "|" + processChild(node.ChildNodes) + "|";
+				}
+				else if(node.Name.ToLowerInvariant() == "a") {
+					string link = "";
+					string target = "";
+
+					if(node.Attributes.Count != 0) {
+						XmlAttributeCollection attribute = node.Attributes;
+						foreach(XmlAttribute attName in attribute) {
+							if(attName.Value.ToString() == "_blank")
+								target += "^";
+							if(attName.Name.ToString() == "href")
+								link += attName.Value.ToString();
+						}
+					}
+					image += processImage(node.LastChild);
+					url = "|" + target + link;
+				}
+			}
+			result = p+image+ url;
+			return result;
+		}
+		/// <summary>
+		/// Processes the child.
+		/// </summary>
+		/// <param name="nodes">The nodes.</param>
+		/// <returns></returns>
+		private static string processChild(XmlNodeList nodes) {
+			string result = "";
+			foreach(XmlNode node in nodes) {
+				bool anchor = false;
+				if(node.NodeType == XmlNodeType.Text) {
+					result += node.Value;
+					//string result = "";
+				}
+				else {
+					switch(node.Name.ToLowerInvariant()) {
+						case "b":
+						case "strong":
+							result += ("'''" + processChild(node.ChildNodes) + "'''");
+							break;
+						case "s":
+							result += ("--" + processChild(node.ChildNodes) + "--");
+							break;
+						case "em":
+						case "i":
+							result += ("''" + processChild(node.ChildNodes) + "''");
+							break;
+						case "u":
+							result += ("__" + processChild(node.ChildNodes) + "__");
+							break;
+						//break;
+						case "h1":
+							result += ("==" + processChild(node.ChildNodes) + "==");
+							break;
+						//break;
+						case "h2":
+							result += ("===" + processChild(node.ChildNodes) + "===");
+							break;
+						//break;
+						case "h3":
+							result += ("====" + processChild(node.ChildNodes) + "====");
+							break;
+						//break;
+						case "h4":
+							result += ("=====" + processChild(node.ChildNodes) + "=====");
+							break;
+						case "pre":
+							result += ("(((" + processChild(node.ChildNodes) + ")))");
+							break;
+						case "code":
+							result += ("@@" + processChild(node.ChildNodes) + "@@");
+							break;
+						case "hr":
+						case "hr /":
+							result += ("----" + processChild(node.ChildNodes));
+							break;
+						case "\t":
+							result += (":" + processChild(node.ChildNodes));
+							break;
+						case "éé":
+							result += ("~~~~" + processChild(node.ChildNodes));
+							break;
+						case "span":
+							if(node.Attributes.Count != 0) {
+								XmlAttributeCollection attribute = node.Attributes;
+								foreach(XmlAttribute attName in attribute) {
+									if(attName.Value.ToString() == "italic")
+										result += "''" + processChild(node.ChildNodes) + "''";
+								}
+							}
+							break;
+						case "\n":
+						case "br":
+							result += ("{br}" + processChild(node.ChildNodes));
+							break;
+						case "ol":
+							result += "{br}" + processChild(node.ChildNodes) + "{br}";
+							break;
+						case "ul":
+							result +=  "{br}" + processChild(node.ChildNodes) + "{br}";
+							break;
+						case "table":
+							result +=  processChild(node.ChildNodes);
+							break;
+						case "tbody":
+							result += processChild(node.ChildNodes);
+							break;
+						case "tr":
+							result += processChild(node.ChildNodes);
+							break;
+						case "td":
+							result += processChild(node.ChildNodes);
+							break;
+						case "li":
+							if (node.ParentNode.Name.ToLowerInvariant() == "ol")
+								result += ("# " + processChild(node.ChildNodes) + "{br}");
+							else if (node.ParentNode.Name.ToLowerInvariant() == "ul")
+								result += ("* " + processChild(node.ChildNodes) + "{br}");
+							break;
+						case "sup":
+							result += ("<sup>" + processChild(node.ChildNodes) + "</sup>");
+							break;
+						case "sub":
+							result += ("<sub>" + processChild(node.ChildNodes) + "</sub>");
+							break;
+						case "p":
+							if(node.Attributes.Count != 0) {
+								XmlAttributeCollection attribute = node.Attributes;
