PHP: preg_replace_callback

Description

mixed preg_replace_callback ( mixed pattern, callback callback, mixed subject [, int limit] )

The behavior of this function is almost identical to preg_replace(), except for the fact that instead of replacement parameter, one should specify a callback that will be called and passed an array of matched elements in the subject string. The callback should return the replacement string.

Przykład 1. preg_replace_callback() example


<?php

// this text was used in 2002

// we want to get this up to date for 2003

$text = "April fools day is 04/01/2002\n";

$text.= "Last christmas was 12/24/2001\n";

// the callback function

function next_year($matches) 

{

  // as usual: $matches[0] is the complete match

  // $matches[1] the match for the first subpattern

  // enclosed in '(...)' and so on

  return $matches[1].($matches[2]+1);

}

echo preg_replace_callback(

            "|(\d{2}/\d{2}/)(\d{4})|",

            "next_year",

            $text);

// result is:

// April fools day is 04/01/2003

// Last christmas was 12/24/2002

?>

You'll often need the callback function for a preg_replace_callback() in just one place. In this case you can use create_function() to declare an anonymous function as callback within the call to preg_replace_callback(). By doing it this way you have all information for the call in one place and do not clutter the function namespace with a callback functions name not used anywhere else.

Przykład 2. preg_replace_callback() and create_function()


<?php

/* a unix-style command line filter to convert uppercase

 * letters at the beginning of paragraphs to lowercase */

$fp = fopen("php://stdin", "r") or die("can't read stdin");

while (!feof($fp)) {

    $line = fgets($fp);

    $line = preg_replace_callback(

        '|<p>\s*\w|',

        create_function(

            // single quotes are essential here,

            // or alternative escape all $ as \$

            '$matches',

            'return strtolower($matches[0]);'

        ),

        $line

    );

    echo $line;

}

fclose($fp);

?>

See also preg_replace(), create_function(), i informacje o typie callback.

User Contributed Notes


I've found php will crash when I store a reference to the $match parameter but there's an easy work around...



$mylist=array();



function my_callback($match) {

    global $mylist;

    

    // this crashes PHP...

    // $mylist[]=$match;

    

    // this copies the $match array so is safe

    $mylist[]=array_merge($match);

    

    return 1;

}



$subject="The rain in Spain falls mainly on the plane.";



echo "<p>".preg_replace_callback("/[aeiou]/i", "my_callback", $subject);

echo "<p>"; print_r($mylist);


Instead of using preg_replace_callback() you could use preg_replace() like below:



<?php



  // ** Simple function: **

     function my_function( $my_param_one, $my_param_two ) {

             // ...

     }

     preg_replace( "! \{(.+)-(.+)\} !e", "my_function('\\1', '\\2')", $input_string );



  // ** Class: **

    class my_class {



            function my_class() {

              preg_replace( "! \{(.+)-(.+)\} !e", "\$this->my_function('\\1', '\\2')", $input_string );

            }



            function my_function( $my_param_one, $my_param_two ) {

              // ...

            }

    }



?>


A powerful php template parser may be written in one line of code :)

(Of course there are many ways of it, this is my own just developed)



print preg_replace_callback("/<\\?(.*?)\\?>/",

create_function('$m','extract($GLOBALS);

return(eval("return($m[1]);"));'),$s);



Template example:

<?$var["x1"]?><br>                          

<?$var["x2"]?><br>                          

<?$var["x1"]." php foreva ".$var["x2"]?><br>



So, as you can see, <? and ?> are template parser tags and it is even possible to make simple php expressions inside such tags. However not too complex to confuse designer :)

You may add \" around inner $m[1] at your taste and templates will be little different ;)

Another advantage of such parser that you do not need to worry about variables must be available in templates - all globals and global arrays will be :) Very handy in fact.



As I saw above, this can be optimized by making separate (not inline) callback function, and I will of courde do it in real program, but this is less gracefull than I wrote here :)


There is a small typo with the code (at least when I try to implement it) in the first comment.  



<?php 

echo preg_replace("/(<\/?)(\w+)([^>]*>)/e", "html_body('\\1', '\\2', '\\3', '$extra_info')", $html_body);

?>



should really be



<?php 

echo preg_replace("/(<\/?)(\w+)([^>]*>)/e", html_body('\\1', '\\2', '\\3', '$extra_info'), $html_body);

?>



That is, the second argument should not have quotes around it.  This should apply to all of the other examples in the comments as well.


if you want to send some extra parameters to your call back function here's how to do it (using preg_replace instead!). i took the example from the manual, on how to convert html tags to upper case...



if you don't use a class:



<?php

$extra_info = "some extra info";



echo preg_replace("/(<\/?)(\w+)([^>]*>)/e", "html_body('\\1', '\\2', '\\3', '$extra_info')", $html_body);



function html_body($open_tag, $the_tag, $close_tag, $extra_info) {



         return $before . strtoupper($middle) . $after;

}

?>



class example:



<?php



class SomeClass {



     function SomeClass() {

          $extra_info = "some extra info";

          echo preg_replace("/(<\/?)(\w+)([^>]*>)/e", "\\$this->html_body('\\1', '\\2', '\\3', '$extra_info')", $html_body);

     }



     function html_body($open_tag, $the_tag, $close_tag, $extra_info) {



              return $before . strtoupper($middle) . $after;

     }

}

?>



now, if you'd like to pass an array :)



<?php

$extra_info = ("some extra info", "little more", "a bit more");



echo preg_replace("/(<\/?)(\w+)([^>]*>)/e", "html_body('\\1', '\\2', '\\3', \\$extra_info)", $html_body);



function html_body($open_tag, $the_tag, $close_tag, $extra_info) {



         return $before . strtoupper($middle) . $after;

}

?>



you can also pass the array as a reference...cool!



hope this helps someone! ;)


Example given for preg_replace_callback with create_function have one flow - it consume a lot of memory. Each call to create_function allocate some memory to create a new function, and to the end of the script a lot of memory could be consumed, if there are many lines recieved from stdin. Of course after ending all memory will be freed, but I have been hanged up my computer processing 20Mb file with similar code on Windows.



