parsing for URLs and cutting down long ones...

dbmasters

I built another forum and am trying to correct an issue I have had in others...I have always looked for URLs that start with [url]http://[/url] and autolink them. No prob:

preg_replace("/(http:\/\/+.[^\s]+)/i",'<a href="\1">\1</a>', $body);

No, reg exp is not my strongest skill, however, what I want to do is find URL's that start with [url]http://[/url] or www. (but not catch the same URL twice such as if it is http://www.

That should be a pieceof cake for somebody that knows reg exp better than I...now the tricky one...

In addition I want to scan a URL's length if it's over a given length, take some of the front and the back of the string, put a "..." in the middle and link it...I tried this:

$url=preg_match("/(http:\/\/+.[^\s]+)/i", $input);
if(strlen($url) > 50)
{
$display_url=substr($url,40)."...".substr($url,-7);
}
else
{
$diplay_url=$url;
}
$input=preg_replace("/$url/",'(<a href="$url" target="_blank">$diplay_url</a>)', $input);

as well as some preg_match_all loops and str_replace in different context, but none of them seem to work...can anyone help me?!

drew010

the expression below will perform a replace on urls beginning with http:// or www.

search:  /\s(http:\/\/|www\.)(\S+)/
replace:  <a href="\\1\\2" target="_new">\\2</a>

you could use that same search code for preg match as well if you want to then check the length of the resulting url for shortening.

here is a set of 3 regexps, you could even add the last one in that i made for someone on here a while back that does what this forum does with [url ] tags.

$patterns = array("/\[url\](.*)\[\/url\]/", "/\s(f|ht)(tps?:\/\/)(\S+)/", "/\[url=(.*)\](.*)\[\/url\]/", "\s(http:\/\/|www\.)(\S+)/");
$replaces = array("<a href=\"\\1\">\\1</a>", " <a href=\"\\1\\2\\3\">\\1\\2\\3</a>", "<a href=\"\\1\">\\2</a>");

$string = preg_replace($patterns, $replaces, $string);

matchs ftp:// or http://, [url ]http: //site[/ url], and [ url=http://site.com]text[/ url]

of course remove the spaces, i just did that so it wouldnt pick up what i typed.

dbmasters

dude, you rock, that's great, thanks.

The problem I think I was having with that is what if there is more than one URL in any given context and checking each one for length, etc...and how to link them all...but I may have just figured that out...maybe not...

dbmasters

Ok, here is what I am trying and it got me a little way:

$chkURL=preg_match('/\s(http:\/\/|www.)(\S+)/',$input,$out);
for($i=0; $i<$chkURL; $i++)
{
if (!eregi('array',$out[1][$i]))
{
// $uStart=substr($chkURL,30);
// $uEnd=substr($chkURL,17);
// $shortURL=$uStart."...".$uEnd;
$linkage="<a href=\"".$out[1][$i]."\" target=\"blank\">".$out[1][$i]."</a>";
$input=str_replace($out[1][$i],$linkage,$input);
}
}

I won't try using my code to shrink the URL until I get the main part working, so it's commented out...but now it just hyperlinks the "h" in the URL...this is driving me batty, what am I missing?

TheDefender

This should get you going with the substr...

$chkURL=preg_match('/\s(http:\/\/|www\.)(\S+)/',$input,$out); 
$lengthofurl = strlen($chkURL);

for($i=0; $i<$chkURL; $i++) 
{ 
if (!eregi('array',$out[1][$i])) 
{ 
  if($lengthofurl > 40) {
   $uStart=substr($chkURL,0,15); 
   $uEnd=substr($chkURL,-15); 
   $shortURL=$uStart."...".$uEnd;
  } else {
    $shortURL=$chkURL;
  } 
$linkage="<a href=\"".$out[1][$i]."\" target=\"blank\">".$out[1][$i]."</a>"; 
$input=str_replace($out[1][$i],$linkage,$input);

} 
}

Or something along those lines... Basically, if it is more than 40 characters, it shows the first 15, followed by ... and the last 15. The syntax should be OK for you, but the method for doing it may need some tweaking...

so
http://www.whatever.com/thispage.php?threadid=hsgajdgs7646dghajgsd73647&postid=HSJHD7987SH

would become:
http://www.what...tid=HSJHD7987SH

dbmasters

thanks for that.

It doesn't through any errors, but it doesn't link the URL either...hmmm

TheDefender

Do this, and see if it works. We'll go from there.