+								foreach(XmlAttribute attName in attribute) {
+									if(attName.Value.ToString() == "imagedescription")
+										result += "";
+								}
+							}
+							break;
+						case "div":
+							if(node.Attributes.Count != 0) {
+								XmlAttributeCollection attribute = node.Attributes;
+								foreach(XmlAttribute attName in attribute) {
+									if (attName.Value.ToString() == "box"){
+										result += "(((" + processChild(node.ChildNodes) + "))){br}";
+									}
+									if(attName.Value.ToString() == "imageleft") {
+										result += "[imageleft" + processChildImage(node.ChildNodes) + "]{br}";											
+									}
+									if(attName.Value.ToString() == "imageright")
+										result += "[imageleft" + processChildImage(node.ChildNodes) + "]{br}";
+									if(attName.Value.ToString() == "imageauto")
+										result += "[imageleft" + processChildImage(node.ChildNodes) + "]{br}";
+								}
+							}
+							else
+								result += (processChild(node.ChildNodes) + "{br}");
+							break;
+
+						case "img":
+							if(node.Attributes.Count != 0) {
+								XmlAttributeCollection attribute = node.Attributes;
+								foreach(XmlAttribute attName in attribute) {
+									//if(attName.Name.ToString() == "src") {
+									//	string[] path = attName.Value.ToString().Split('=');
+										//result += "|" + processChild(node.ChildNodes);
+										result += "";
+									//}
+								}
+							}
+							break;
+						case "a":
+							string link="";
+							string target="";
+							string title="";
+							if(node.Attributes.Count != 0) {
+								XmlAttributeCollection attribute = node.Attributes;
+								foreach(XmlAttribute attName in attribute) {
+									if(attName.Name.ToString() != "id".ToLowerInvariant()) {
+										if(attName.Value.ToString() == "_blank")
+											target += "^";
+										if(attName.Name.ToString() == "href")
+											link += attName.Value.ToString();
+										if(attName.Name.ToString() == "title")
+											title += attName.Value.ToString();
+									}
+									else{
+										anchor = true;
+										result += "[anchor|#" + attName.Value.ToString().ToLowerInvariant() + "]" + processChild(node.ChildNodes);
+										break;
+									}
+								}
+								if(!anchor)
+									result += "[" + target + link + "|" + title + "]" + processChild(node.ChildNodes); //"]");
+							}
+						break;
+
+						default:
+							result += (node.OuterXml);
+							break;
+					}
+				}
+			}
+			return result;
+		}
+
+		/// <summary>
+		/// Froms the HTML.
+		/// </summary>
+		/// <param name="reader">The reader.</param>
+		/// <returns></returns>
+		private static XmlDocument FromHTML(TextReader reader) {
+			
+			// setup SgmlReader
+			Sgml.SgmlReader sgmlReader = new Sgml.SgmlReader();
+			sgmlReader.DocType = "HTML";
+			sgmlReader.WhitespaceHandling = WhitespaceHandling.All;
+
+			sgmlReader.CaseFolding = Sgml.CaseFolding.ToLower;
+			sgmlReader.InputStream = reader;
+
+			// create document
+			XmlDocument doc = new XmlDocument();
+			doc.PreserveWhitespace = true;
+			doc.XmlResolver = null;
+			doc.Load(sgmlReader);
+			return doc;
+		}
+
 		/// <summary>
 		/// Reverse formats HTML content into WikiMarkup.
 		/// </summary>
 		/// <param name="html">The input HTML.</param>
 		/// <returns>The corresponding WikiMarkup.</returns>
 		public static string ReverseFormat(string html) {
-			Match match = null;
 
+			StringReader strReader = new StringReader(html);
+			XmlDocument x = FromHTML((TextReader)strReader);
+			string text = processChild(x.FirstChild.ChildNodes);
+			//StringBuilder t = new StringBuilder(html);
+			//result = "";
+			listText.Clear();
+			return text;
+		}
+
+
+		/// <summary>
+		/// Reverse formats HTML content into WikiMarkup.
+		/// </summary>
+		/// <param name="html">The input HTML.</param>
+		/// <returns>The corresponding WikiMarkup.</returns>
+		public static string ReverseFormatOld(string html) {
+
+			Match match = null;
 			StringBuilder buffer = new StringBuilder(html);
 			if(!html.EndsWith("\r\n")) buffer.Append("\r\n");