<?php

 $fp = fopen("php://stdin", "r") or die("can't read stdin");

 while (!feof($fp)) {

   $line = fgets($fp);

   $line = preg_replace_callback('|<p>\s*\w|',

// this will create as many new functions, as there are lines recieved

     create_function(

       '$matches',

       'return strtolower($matches[0]);'

     ), 

// 

     $line

   );

   echo $line;

 }

 fclose($fp);

?>


a repost (with correction) of a note originally attributed to matthew at de-construct dot com (26-Aug-2003 04:48) - a note i found in a compiled help file but which i do not see on this site.  he was missing a stripslashes call in his example, which i have fixed below:



<?

$outer_data = array('foo'=>'abc','bar'=>'123');

$callback = create_function('$matches',

'

$inner_data = unserialize(stripslashes("'. addslashes(serialize($outer_data)) .'"));

// do something here

return $inner_data[$matches[1]];

'

);

$template_data = preg_replace_callback("/\{[$](.*?)\}/", $callback, $template_data);

?>



i found this very useful for implementing a template engine!


I thought I'd post this as I'd been using the function preg_replace_callback with NP within functions, but as soon as I re-wrote my script as a class, it all fell apart.



When debugging I ended up writing test code and this is what follows. It works as I'd expect, YMMV of course.



<?php



// v v over simplified

class foo

{

  function parse()

  {

    $pattern = "/<a(.*?)href\s*=\s*['|\"|\s*](.*?)['|\"|>](.*?)>(.*?)<\/a>/i";

    $string = "<a class='whatever' href='http://foo.com' target='_blank'>foo</a>";

    print preg_replace_callback($pattern,array($this,'cb'),$string);

  }



  function cb($matches)

  {

    return "<a" . $matches[1] . "href='http://someothersite.com/foo.php?page=" . urlencode($matches[2]) . "'" . $matches[3] . ">" . $matches[4] . "</a>";

  }



}



$bar = new foo();

$bar->parse();



/**

output is

<a class='whatever' href='http://someothersite.com/foo.php?page=http%3A%2F%2Ffoo.com' target='_blank'>foo</a>

*/



?>


You can use this function to display url address as a link or image based on suffix of the url. 





Every string beginning with www or http(s):// will be displayed as an URL address.  But if the string ends with *.jpg or *.gif, it will be displayed as an image.





Please note: This function also verifies the size of the image. If is it more than 600x400, it changes it.


<?php





$pattern_html = "/\b((http(s?):\/\/)|(www\.))([\w\.]+)";


$pattern_html .= "([\#\,\/\~\?\&\=\;\%\-\w+\.]+)\b/i";





$text = preg_replace_callback($pattern_html,'Check_if_Image',$text);


.


.


function Check_if_Image($matches) {


$suffix = strtolower(substr($matches[0],-4));


if ( $suffix == '.jpg' or $suffix == '.gif') {


    $dsn = 'http'.$matches[3].'://'.$matches[4].$matches[5];


    if (list($width, $height, $type, $attr) = getimagesize("$dsn")) {


    if ( ($width > 600) ) {


        $koef = $width / 600;


        $width = 600;


        $height /= $koef;


        }


    if ( ($height > 400) ) {


        $koef = $height / 400;


        $height = 400;


        $width /= $koef;


        }


    }


    else {


        $height=400;


        $width=600;


    }





    $ret = "<img src=\"$dsn\"";


    $ret .= "\" border=0 width=\"$width\" height=\"$height\">";


    }


else {


    $ret = '<a href="http'.$matches[3].'://'.$matches[4].$matches[5]; 


    $ret .='" target="_blank">'.$matches[0].'</a>';


    }


return ("$ret");


}





?>


Here's an example of how to evaluate PHP code inside a HTML block. In my example I have two timestamps which I insert at several places in a HTML-template. The formatting for the dates are put inside the template. When the insert is done the code is evaluated and the dates are formatted. All code inside "<php>" to "</php>"-tags will be evaluated.

-----------------------------------------------------------

function evalcode($matches)

{

eval("\\$matches[2] = $matches[2];");

return $matches[2];

}



$template = "<html><body>This really simple HTML-page that was created at <php>date('Y-m-d',{datecreated})</php> and modified at <php>date('H:i:s',{datemodified})</php> on the <php>date('dS',{datemodified})</php> of <php>date('F Y',{datemodified})</php>.</body></html>";



$datecreated = 1065162746;

$datemodified = time();

$output = $template;

$output = str_replace('{datecreated}', $datecreated, $output);

$output = str_replace('{datemodified}', $datemodified, $output);

$output = preg_replace_callback("/(<php>)(.*)(<\\/php>)/U", "evalcode", $output);

echo $output;

----------------------------------------------------------

The output will be something like "This really simple HTML-page that was created at 2003-10-03 and modified at 16:58:31 on the 27th of October 2003."