ansaurus

Question

preg_replace array: mask regex characters problem

Answer 1

A:

Im not entirely sure what you are trying to do but i saw "breaks when the word groups include regex syntax characters" which makes me think that all you need to do is escape these characters... ie put a \ before them.

EDIT:

Im getting pretty stuck with this as well, but if if show you what I've got, maybe it could help you out:

<?php

$text = "<html><body><pre>
Replace all foo / bar / baz cases here:
Case 1: Text Foo text.
Case 2: Text 'Foo' Bar text Foo.
Case 3: Text Foobar (2) text.
Case 4: Text Bar & Baz.
Case 5: Text Bar Baz?
Case 6: Text Bar? & Baz?
Case 7: Text Bar-X.

Replace nothing here (text followed by brackets) or [inside square brackets]: 
Case 1: Text Foo (text).
Case 2: Text 'Foo' Bar (text) Foo (text).
Case 3: Text Foobar (2) (text).
Case 4: Text Bar & Baz (text).
Case 5: Text Bar Baz (text).
Case 6: Text Bar? & Baz (text).
Case 7: Text Bar-X (text).
Case 8: [Text Foo]
</pre></body></html>";

function convertRegexChars($string)
{
    $converted = str_replace("?","&#63;",$string);
    $converted = str_replace(".","&#46;",$converted);
    $converted = str_replace("*","&#42;",$converted);
    $converted = str_replace("+","&#43;",$converted);
    return $converted;
}

$s = array(
  array("t" => "Foo",         "u" => "http://www.foo.net"),
  array("t" => "'Foo' Bar",   "u" => "http://www.foo.net"),
  array("t" => "Foobar (2)",  "u" => "http://www.foo.net"),
  array("t" => "Bar & Baz",   "u" => "http://www.foo.net"),
  array("t" => "Bar Baz?",    "u" => "http://www.foo.net"),
  array("t" => "Bar? & Baz?", "u" => "http://www.foo.net"),
  array("t" => "Bar-X",       "u" => "http://www.foo.net")
 );

$replaced = convertRegexChars($text);
foreach ($s as $i => $row) {
    $txt = convertRegexChars($row['t']);
    $replaced = preg_replace('/(?='.$txt.'[^\]][^(])\b'.$txt.'\b/mS',
                           '<a href="'.$row["u"].'">'.$txt.'</a>',
                           $replaced);
 }
echo $replaced;

?>

Chief17 2010-06-19 12:22:56

Thanks for the fast reply. I thought, that's what preg_quote() and \Q \E are there for. I can't replace them in $text and don't want to in $s.So far I tried '.preg_quote($row["t"]).', '.preg_quote($row["t"], '/').' and \Q'.$row["t"].'\E but the result is never as desired.

Martin 2010-06-19 12:35:18

Chief17 2010-06-19 12:37:53

Martin 2010-06-19 12:45:33

Check my **Edit**

Chief17 2010-06-19 12:52:04

Unfortunatly that's not very useful for me. I need the regex because it shouldn't apply when the term is followed by brackets or inside square brackets.Also, there will be thousands of terms to be replaced, so an unindexed, light-weight and multi-dimensional array as I set up on top is needed. The regex itself works fine, it just doesn't work on word groups using regex characters like + ? / ( etc. So I need to mask them. I tried all variations I can think of but it won't work for all cases. And as mentioned I can't mask them in $text or $s.

Martin 2010-06-19 12:56:19

Ahh ok, i see what your trying to do now :oPIve just run your script and it seems to work fine as far as i can see. Which particular array (`array("t" => "Bar-X", "u" => "http://www.foo.net")`) isnt doing as you expected?(just read your edit above, gimme a min and ill see)

Chief17 2010-06-19 13:08:58

Martin 2010-06-19 13:17:57

I see your problem now. Tricky one, let me think about it for a bit and ill get back to you in a min (to code format something in a reply enclose code in ` the key before the number 1 along the top)

Chief17 2010-06-19 13:22:21

Thanks, no hurry. I'm struggling with this for 2 days already ;)

Martin 2010-06-19 13:24:56

Made another **EDIT**, it doesn't work, but maybe a different approach might help

Chief17 2010-06-19 16:23:21

I can't believe it's that hard. Do we oversee some major thing?

Martin 2010-06-20 08:57:17

Yeah looks like we did, serg555 seemes to have got it.

Chief17 2010-06-20 10:47:47

I've overseen his reply. Thanks to you too :)

Martin 2010-06-20 11:08:44

Answer 2

+1 A:

This should work, at least at provided test cases:

$replaced = preg_replace('/([.,\s!^]+)('.preg_quote($row["t"],'/').')([.,\s!$]+)(?!\()/mS',
                           '$1<a href="'.$row["u"].'">$2</a>$3',
                           $replaced);

\b doesn't work as expected when your match itself is wrapped inside some boundaries (like in Foobar (2)), so you should specifically provide a list of allowed characters. I quickly put [.,\s!^] and [.,\s!$] there, you probably will have to add some more allowed characters according to your specs (like -, _?)

serg 2010-06-19 17:02:19

Nice one! you have an extra `)` where you are closing `preg_quote` btw

Chief17 2010-06-19 17:07:55

You are right, thanks.

serg 2010-06-19 17:08:47

Good job, I was running out of ideas :oP

Chief17 2010-06-19 17:10:33

Overseen this answer. Thanks so much!

Martin 2010-06-20 10:57:19

ansaurus

tags:

views:

answers:

preg_replace array: mask regex characters problem

related questions