Re: String manipulation out of the box solution?

[Merlin <news.groups@xxxxxx>]

Richard Lynch wrote:
> On Tue, May 10, 2005 5:58 am, Merlin said:
>
> > I am writing an internal full text search engine and do have trouble in
> > outputting the text in an apropriate way.
> >
> > Problem is that if there is more than one word I cant handle the text
> > cropping.
> >
> > For example:
> > Search term: php germany
> > Text from database: There is no such great language than php. Amongh those
> > countries using this language is Germany with a good percentage of users.
> > Text output should be:
> > ...language than php. Amongh... language is Germay with a good...
> >
> > Similar to the way google does it. I tried it now with a couple of ways
> > but
> > failed for most (I tried with strpos and substr).
> >
> > Is there a solution out of the box with php, or maybe anybody knows a good
> > script which does that. This does sound like a standard feature to me.
>
>
> Here's a quickie, untested, and probably with some kind of logic errors,
> or at least things not quite what you want.
>
> $fulltext = "There is no such great language than php. Amongh those
> countries using this language is Germany with a good percentage of
> users.";
> $words = explode(" ", "php germany");
> $snippets = '';
> while (list(, $word) = each($words)){
>   if (!stristr($snippets, $word)){ //skip this if we already got the word.
>     $start = strpos($fulltext, $word);
>     if ($start !== false){
>       $end = $start + strlen($word);
>       $jumpback = strpos(' ', $fulltext, $start - 20);
>       $jumpforward = strpos(' ', $fulltext, $end + 20);
>       $snippet = substr($fulltext, $jumpback, $jumpforward);
>       $snippets .= " $snippet ";
>     }
>   }
> }
> reset($words);
> while (, $word) = each($words)){
>   $snippets = str_replacei($word, "<b>$word</b>", $snippets);
> }
> echo $snippets;
>
>
> To Do:
>
> Might wanna store an array of start/end numbers for snippets, then sort by
> start, then combine those that "overlap" one end to the next start, *THEN*
> combine those snippets, so you don't have snippets out or order, nor
> overlapping.
>
> Still, I got ya started...

Hi Richard,

thank you for the jump start! I have fixed some errors within the script and now 
it works ( I am attaching the script). You are right, the overlap is a to do.

Best regards, Merlin

<?php

$fulltext = "There is no such great language than php. Amongh those
countries using this language is Germany with a good percentage of
users.";
$words = explode(" ", "php Germany");
$snippets = '';
while (list(, $word) = each($words)){
  if (!stristr($snippets, $word)){ //skip this if we already got the word.
    $start = strpos($fulltext, $word);
    if ($start !== false){
      $end = $start + strlen($word);
      $jumpback = strpos($fulltext, ' ',$start - 20);
      $jumpforward = strpos( $fulltext,' ', $end + 20);
      $snippet = '...'.substr($fulltext, $jumpback, $jumpforward).'...';
      $snippets .= " $snippet ";
    }
  }
}
reset($words);

while (list(, $word) = each($words)){
  $snippets = str_replace($word, "<b>$word</b>", $snippets);
}

echo $snippets;

?>

--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php