Can this be optimized?

[Michdd]

I created something that searches for links within a chat row, and outputs then as a link. The search method isn't the best, but that doesn't matter much. It also shortens the link if it's a certain width, Eg: http://www.domain.com/something/somethingelse.php?q=something might show as http://www.domai ... mething . It works fine, I'm just wondering if it could be written better since it's a bit..sketchy.

 

$split = explode(' ', $row['chat']);
	for($i = 0;$i < count($split);$i++)
	{
		if(substr($split[$i], 0, 4) == 'http')
		{
			$link_name = $split[$i];
			if(strlen($split[$i]) >= 40)
			{
				$link_name{(floor(strlen($split[$i])/2))} = ' ';
				$link_parts = explode(' ', $link_name);
				$link_parts[0] = substr($link_parts[0], 0, 19);
				$link_parts[1] = substr($link_parts[1], 19);
				$link_name = $link_parts[0] . ' ... ' . $link_parts[1];
			}
			$split[$i] = '<a class="link" href="' . $split[$i] . '" target="_blank">' . $link_name . '</a>';
		}
	}

[laffin]

I wudda opted for preg_replace or a simpler substr system.

 

I went with a preg_replace system and got this.

<?php
$line="this is a sample website http not working while http://www.google.com shud work.
Now a more complicated example http://www.google.com/search?q=preg_replace";

$pr = preg_replace('@(http://([^\s?/]*)[^\s]*)@i','<a class="link" href="\1" target="blank">\2</a>',$line);
header('Content-type: text/plain');
echo "{$pr}\n";

?>

 

but I belive a substr system wud be a lot faster. But this is simple and works nice

 

output was

this is a sample website http not working while <a class="link" href="http://wwww.google.com">www.google.com</a> shud work.

Now a more complicated example <a class="link" href="http://www.google.com/search?q=preg_replace">www.google.com</a>

 

 

[Ken2k7]

I don't know what you're trying to do with $link_name, but if it's greater than 40 characters and you want to concatenate the first 19 characters, the last 19 characters and 3 dots, that will end up being more than 40 characters. I suggest you reduce it down to like 10

 

<?php
$line="this is a sample website http not working while http://www.google.com should work. Now a more complicated example http://www.google.com/search?q=preg_replace";

$parts = explode(' ', $line);
foreach ($parts as $key => $part) {
     if (strpos($part, 'http') === 0 && strlen($part) > 39) {
          $parts[$key] = '<a class="link" href="' . $part . '" target="_blank">' . substr($part, 0, 10) . '...' . substr($part, -4, 4) . '</a>';
     }
}
echo $implode(' ', $parts);

[laffin]

Wrote up some more code

<?php
header('Content-type: text/plain');

$line="this is a sample website http not working while http://www.google.com shud work.
Now a more complicated example http://www.google.com/search?q=preg_replace";

function CutName($var,$len=40)
{
    return strlen($var)>$len?(substr($var,0,$len-3).'...'):$var;
}

function prlinks($matches)
{
    return '<a class="link" href="'. $matches[1] .'" target="_blank">'. CutName($matches[1]) .'</a>';
}

    $pr = preg_replace_callback('@(http://[^\s]*)@i','prlinks',$line);

echo "Preg Match\n";
echo "{$pr}\n";

?>

 

Theres my entry :-)

 

preg match is pretty flexible, it shud be faster or just as fast as a comparable strpos/substr/implode/explode routine...

 

[Ken2k7]

One reason I chose to not use preg_match() was because it got a bit messy as you had it there.

 

laffin: what about https:// URLs? Just picking on your code.

[laffin]

The reason I kept it as different functions, not to make it messy, but to add more granularity to the functions...

Having the CutName as its own function, allows u to use it elsewhere besides the links routine.

 

So I dun see where Messy comes from?

 

I used the fulll schema of http:// so it grabs full urls, instead of of someome saying just http

 

how hard is it to add https thats simple for me

$pr = preg_replace_callback('@(https?://[^\s]*)@i','prlinks',$line);

 

Now how hard is it for the strpos routine u have?

