C# FlowDocument to HTML conversion

The general technique is to use a XamlWriter to convert the FlowDocument content to a stream of XML, and then to use an XSLT transform to convert the XML to HTML. That's not much of an answer, but that's because there's a huge range of possible HTML representations of any given FlowDocument.

This transform, for instance, converts every top-level Section to a div, every Paragraph to a p, and every Run to a span whose class tells you whether or not it's italicized, bold-faced, or underlined, or any combination of the above. It was useful for the purpose I wrote it for, but to call it a lossy transformation is an understatement:

<?xml version="1.0" encoding="utf-8"?>
<xsl:stylesheet
    version="1.0"
    xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
    xmlns:x="http://schemas.microsoft.com/winfx/2006/xaml/presentation"
    xmlns:msxsl="urn:schemas-microsoft-com:xslt"
    exclude-result-prefixes="msxsl x">

  <xsl:output method="xml" indent="yes" omit-xml-declaration="yes"/>

  <xsl:template match="x:Section[not(parent::x:Section)]">
    <div>
      <xsl:apply-templates select="node()"/>
    </div>
  </xsl:template>

  <xsl:template match="x:Section">
    <xsl:apply-templates select="node()"/>
  </xsl:template>

  <xsl:template match="x:Paragraph">
    <p>
      <xsl:apply-templates select="node()"/>
    </p>
  </xsl:template>

  <xsl:template match="x:Run">
    <xsl:variable name="class">
      <xsl:if test="@FontStyle='Italic'">
        <xsl:text>i </xsl:text>
      </xsl:if>
      <xsl:if test="@FontWeight='Bold'">
        <xsl:text>b </xsl:text>
      </xsl:if>
      <xsl:if test="contains(@TextDecorations, 'Underline')">
        <xsl:text>u </xsl:text>
      </xsl:if>
    </xsl:variable>
    <span>
      <xsl:if test="normalize-space($class) != ''">
        <xsl:attribute name="class">
          <xsl:value-of select="normalize-space($class)"/>
        </xsl:attribute>
      </xsl:if>
      <xsl:value-of select="text()"/>
    </span>
  </xsl:template>

</xsl:stylesheet>

Here's a value converter I wrote to do the conversion - note that in order to use the value converter, you also have to hack around and implement a version of RichTextBox that exposes the content as a dependency property. Really this whole project was a pain.

public class FlowDocumentToHtmlConverter : IValueConverter
{
    private static XslCompiledTransform ToHtmlTransform;
    private static XslCompiledTransform ToXamlTransform;

    public FlowDocumentToHtmlConverter()
    {
        if (ToHtmlTransform == null)
        {
            ToHtmlTransform = LoadTransformResource("Converters/FlowDocumentToXhtml.xslt");
        }
        if (ToXamlTransform == null)
        {
            ToXamlTransform = LoadTransformResource("Converters/XhtmlToFlowDocument.xslt");
        }
    }
    private static XslCompiledTransform LoadTransformResource(string path)
    {
        Uri uri = new Uri(path, UriKind.Relative);
        XmlReader xr = XmlReader.Create(Application.GetResourceStream(uri).Stream);
        XslCompiledTransform xslt = new XslCompiledTransform();
        xslt.Load(xr);
        return xslt;
    }

    #region IValueConverter Members

    public object ConvertBack(object value, Type targetType, object parameter, System.Globalization.CultureInfo culture)
    {
        if (!(value is FlowDocument))
        {
            return null;
        }
        if (targetType == typeof(FlowDocument))
        {
            return value;
        }

        if (targetType != typeof(string))
        {
            throw new InvalidOperationException(
                "FlowDocumentToHtmlConverter can only convert back from a FlowDocument to a string.");
        }

        FlowDocument d = (FlowDocument)value;

        using (MemoryStream ms = new MemoryStream())
        {
            // write XAML out to a MemoryStream
            TextRange tr = new TextRange(
                d.ContentStart,
                d.ContentEnd);
            tr.Save(ms, DataFormats.Xaml);
            ms.Seek(0, SeekOrigin.Begin);

            // transform the contents of the MemoryStream to HTML
            StringBuilder sb = new StringBuilder();
            using (StringWriter sw = new StringWriter(sb))
            {
                XmlWriterSettings xws = new XmlWriterSettings();
                xws.OmitXmlDeclaration = true;
                XmlReader xr = XmlReader.Create(ms);
                XmlWriter xw = XmlWriter.Create(sw, xws);
                ToHtmlTransform.Transform(xr, xw);
            }
            return sb.ToString();
        }
    }

    public object Convert(object value, Type targetType, object parameter, System.Globalization.CultureInfo culture)
    {
        if (value == null)
        {
            return new FlowDocument();
        }
        if (value is FlowDocument)
        {
            return value;
        }
        if (targetType != typeof(FlowDocument))
        {
            throw new InvalidOperationException(
                "FlowDocumentToHtmlConverter can only convert to a FlowDocument.");
        }
        if (!(value is string))
        {
            throw new InvalidOperationException(
                "FlowDocumentToHtmlConverter can only convert from a string or FlowDocument.");
        }

        string s = (string)value;

        FlowDocument d;

        using (MemoryStream ms = new MemoryStream())
        using (StringReader sr = new StringReader(s))
        {
            XmlWriterSettings xws = new XmlWriterSettings();
            xws.OmitXmlDeclaration = true;
            using (XmlReader xr = XmlReader.Create(sr))
            using (XmlWriter xw = XmlWriter.Create(ms, xws))
            {
                ToXamlTransform.Transform(xr, xw);
            }
            ms.Seek(0, SeekOrigin.Begin);

            d = XamlReader.Load(ms) as FlowDocument;
        }
        XamlWriter.Save(d, Console.Out);
        return d;
    }

    #endregion
}