Handling html in strings pasted from Text Editor (Word)

wmcelwee

I am working on a site with a CMS back end. The people writing the articles are typically using Word and then copy/pasting into (I think) FCK editor (maybe Tiny MCE) on the back end. The issue is that many times, depending on what they did in word, I am getting a lot of extra, unwanted or unnecessary HTML tags in the article.

For instance, and article may start like this and/or contain similar strings within the field
<p class="ecxMsoNormal" style="text-indent: 13.5pt"> Where is...

or this
<div style="margin: 0in 0in 0pt">

I want to remove mostly the attributes of the tags, but I want to leave the attributes on a tag such as 'a' where the 'href' is important. Of course I also need to keep the formatting tags like 'p' and 'div' (maybe convert div to p).

I've looked for some scripts to strip tags/attributes but nothing I have come across has worked well or at all. I'd really like to be able to clean these articles better.

Any tips?

halojoy

What you describe
is just how the php [man]strip_tags/man function is working
1. function strips all HTML <tags>
2. except for those tags <p> <div> you tell you want to keep

The second example below is:

$filtered_text = strip_tags( $text, '<p><a><div>' );

In this case we keep all paragraphs <p> , links <a href> and all <div> that is found in $text

These are the examples from PHP Manual

<?php

$text = '<p>Test paragraph.</p><!-- Comment --> <a href="#fragment">Other text</a>';

// strip all tags
echo strip_tags($text);
echo "\n";


// Allow <p> and <a>
echo strip_tags($text, '<p><a>');


?>

wmcelwee

Thanks for the reply, however, what I am looking to do is remove attributes from the tags (except href from 'a' tags). I assume this is done through a regular expression but none that I found seemed to do the job correctly. I didn't know if there was another method I could use or help with a regular expression.

Weedpacket

Depending on how sane the Word-generated markup (and any change by the editor) is you may be able to use [man]DOM[/man] to read and parse the content. Then you can traverse the parsed document and strip out any unwanted parts.

Try both HTML and XML for both input and output (XML is much stricter about input; both forms of output may add additional cruft to yield a valid document).