 

complex pattern matching is best for preg, strpos/substr will just add a lot more code. learn the basics of regex, it may take some time, but its well worth it....

 

So u should look at the difference between both codes again, to me the preg is a lot cleaner than the strpos method.

 

[Ken2k7]

If you want to be technical, your regex doesn't do much more than my simple strpos(). So really, where's the "complex" part of this? The only use I see of using regex is if you were to check if a string is of URL format. Your regex doesn't handle that case.

 

And you ended up using substr() too.

[corbin]

Ken, if I had to give it a blind guess, I would guess the preg_replace_callback method is faster than exploding everything.

[Ken2k7]

Here are 5 tests results with this file:

<?php

$str = 'aslkdfjs awe werqwotw witowt kwl tow wietqowingtanfgabga,wk eonw aekds goweg nwgna http://asdfwoiet.com/asdfiwe ahtwew https://123.123.123 wer http://agoiww.net/awerqwT?Asdasfoie/aiowe iuotwkl i oawet awltoagl ksbigwb kia giawgbwiabgkagubaibg kb dkljg s wsiowoeriwor a dm e rt yqert qwe t qw er w r wqer wqe r qw er qwe r wqer';

$start = microtime(true);
function CutName($var,$len=40)
{
    return strlen($var)>$len?(substr($var,0,$len-3).'...'):$var;
}

function prlinks($matches)
{
    return '<a class="link" href="'. $matches[1] .'" target="_blank">'. CutName($matches[1]) .'</a>';
}

$pr = preg_replace_callback('@(http://[^\s]*)@i','prlinks',$line);

$end = microtime(true);
$diff1 = $end - $start;

echo 'Test 1: ' . $diff1 . " seconds.\n";

$start = microtime(true);
$parts = explode(' ', $line);
foreach ($parts as $key => $part) {
if (strpos($part, 'http') === 0 && strlen($part) > 39) {
	$parts[$key] = '<a class="link" href="' . $part . '" target="_blank">' . substr($part, 0, 10) . '...' . substr($part, -4, 4) . '</a>';
}
}
$end = microtime(true);
$diff2 = $end - $start;

echo 'Test 2: ' . $diff2 . " seconds.\n";

if ($diff2 > $diff1) echo "Test1 is faster.\n";
else echo "Test2 is faster.\n";

 

Results:

[bash]$ php test.php
Test 1: 0.00012400000000001 seconds.
Test 2: 3.6999999999954E-05 seconds.
Test2 is faster.
[bash]$ php test.php
Test 1: 0.0001239999999999 seconds.
Test 2: 3.7999999999982E-05 seconds.
Test2 is faster.
[bash]$ php test.php
Test 1: 0.00011700000000003 seconds.
Test 2: 3.6000000000036E-05 seconds.
Test2 is faster.
[bash]$ php test.php
Test 1: 0.000112 seconds.
Test 2: 3.7000000000065E-05 seconds.
Test2 is faster.
[bash]$ php test.php
Test 1: 0.00012299999999998 seconds.
Test 2: 3.6999999999954E-05 seconds.
Test2 is faster.

 

Maybe perhaps the string is too short?

 

I also noticed I used $implode on my previous code. Here's the corrected code. Sorry about that.

<?php
$line="this is a sample website http not working while http://www.google.com should work. Now a more complicated example http://www.google.com/search?q=preg_replace";

$parts = explode(' ', $line);
foreach ($parts as $key => $part) {
     if (strpos($part, 'http') === 0 && strlen($part) > 39) {
          $parts[$key] = '<a class="link" href="' . $part . '" target="_blank">' . substr($part, 0, 10) . '...' . substr($part, -4, 4) . '</a>';
     }
}
echo implode(' ', $parts);

 

 

Edit: defining the functions outside of $start = microtime(true); for Test1 didn't improve the time by much. Test2 is still faster. Well those are my test. Someone else is free to post up their results as well.