$chkURL = "whatever.com/thisandthat/yaddayadda/index.php?id=123&yadda=BAD";
$lengthofurl = strlen($chkURL);
if($lengthofurl > 40) {
  $uStart=substr($chkURL,0,15);
  $uEnd=substr($chkURL,-15);
  $shortURL=$uStart."...".$uEnd;
  } else {
    $shortURL=$chkURL;
  }
echo $shortURL;

// OUtput:
// whatever.com/th...d=123&yadda=BAD

dbmasters

yeah, actually I tried something very similar myself, it builds the short URL properly:

whatever.com/th...d=123&yadda=BAD

I am thinking the error is somewhere in:

$chkURL=preg_match('/s([url]http://[/url]|www.)(S+)/',$input,$out);

but what do I know...

TheDefender

I will look at the regular expression part a little later... (RegExp is not my strong suit) Hopefully until then, drew010 will pop back in and offer his help...

Right now, I am going to go get some dinner with the family...

drew010

here is something i wrote up that seems to do the trick. it may need some minor modifications for certain causes because i didnt play with it too much. here was my test code.

$link = <<<EOD
this is a test long link
http://www.drew-phillips.com/scripts/download.php?file=This.zip&dir=blah
here is another slightly big
link with no http:// www.drew-phillips.com/scripts/download.php[/url] fdsafasdfasfwefwe
go to www.google.com 
please, its short!!!
EOD;

$count = preg_match_all("/\s(http:\/\/|www\.)(\S+)/", $link, $matches);

for ($i = 0; $i < sizeof($matches); $i++) {

if (strlen($matches[2][$i]) > 30) {
	$beg = substr($matches[2][$i], 0, 15);
	$end = substr($matches[2][$i], -15);
	$link = preg_replace("/\s".preg_quote($matches[1][$i].$matches[2][$i], "/")."/", " <a href=\"http://{$matches[2][$i]}\">$beg...$end</a>", $link);
} else {
	$link = preg_replace("/\s".preg_quote($matches[1][$i].$matches[2][$i], "/")."/", "<a href=\"http://{$matches[2][$i]}\">{$matches[2][$i]}</a>", $link);
}	
}

echo $link;

dbmasters

thanks, that close, but not quite...I'll have to mess with it some more, it doesn't complete the a tags...argh...

TheDefender

That's beautiful drew010, straight into the snippet library that one goes...

I would make one suggestion though... I would change the conditional statement:

if (strlen($matches[2][$i]) > 30) {

To 40 rather than 30. The reason is, if it's 31, then the split will occur at 15, and 17, thus only getting rid of 1 character, and adding 3 (the 3 periods) so essentially, your original URL is actually shorter than your result. Give it a bit of cushio. With 40, then you have a 15 character removal every time, before the priods are inserted... Just my observation though... Otherwise, this is good stuff, thanks drew010!

P.S. That worked perfectly for me.

dbmasters

OK, well, the only string in my text was:

http://www.dytzgfbfghbfghhjsth.agag/sgfagaergarg/aergaergaerg/aergaergaerg/aerg.gif

and what it returned was:

http://www.dytzgfbfghbfghhjsth.agag/sgfagaergarg/aergaergaerg/aergaergaerg/aerg.gifhref="http://">href="http://">

Additionally, if there is not a space at the end of the post (or the URL at least) the URL doesn't link at all...

Whereas "/((http:\/\/|www.)+.[^\s]+)/i" does...but neither seem to work for me when shorteneing the URL.

drew010

defender, i took your advice and changed it to 40, good call. i also played with it so you can have a link at the end of the post without whitespace after it. and fixed up the replacement a little bit. it seemed to work ok for me. here's the new snipped and the example output

<?php

$link = <<<EOD
this is a test long link
http://www.dytzgfbfghbfghhjsth.agag/sgfagaergarg/aergaergaerg/aergaergaerg/aerg.gif
lines
http://www.drew-phillips.com/scripts/download.php?file=This.zip&dir=blah
here is another slightly big
link with no http:// www.drew-phillips.com/scripts/download.php fdsafasdfasfwefwe
go to www.google.com please, its short!!!
www.someurl.com/at_end_of_string
EOD;

$count = preg_match_all("/(\s)(http:\/\/|www\.)(\S+)/i", $link, $matches);

echo $count;

for ($i = 0; $i < $count; $i++) {

if (strlen($matches[3][$i]) > 40) {
	$beg = substr($matches[3][$i], 0, 15);
	$end = substr($matches[3][$i], -15);
	$link = preg_replace("/".preg_quote($matches[1][$i].$matches[2][$i].$matches[3][$i], "/")."/", "{$matches[1][$i]}<a href=\"http://{$matches[3][$i]}\">$beg...$end</a>", $link);
} else {
	$link = preg_replace("/".preg_quote($matches[1][$i].$matches[2][$i].$matches[3][$i], "/")."/", "{$matches[1][$i]}<a href=\"http://{$matches[3][$i]}\">{$matches[3][$i]}</a>", $link);
}	
}

echo $link;

