How do I check for bad/good words?

[Anon]

Hi, what is the best way to modify the code below to prevent the word ass from matching a list of non bad words like harass or brass? Basically I need some code that says if ass is in $search AND $search does not contain harass or brass, then go to bad word page otherwise go to non bad word page. Thanks!

<?php
$badword = eregi("ass", "$search");
if ($badword) {
header("Location:http://www.domain.com/badwordpage.htm");
exit();
} else {
header("Location:http://www.domain.com/nobadwordpage.htm");
exit();
}
?>

[Anon]

Well I don't know if this is the most efficient solution but the following appears to work:

<?php

$badwords = "ass|whatever";
$exceptions = "harass|brass";

$bw = eregi("$badwords", "$search");
$ex = !eregi("$exceptions", "$search");

if (($bw) && ($ex)) {
header("Location:http://www.domain.com/badwordpage.htm");
exit();
} else {
header("Location:http://www.domain.com/nobadwordpage.htm");
exit();
}
?>

[Anon]

i'll be witing the same code very soon, and have yet to give it any thought. i know hardly anyting about regular expressions, but looking at the basics it would seem pretty easy to look for " ass" with a space in front of it and no alpha-numeric characters after it, to catch " ass.". or maybe i'm missing something...

[Anon]

Well, the "solution" I posted was ok for words that do not have many exceptions but in the case of ass I soon realized there are 100's of potential good words that contain the letters "ass" that would need to be added to the exceptions variable so I came up with a simpler solution that will only match ass no matter how it is entered.

<?php

$originalsearchterm = "$search";
$modifiedsearchterm = ". $search .";

$badwords = " ass |@!#$";
$testbad = eregi("$badwords", "$modifiedsearchterm");
if ($testbad) {
header("Location:http://www.domain.com/badwords.php");
exit();
} else {
header("Location:http://www.domain.com/nobadwords.php?searchterm=$originalsearchterm");
exit();
}

?>

So now words like brass, association, passage will not trigger it and they don't have to be listed/updated. 🙂

[Anon]

OMG, it is getting late, after re-reading your reply I see what I posted is basically what you were suggesting - that is to add a space around the search term. BTW, the code I posted should have been ...

Should have been ...

<?php

$originalsearchterm = "$search";
$modifiedsearchterm = ". $search .";

$badwords = " ass ";
$testbad = eregi("$badwords", "$modifiedsearchterm");
if ($testbad) {
header("Location:http://www.domain.com/badwords.php");
exit();
} else {
header("Location:http://www.domain.com/nobadwords.php?searchterm=$originalsearchterm");
exit();
}

?>

[reagent]

Try using this code snippet and modify it to your needs:

$badwords = array('ass', '@!#$');

$strText = '@!#$, this ass keeps harrassing me!';

$strSearch = '';
for ($iIndex = 0; $iIndex < sizeof($badwords); $iIndex++)
{
if ($iIndex == 0)
$strSearch.= '\b';
else
$strSearch.= '\b|\b';

$strSearch.= $badwords[$iIndex];

}
$strSearch.= '\b';

$blnBadWords = preg_match("/$strSearch/i", $strText);

if ($blnBadWords)
{
do something
}
else
{
something else
}

You can easily change this to replace the offending text as well by changing these lines:

$blnBadWords = preg_match("/$strSearch/i", $strText);

if ($blnBadWords)
{
do something
}
else
{
something else
}

to:

$strText = preg_replace("/$strSearch/i", '%@^$', $strText);

now $strText contains a censored version of the input text

You can take it one step further by creating a file with a list of bad words, and then reading that into an array. Or you can just have an include file that contains an array of all the 'bad' words.

hth
p.

[reagent]

looks like phpbuilder has a similar filter

this part:
$badwords = array('ass', '@!#$');

$strText = '@!#$, this ass keeps harrassing me!';

should have had the 's' word instead of '@!#$' in both places.

p.

[Anon]

assuming that stristr() recognizes a space as a character.

$str = 'this ass keeps harrassing me!';

$BadWords = array(" ass ","badword2 "," badword3 ". ect, ect);

foreach($BadWords as $value){
(stristr($str,$value)) ? badwords() : nobadwords();
}

stristr() is immensly faster than eregi. Something near 1000 times faster. The drawback is that it would miss the word if it didn't have a space before it, as in

"ass lovers everywhere"

eregi would be slightly more reliable, but stristr() is noticably faster on large strings of text. Whenever you know exactly what you are looking for, and don't need to ferret out variable characters, stristr() should be considered over eregi.

[Anon]

Thanks for the info, I will look into that as I am all for speed!

"The drawback is that it would miss the word if it didn't have a space before it"

That can be solved by first adding a space before and after the search term or phrase that is being considered.