Locating a < symbol in HTML that isn't part of a tag

Question

I'm trying to find a way to reliably locate and replace < and > symbols within an HTML/XML formatted string that do not belong to tags.

Basically I start with an HTML string and convert it into something usable by PDFLib, which uses a form of XML to describe documents to be written as PDF's. However if there is a < within in the content it sees it as the opening of a tag and throws a parse exception.

Example input:

<p>This is a test where 6 < 9</p>
<p>This is part of <strong>The same test</strong></p>
<p>This should also work 6<99999</p>

The text surrounding the < is not always numbers, it is user entered and could be anything such as Grade<C, Blue<Red<Green, Test < Test2.... just about anything really

Required output

This is a test where 6 <charref fontname=Helvetica encoding=unicode>&lt;<resetfont> 9\n
This is part of <fontname=Helvetica fontstyle=bold encoding=unicode>The same test<resetfont>\n
This should also work 6<charref fontname=Helvetica encoding=unicode>&lt;<resetfont>99999\n

I've tried a str_replace and preg_replace, but can't find a solution that will reliably leave the tags alone and replace just the < in context.

Parsing the DOM also seems to fail as the DOMDocument sees the < as an opening tag as well

Using htmlspecialchars on the string converts all the tags <> into <> as well which is no good.

Does anyone have any ideas?

Answer 1

Try using the answer from this question:

how to repair malformed xml

I tried to add this as as it stands, but StackOverflow requires me to add some description to the answer, or it automatically gets converted into a comment, which can't be accepted as an answer.

Answer 2

try reading the string from start char by char if it encounters a < push it in a buffer if > is found without a space then its a tag else if it encounter a < again mark the previous as < and put next in buffer ... and repeat until the end of string

Answer 3

While it's no longer maintained, I think the php port of html5lib is probably your best bet for parsing bad markup.

A simple call like this:

require_once 'your-path-path-to-html5lib/Parser.php';
$dom = HTML5_Parser::parse($input);

will take bad markup in $input and return a valid php DOMDocument.

From there you can save it back to a string with $dom->saveHTML() or $dom->saveXML, or extract the bits you want with the DOM API.

Note that this will produce a full HTML document with head and body etc. even if your original data didn't include that.

If you just want to parse an HTML fragment, you can do:

$dom = HTML5_Parser::parseFragment($input);

which will return a DOMNodeList.

Answer 4

HTML entities are the best way to do such things <> are the entities used to replace <> in HTML. Even using the <code> tag. You can use these entities and replace them with <> in your HTML Tags. www.w3schools.com/html/html_entities.asp