Regular expression woes

dacloo

Hi!

I have quite a difficult problem.
The HTML page is stored in one big string.
I need to convert clean urls to traditional ones:

index.php/read/2/4

to:

index.php?action=read&id=2&parid=4

So basicly, between the content of the HREF tag <A HREF="[changes]">
...I have to change the URL.

Rules:
1. Replace first / character with ?action=
2. Replace second / character with &id=
3. Replace third / character with &parentid=

Does anyone has a clue if this is possible with preg_replace, and how? I find it very hard.

bubblenut

Something like this should do the trick

preg_replace('/index.php\/(.*)\/(.*)\/[^"\' ]/G', 'index.php?action=\1&id=\2&parid=\3', $string);

dacloo

Hi!

Thank you for your reply!

I had to add a \ because of a PHP parse problem.

PHP outputs:
Warning: Unknown modifier 'G' in m:\internet\websites\cms3\test_url.php on line 6

$message = preg_replace('/index.php\/(.*)\/(.*)\/[^"\' ]/G', 'index.php?action=\1&id=\2&parid=\3', $HTML);

And...you entered static values for action,id and parid. But, those values are dynamic, they can have any value.

EDIT:

$message = preg_replace('/index.php\/(.*)\/(.*)\/(.*)\/[^"\' ]/', 'index.php?action=\\1&id=\\2&parid=\\3', $HTML);

The above is coming closer to what I want.

bubblenut

Hey, sorry, couldn't remember what the modifier for ungreedy matching was and couldn't be bothered to look it up :o so I just put the G in the modifier is supposed to be U, you need this otherwise you'll have some very odd results if you have more than one occurence of the pattern on any one line of the string. Also, you need a * ofter the character class. So it should look like this

$message = preg_replace('/index.php\/(.*)\/(.*)\/(.*)\/[^"' ]*/U', 'index.php?action=\\1&id=\\2&parid=\\3', $HTML);

Also, in retrospect, it might be wise to be little more specific than just .* on the three parts because there are certain characters which should not be accepted in that, take this as an example.

<a href="index.php/one" title="dog/cat">Domestic Pets</a>

would be transformed to

<a href="index.php?action=one" title="dog&id=cat">Domestic Pets<&parid=a>

Not what you're wanting I'd imagine.

dacloo

Hey!

Thank you so much. Much appriciated.

	$cnt = preg_replace('/index.php\/(.*)\/(.*)\/(.*)\/[^" ]*/U', 'bb.php?action=\\1&id=\\2&parid=\\3', $cnt);
	$message = preg_replace('/index.php\/(.*)\/(.*)\/[^" ]*/U', 'index.php?action=\\1&id=\\2', $message);
	$message = preg_replace('/index.php\/(.*)\/[^" ]*/U', 'index.php?action=\\1', $message);

I did three seperate expressions; some links may contain less than three parameters.

Indeed the same problem as you described is happening.
in the closing </a> tag, I get garbage.
But, how can I tell the expression to only change stuff between the a contents? (<a href="ONLYCHANGEHERE"></a>)

bubblenut

Well, we can put the constrint straight in there. Basically, what you want is to match any character except the cclosing double quote of the src tag (ie [^"])
So..

$cnt = preg_replace('/index.php\/(.*)\/(.*)\/(.*)\/[^" ]*/U', 'bb.php?action=\\1&id=\\2&parid=\\3', $cnt);

becomes

$cnt = preg_replace('/index.php\/([^"]*)\/([^"]*)\/([^"]*)\/[^" ]*/U', 'bb.php?action=\\1&id=\\2&parid=\\3', $cnt);

HTH
Bubble

dacloo

hmmmm weird, but it doesn't seem to work; nothing is "found" by the regular expression. Are these [] signs right?

tsinka

Hi,

I messed around a little bit with just one preg_replace but finally took that aside and came to the following code example:

<?PHP
$test = '<a href="index.php/one">one</a>'
       .'<a href="index.php/only/two">two</a>'
       .'<a href="index.php/aaa/bbb/ccc">three</a>'
       .'<a href="index.php/111/222/333/444">four</a>'
       .'<a href="index.php/111/222/333/444/555">five</a>';

$arrSubst = array("bb.php","action","id","parid","parid2","parid3");

preg_match_all('/index.php[^"> ]*/',$test,$matches);

print_r($matches);

if (!empty($matches)) {
  foreach($matches[0] as $url) {
    $newurl = "";
    $arrParts = explode("/",$url);

foreach($arrParts as $key => $value) {
  if ($key != 0) {
    $newurl .= ($key == 1) ? "?" : "&";
    $newurl .= $arrSubst[$key]."=".$value;
  } else {
    $newurl .= $arrSubst[$key];
  }
}

$test = str_replace($url.'"',$newurl.'"',$test);
  }
}

echo $test;
?>

This works only with a hrefs where the url is surrounded by double quotes...there might be an easier solution but I don't find it.

Thomas

tsinka

Ok,

just after submitting the post I think that there is an "easier" solution by using preg_replace_callback ...

Thomas

dacloo

Tsinka, that is a very smart solution. Thank you.
I can't imagine situations where the href is not surrounded by quotes, so this is a pretty good solution!

tsinka

Hi,

here's the callback one:

<?PHP
$test = '<a href="index.php/one">one</a>'
       .'<a href="index.php/only/two">two</a>'
       .'<a href="index.php/aaa/bbb/ccc">three</a>'
       .'<a href="index.php/111/222/333/444">four</a>'
       .'<a href="index.php/111/222/333/444/555">five</a>';

$arrSubst = array("bb.php","action","id","parid","parid2","parid3");

$test = preg_replace_callback('/index.php[^"> ]*/',"callb",$test);

echo $test;

function callb($matches) {
  $newurl = "";
  $arrParts = explode("/",$matches[0]);

  foreach($arrParts as $key => $value) {
    if ($key != 0)
      $newurl .= ($key == 1) ? "?" : "&";

$newurl .= $GLOBALS['arrSubst'][$key];

if ($key != 0)
  $newurl .= "=".$value;
  }

  return $newurl;
}
?>

Thomas

tsinka

Hint:

You want to have this code inside a class and the callback function is inside the class, too ?

Replace

$test = preg_replace_callback('/index.php[^"> ]*/',"callb",$test);

$test = preg_replace_callback('/index.php[^"> ]*/',array($this,"callb"),$test);

in that case.

Thomas

dacloo

Thanks. For now, I haven't done much with classes (which is my next goal for PHP in general).
Only the template engine is classed based, and the rest highly procedural.