[laffin]

I think something is wrong I get Some completely different results

<?php
header('Content-type: text/plain');

$line="this is a sample website http not working while http://www.google.com shud work.
Now a more complicated example http://www.google.com/search?q=preg_replace
secure connection https://docs.google.com";

function CutName($var,$len=40)
{
    return ;
}

function prlinks($matches)
{
    return '<a class="link" href="'. $matches[1] .'" target="_blank">'. (strlen($matches[1])>20?(substr($matches[1],0,17).'...'):$matches[1]) .'</a>';
}
$loops=500;

$at=0;
for($i=0;$i<$loops;$i++)
{
    $st=microtime(TRUE);
    $pr = preg_replace_callback('@(https?://[^\s]*)@i','prlinks',$line);
    $at+=(microtime(TRUE)-$st);
}
$at/=$loops;
$at=number_format($at,6);

echo "Preg Match (Avg time = {$at} secs)\n";
echo "{$pr}\n";

$at=0;
for($i=0;$i<$loops;$i++)
{
		$st=microtime(TRUE);
	$parts = explode(' ', $line);
	foreach ($parts as $key => $part) {
   		if (strpos($part, 'http') === 0 && strlen($part) > 39) {
      	$parts[$key] = '<a class="link" href="' . $part . '" target="_blank">' . substr($part, 0, 10) . '...' . substr($part, -4, 4) . '</a>';
   		}
	}
	$ss=implode(' ',$parts);
	$at+=(microtime(TRUE)-$st);
}
$at/=$loops;
$at=number_format($at,6);
echo "\nSubstr (Avg time = {$at} secs)\n";
echo "{$ss}\n";
?>

 

Result

Preg Match (Avg time = 0.000285 secs)

this is a sample website http not working while <a class="link" href="http://www.google.com" target="_blank">http://www.google...</a> shud work.

Now a more complicated example <a class="link" href="http://www.google.com/search?q=preg_replace" target="_blank">http://www.google...</a>

secure connection <a class="link" href="https://docs.google.com" target="_blank">https://docs.goog...</a>

 

Substr (Avg time = 0.000592 secs)

this is a sample website http not working while http://www.google.com shud work.

Now a more complicated example <a class="link" href="http://www.google.com/search?q=preg_replace

secure" target="_blank">http://www...cure</a> connection https://docs.google.com

 

 

Looks like something went sour on yer code, so time to go back and debug, and make yer code more complex.

 

 

[laffin]

Just some updated code

<?php
header('Content-type: text/plain');

$line="this is a sample website http not working while http://www.google.com shud work.
Now a more complicated example http://www.google.com/search?q=preg_replace
secure connection https://docs.google.com
http://www.php.net/ on a newline
mailto links mailto:someone@gmail.com
ftp://yoursite.com
telnet:myfunky.bbs";

function CutName($var,$len=40)
{
    return strlen($var)>$len?(substr($var,0,(($len-3)/2)).'...'.substr($var,(($len-3)/2)-$len)):$var;
}

function prlinks($matches)
{
    return '<a class="link" href="'. $matches[1] .'" target="_blank">'. CutName($matches[4]) .'</a>';
}

$pr = preg_replace_callback('@((https?|ftp|mailto|telnet):(//)?([^\s]{4,}))@i','prlinks',$line);

echo "Preg Match\n";
echo "{$pr}\n";

?>

 

Preg Match

this is a sample website http not working while <a class="link" href="http://www.google.com" target="_blank">www.google.com</a> shud work.

Now a more complicated example <a class="link" href="http://www.google.com/search?q=preg_replace" target="_blank">www.google.com/search?q=preg_replace</a>

secure connection <a class="link" href="https://docs.google.com" target="_blank">docs.google.com</a>

<a class="link" href="http://www.php.net/" target="_blank">www.php.net/</a> on a newline

mailto links <a class="link" href="mailto:someone@gmail.com" target="_blank">someone@gmail.com</a>

<a class="link" href="ftp://yoursite.com" target="_blank">yoursite.com</a>

<a class="link" href="telnet:myfunky.bbs" target="_blank">myfunky.bbs</a>

[/code]

 

thats pretty feature rich for making links now Enjoy