?>

this is a test long link
<a href="http://www.dytzgfbfghbfghhjsth.agag/sgfagaergarg/aergaergaerg/aergaergaerg/aerg.gif">www.dytzgfbfghb...rgaerg/aerg.gif</a>
lines
<a href="http://www.drew-phillips.com/scripts/download.php?file=This.zip&dir=blah">www.drew-philli...is.zip&dir=blah</a>
here is another slightly big
link with no http:// <a href="http://drew-phillips.com/scripts/download.php">drew-phillips.com/scripts/download.php</a> fdsafasdfasfwefwe
go to <a href="http://google.com">google.com</a> please, its short!!!
<a href="http://someurl.com/at_end_of_string">someurl.com/at_end_of_string</a>

dbmasters

Ooo Ooooo, Oooo, now we're cookin with gas! OK...not it's just tacking a "www." on the end of the link and it didn't shrink it, but hey, it's as close as it has been yet THANKS! I'll start tweaking!

Where does matches[3] come from?

This is a cool little bit of code...

drew010

matches[3] is the actual domain portion of the url. matches[1] is the type of whitespace before the url, either newline or space most likely, matches[2] is http:// or www. and matches[3] is the remaining part of the url.
all i can suggest to get it working is check your variables because im not sure why the example would work here but not live on your site unless something more than the variable changes.

dbmasters

Well, don't understand this at all...I just can not get this to work properly, but for now I did get a:

$input=preg_replace("/[link=(.[^[]+)](.[^{[]+)[\/link]/is",'<a} href="\1" target="_blank">\2</a>', $input);

to give a linking option to my forum users, so that'll work until I can get this other function working properly. I don't get it, the code looks good, it works in a given test senario oin your example, but not in my forum contents...argh

drew010

i just made a test php page on my site. you can go there and try typing in there what doesn't work on your site and see if the same thing occurs. links seemed to work fine for me. http://www.drew-phillips.com/php/test.php

source code of it is this:

<form method="POST">
<textarea name="text" rows="15" cols="45"></textarea>
<input type="submit">
</form>


<?php


if (isset($_POST['text'])) {

$link = $_POST['text'];


$count = preg_match_all("/(\s)(http:\/\/|www\.)(\S+)/i", $link, $matches);

echo $count;

for ($i = 0; $i < $count; $i++) {

if (strlen($matches[3][$i]) > 40) {
	$beg = substr($matches[3][$i], 0, 15);
	$end = substr($matches[3][$i], -15);
	$link = preg_replace("/".preg_quote($matches[1][$i].$matches[2][$i].$matches[3][$i], "/")."/", "{$matches[1][$i]}<a href=\"http://{$matches[3][$i]}\">$beg...$end</a>", $link);
} else {
	$link = preg_replace("/".preg_quote($matches[1][$i].$matches[2][$i].$matches[3][$i], "/")."/", "{$matches[1][$i]}<a href=\"http://{$matches[3][$i]}\">{$matches[3][$i]}</a>", $link);
}	
}
echo "<hr/>";
echo $link;
}

?>

TheDefender

Tweaking? Are you sure you are copying and pasting this thing correctly, because it works fine for me as is?!? I would guess your issue may come from your input to the preg_match_all() function or something... Post the code exactly as you have it dbmasters, so we can take a look at it. Be sure when you post it, that you use the code or php tags (using the # or PHP buttons above).

drew010, this code snippet rocks. I have all of the revs in my snippet library, so I can utilize them fully, because they can and will come in handy in some codes I am writing for a client very soon...

drew010

thanks for the comment. in case you didnt grab it before i posted the other snippet that works the same as the bb code on this board letting you use [url ] tags and [url= tags and just detects links, here it is.

$patterns = array("/\[url\](.*)\[\/url\]/", "/(\s)(f|ht)(tps?:\/\/)(\S+)/", "/\[url=(.*)\](.*)\[\/url\]/", "(\s)(http:\/\/|www\.)(\S+)/");
$replaces = array("<a href=\"\\1\">\\1</a>", " <a href=\"\\1\\2\\3\\4\">\\2\\3\\4</a>", "\\1<a href=\"\\2\">\\3</a>");

$string = preg_replace($patterns, $replaces, $string);

this doesnt do the shortening but something similar to the topic of this post could easily be done.