MS-Word-to-HTML document cleaner

This is a very rudimentary script for stripping gack from MS Word documents that have been converted to HTML.

I used preg_replace(), and a series of patterns to remove excess styling from the document.

It’s not perfect by any means, and I make no promises or warranties that it will work for you. This code (which starts and stops between the <pre> and </pre> tags) is however, is free for you to use in any way you wish.

<?

$input=file_get_contents('prog.html');

# i at the end of the pattern indicates a case-insensitive search.

$find=array(
    '/mso-[a-z]{1,10}:/',
    '/mso-[a-z]{1,10}-[a-z]{1,10}:/',
    '/mso-[a-z]{1,10}-[a-z]{1,10}-[a-z]{1,10}:/',
    '/<!\[if !supportEmptyParas\]>/',
    '/<!\[endif\]>/',
    '/<!\[if !supportLists\]>/',
    '/l3/',
    '/level1/',
    '/lfo5/',
    '/Mso[a-zA-Z]{1,10}/i',
    '/<o:p><\/o:p>/',
    '/border:/',
    '/border-[a-zA-Z]{3,9}:/',
    '/margin:/',
    '/margin-[a-zA-Z]{3,9}:/',
    '/ class=/',
    '/( ){1,3}/',
    '/[0-9]{0,3}\.[0-9]{1,2}in/',
    '/[0-9]{0,3}\.[0-9]{1,2}pt/',
    '/[0-9]{0,3}\.[0-9]{1,2}%/',
    "/[0-9]{1,3}pt/",
    "/[0-9]{1,3}in/",
    "/[0-9]{1,3}%/",
    '/solid windowtext/',
    '/text-indent:/',
    '/tab-stops:[a-zA-Z0-9]{0,8}/',
    '/font-[a-zA-Z]{4,7}:/',
    '/font:/',
    '/width:/',
    '/height:/',
    '/padding:/',
    '/padding-[a-zA-Z]{3,9}:/',
    '/color:/',
    '/none;/',
    '/collapse;/',
    '/background:/',
    '/Tahoma/',
    '/"Times New Roman"/',
    "/<span style='.{0,20}'>/",
    '/<span style=".{0,20}">/',
    '/<\/span>/',
    '/<span\s{0,20}style=".{0,10}">/',
    "/<span\s{0,20}style='.{0,10}'>/",
    #removes colors
    '/#[a-zA-Z0-9]{6};/',
    #may remove most style attributes
    "/style='\p{P}{0,10}'/",
    '/style="\p{P}{0,10}"/m',
    #cleaning up stray bits
    "/style='/",
    "/;\s;/",
    "/;\s{0,10}'/",
    "/\s;\s{0,20}'/",
    '/<p><b><u><\/u><\/b><\/p>/'
);

$replace='';

$input = preg_replace($find,$replace,$input);
$input = preg_replace('/\r >/','>',$input);

print $input;

?>