[RESOLVED] converting bad urls to links

redrage

here is a sample string

$bob = "Here is a bunch of text.  and now there is some link here www. boblikes cheese.com and cheese is good for you";

and the same for email addresses

$bob = "Here is a bunch of text.  and now there is some mailto here bob@ cheesypoofs.com and cheese is good for you";

i need to remove them spaces in the url and make it a link. cannot seem to get it to work they way i want to. everything i've done just blows up in my face.

madwormer2

I don't have a solution, but I have some pointers that you will need to consider when writing the script.

When you remove these spaces, you'll need to be careful where you remove them from. You need to consider possible false positives. Also, as a human YOU may be able to distinguish a link from some letters, but you know this because of experience.

A script however may see this as a positive: 'We keep documents in www. Then it is displayed at our .co.uk adress, like so:'

As you can see, some technobabble about some websetup could easily be turned into a link.

I'd consider just asking whoever's inputting this data to make nice URLs, instead of bad ones :p

redrage

madwormer2 wrote:
I don't have a solution, but I have some pointers that you will need to consider when writing the script.

When you remove these spaces, you'll need to be careful where you remove them from. You need to consider possible false positives. Also, as a human YOU may be able to distinguish a link from some letters, but you know this because of experience.

A script however may see this as a positive: 'We keep documents in www. Then it is displayed at our .co.uk adress, like so:'

As you can see, some technobabble about some websetup could easily be turned into a link.

I'd consider just asking whoever's inputting this data to make nice URLs, instead of bad ones :p

Exactly why i'm having issues. I've been dicking with it for a day or two.

ya the bad urls thing. cannot fix that because of the software thats used to enter them. it inserts spaces after "."'s and "@"'s to make it wrap to the next line instead of wrapping the whole address. I almost had it working a few times but still had some false positives and some others that just didn't work. The previous person who designed the page had the same issue using perl instead and never found a solution. (or gave up not sure)

I'm so frustrated with it. was hopeing new eyes would see a solution with out to much hassle.

madwormer2

Well, what have you got currently, I'm sure there's ways to improve the accuracy.

sneakyimp

when you talk about the 'software thats used to enter them', that presents an opening. in the general case, you can't really solve your problem reliably. for instance if you were parsing what a zillion monkeys were typing on typewriters.

however, if you are dealing with input that is consistently corrupted in the same way, there is some hope. you just have to reverse the corruption of the software. for instance, it always puts a space after '@'? do str_replace() and remove all those spaces.

You could do the same for periods but that's trickier. you'd have to make sure they ended up being part of some domain. preg_replace() might be helpful. replace all periods preceded by 'www' and followed by a space with just 'www.'
similarly, replace all periods followed by a space and then either edu, com, net, or whatever with just a period then whatever came next.

but, like i said, i don't think you can solve this problem generally. you might be able to make some headway in specific cases though.

redrage

Got it working. After much consideration i decided to dump the idea of making the emailaddresses linkable because robots and such. Also decided to put the link underneith the original string so its not altered (except the at sign). few minor additions and someone else could get the email links working too.

so this is what i ended up with. still needs cleaned up but it works. Spent 6 hours on it yesterday and didn't get jack. spent an hour on it today and got it, darn mondays.

Thanks for the help 🙂

function back_class($cat) {

    $res = mysql_query("QUERY REMOVED");
$cadc = mysql_num_rows($res);
while($row = mysql_fetch_array($res)) {
	$t = trim(ereg_replace(" +", " ",$row[0]));
	$class[] = eregi_replace("@", "[at]",$t);
}

foreach ($class as $str) {
unset($splen);
unset($link);
if (eregi("www.",$str) || eregi("\@",$str)) {
	$sstr = split(" ",$str);

	for ($i = 0; $i <= count($sstr); $i++) {
		if (eregi("www",$sstr[$i]) || eregi("\@",$sstr[$i])) {
			$akey = $i;
		}
	}

	for ($i = 0; $i <= count($sstr); $i++) {
		if(eregi(".com",$sstr[$i]) || eregi(".net",$sstr[$i]) || eregi(".org",$sstr[$i]) || eregi(".us",$sstr[$i]) || eregi(".uk",$sstr[$i]) || eregi(".edu",$sstr[$i])) {
			$bkey = $i;
		}
	}
	$splen = $bkey - $akey;
	$splen++;
	$lnkarry = array_slice($sstr,$akey,$splen);
	$url=implode($lnkarry);
	$link = "<br><a href=http://$url>$url</a>";
}
$piss[] = (isset($link)) ? $str . $link : $str;
 } // end for
return $piss;
}