Trouble with simple PHP profanity filter

Question

Hey everyone,

I am writing a simple profanity filter in PHP. Can anyone tell my why, in the following code, the filter works (it will print [explicit]) for the $vowels array and not the $lines array which I constructing from a text file?

 function clean($str){

$handle = fopen("badwords.txt", "r");
if ($handle) {
   while (!feof($handle)) {
       $array[] = fgets($handle, 4096);
   }
   fclose($handle);
}

$vowels = array("a", "e", "i", "o", "u", "A", "E", "I", "O", "U");

$filter = "[explicit]";
$clean = str_replace($array, $filter, $str);
return $clean;
 }

When using $vowels in replace of $array, it works except for lowercase vowels which return:

 [[expl[explicit]c[explicit]t]xpl[explicit]c[explicit]t]

 instead of 

 [explicit]

Not sure why that is going on, either.

Any ideas?

Thanks!

Answer 1

Because the output of the filter contains lower case vowels, which are also the characters you're filtering. Namely you're creating a feedback loop.

Answer 2

First off, file_get_contents is a much simpler function to read a file into a variable.

$badwords = explode("\n", file_get_contents('badwords.txt');

Second, preg_replace offers much more flexible string replacement options. - http://us3.php.net/preg_replace

foreach($badwords as $word) {
    $patterns[] = '/'.$word.'/';
}

$replacement = '[explicit]';

$output = preg_replace($patterns, $replacement, $input);

Answer 3

Make sure you read:

Coding Horror: Obscenity Filters: Bad Idea, or Incredibly Intercoursing Bad Idea?

before you choose to continue on the road of string replacement...

Answer 4

I modified Davethegr8's solution to get the following working example:

 function clean($str){

global $clean_words; 

$replacement = '[explicit]';

if(empty($clean_words)){
 $badwords = explode("\n", file_get_contents('badwords.txt'));

 $clean_words = array();

 foreach($badwords as $word) {
     $clean_words[]= '/(\b' . trim($word) . '\b)/si';
 }
}

$out = preg_replace($clean_words, $replacement, $str);
return $out;
 }

Answer 5

there are web services out there now to filter profanity. For example: WebPurify