private void ParseHtmlChildren(HtmlFormatPlainText format, HtmlNode node, StringBuilder sb)
{
foreach (HtmlNode child in node.ChildNodes)
{
ParseHtml(format, child, sb);
}
}
private string ConvertHtmlToPlainText(HtmlFormatPlainText format, string text)
{
StringBuilder sb = new StringBuilder();
HtmlDocument html = new HtmlDocument();
html.LoadHtml(text);
ParseHtml(format, html.DocumentNode, sb);
return(TrimLines(sb.ToString()));
}
private void ParseHtml(HtmlFormatPlainText format, HtmlNode node, StringBuilder sb)
{
string html;
switch (node.NodeType)
{
case HtmlNodeType.Comment:
// Don't output comments
break;
case HtmlNodeType.Document:
ParseHtmlChildren(format, node, sb);
break;
case HtmlNodeType.Text:
string parentName = node.ParentNode.Name;
if (parentName.Equals("script") || parentName.Equals("style"))
{
// Ignore scripts and styles
break;
}
html = ((HtmlTextNode)node).Text;
if (HtmlNode.IsOverlappedClosingElement(html))
{
// Is it in fact a special closing node output as text?
break;
}
sb.Append(HtmlEntity.DeEntitize(html));
break;
case HtmlNodeType.Element:
HtmlFormatPlainText nextFormat = format;
switch (node.Name)
{
case "p":
sb.Append(Environment.NewLine);
break;
case "ol":
nextFormat = new HtmlFormatPlainText(HtmlFormatMode.OrderedList)
{
Indent = format.Indent + 1
};
break;
case "ul":
nextFormat = new HtmlFormatPlainText(HtmlFormatMode.UnorderedList)
{
Indent = format.Indent + 1
};
break;
case "li":
nextFormat.Counter += 1;
sb.Append(' ', nextFormat.Indent);
switch (nextFormat.Mode)
{
case HtmlFormatMode.OrderedList:
sb.Append(string.Format("{0}. ", nextFormat.Counter));
break;
case HtmlFormatMode.UnorderedList:
sb.Append("* ");
break;
}
break;
}
if (node.HasChildNodes)
{
ParseHtmlChildren(nextFormat, node, sb);
}
break;
}
}