views:

149

answers:

2

I have a paragraph of text and i want to replace some words using PHP (preg_replace). Here's a sample piece of text:

This lesson addresses rhyming [one]words and ways[/one] that students may learn to identify these words. Topics include learning that rhyming words sound alike and these sounds all come from the [two]ending of the words[/two]. This can be accomplished by having students repeat sets of rhyming words so that they can say them and hear them. The students will also participate in a variety of rhyming word activities. In order to demonstrate mastery the students will listen to [three]sets of words[/three] and tell the teacher if the words rhyme or not.

If you notice there are many occurances of the word 'words'. I want to replace all the occurances that don't occur inside any of the tags with the word 'birds'. So it looks like this:

This lesson addresses rhyming [one]words and ways[/one] that students may learn to identify these birds. Topics include learning that rhyming birds sound alike and these sounds all come from the [two]ending of the words[/two]. This can be accomplished by having students repeat sets of rhyming birds so that they can say them and hear them. The students will also participate in a variety of rhyming word activities. In order to demonstrate mastery the students will listen to [three]sets of words[/three] and tell the teacher if the birds rhyme or not.

Would you use regular expressions to accomplish this?
Can a regular expression accomplish this?

+1  A: 

The tags you have don't define a regular language and thus regular expressions won't work well for this. The best I think you can do is remove all the tags (replace them with something else), replace "words" with "birds" and then put the content back.

$str = preg_replace_callback('!\[one\].*?\[/one\]!s', 'hash_one', $input);
$str = str_replace('words', 'birds', $str);
$output = preg_replace_callback('!:=%\w+%=:!', 'replace_one', $str);

$hash = array();

function hash_one($matches) {
  global $hash;
  $key = ':=%' . md5($matches[0]) . '%=:'; // to ensure this doesn't occur naturally
  $hash[$key] = $matches[0];
  return $key;
}

function replace_one($amtches) {
  global $hash;
  $ret = $hash[$matches[0]];
  return $ret ? $ret : $matches[0];
}
cletus
Ha, this is in fact the method i've used!
Gary Willoughby
+2  A: 

You can use the 'isolation' technique (my invention! :)))), outlined here

$c = new Clipboard; // see the link
$text = $c->cut($text, '~\[(\w+)\].+?\[/\1\]~');
$text = str_replace('words', 'birds', $text);
$text = $c->paste($text);
echo $text;

just for fun, a pure regexp way

$re = "~
    words
    (?=
        [^\[\]]*
        (
            (\[\w)
            |
            $
        )
    )
~x";

echo preg_replace($re, 'birds', $text);
stereofrog