Using C# to convert incorrect html string to real html

Question

My original issue is that I am trying to serialize a string containing html tags to an XML element.

hello <a href="world.php">World</a>, this

is
a nice
test

<ul>
  <li>to demonstrate my issue</li>
  <li>and find a solution</li>
</ul>

However, I have 2 issues

1. Serializing HTML to XML: I did not succeed in defining the Serializable class to correctly serialize with XmlSerialze, so I decided that, using CDATA sections might be the better way. This is however not correctly deserialized by the target tool (that I have no influence on). What I need is plain and correct html (XHMTL?) within the xml output file.

---

2. The string looks e.g. as above, but is not fully correct html (no <p> tags, no <br> tags). Now I would like to replace the newlines by a p or br tag. I have had a look here and used the suggested solution:

    string result = "<p>" + text
     .Replace(Environment.NewLine + Environment.NewLine, "</p><p>")
     .Replace(Environment.NewLine, "<br />")
     .Replace("</p><p>", "</p>" + Environment.NewLine + "<p>") + "</p>";

However, this does not in all cases generate valid html. In the example above, it would create <br />s between the <li> tags or cause <ul> tags within <p> tags - which is both not allowed.

Target would be to have a result like the following (line breaks are only for better readability and don't matter here)

<p>hello <a href="world.php">World</a>, this</p>
<p>is<br/>
a nice<br/>
test<br/></p>
<ul>
  <li>to demonstrate my issue</li>
  <li>and find a solution</li>
</ul>

Do you have any suggestion how to solve this either with a string.Replace, Regex, or better solution (HtmlDocument)?

Please note: I have no influence on deserialization, the XML output is evaluated by I tool I have no influence on, and it has to be UTF-8 encoded.

Thank you!

EDIT: Clearly separated the 2 issues

EDIT2: No influence on deserialization

EDIT3: Added target output

Answer 1

What you're trying to do is implement a "tag soup parser", which takes text that may or may not be HTML as input and transforms that into a valid DOM, that a HTML parser can handle.

You don't want to reinvent this wheel, most definitely not with simple string replaces. See How to parse bad html? for some hints.

---

Or you can just encode the input HTML in such a way that it doesn't interfere with the XML that you're trying to put it in, like a CDATA section or base64-encoding the input would also suffice. Don't use "entity encoding", as your XML parser is going to complain about HTML entities that aren't XML entities.

Answer 2

I've had to do similar (ensuring 3rd party content has valid HTML). If I was doing this, I'd do the following:

1) Replace line breaks with HTML line breaks

string result = text.Replace(Environment.NewLine, "<br />");

2) Use the HTMLAgility pack to fix any invalid HTML

    var doc = new HtmlDocument();
    HtmlNode.ElementsFlags["p"] = HtmlElementFlag.Closed;
    doc.OptionFixNestedTags = false;
    doc.LoadHtml(result);

    if (doc.ParseErrors.Count() > 0)
    {
                // throw error
    }else{
                // get fixed html
                 result= doc.DocumentNode.OuterHtml;
    }