Regex highlighting of keywords in HTML docs?

Anon

I have a number of HTML fragments in a database, similar to this:

$html = "I went to the shopping center yesterday.";

I have created a search page which allows users to do keyword searches, and I would like to apply bold formatting to matching keywords within the HTML fragments. Here's an example illustrating what I have so far:

$keyword = "shopping";
$html = eregi_replace("($keyword)", "\1", $html);

This works fine. It will put bold tags around the word 'shopping'.

The problem is that this regex also matches keywords that appear inside of HTML tags.

If you change $keyword to "center" in the above example, it will put bold tags inside of the tag. It screws everything up.

Is there some way I can restrict my regex so that it will not bold keyword matches that appear within HTML tags?

I have been struggling with this for hours, and I normally don't like to ask for help... but I really could use some ideas or direction in this case. Any suggestions would be vastly appreciated.

Anon

OK, I actually have it working somewhat now:

$html = "I went to the shopping center yesterday.";
$keyword = "center";

$html = eregi_replace("(<[^{>]+>[^{<>])($keyword)([^<>]<[^>]+)>",}} "\1\2\3",$html);

echo $html;

Outputs:

<PRE>
I went to the shopping center yesterday.

It's still not working quite right though... it will not match more than one occurrence of the same keyword between any given set of HTML tags... for example, if i chage $keyword to 'nt', it will match that string in the word 'center', but not in the word 'went'.

I think I am going to have to alternate between matching HTML end tags and matching keywords... it IS possible to take turns matching expressions within a regex, right?

Any other suggestions?

Anon

there are a couple of ways you can do this; this is probably the simplest.

first, run preg_match_all("/(<[^{>]+>)([^<>]*)/",} $text, $matches)

this will give you an array of tags, and an array of non-tags (some of which will be empty) (note: anything before the first tag will be lost, but your files should start with anyway)

you can then highlight the non in the array, and then glue the whole lot back together.

--
rad