Worry less about maintaining a list of every possible spelling of every bad word.
Ensuring that text entered by a user is polite can be difficult. Generating memorable passwords automatically using letters and numbers, without creating an offensive combination is challenging as well.
This project attempts to help by taking a piece of text, and replacing symbols that are commonly used to hide bad words with the letters they may represent. The result can be scanned against a far simpler bad words file.
npm install @cityssm/unleet
unleet("b@d w0rd"); // => [ "bad word" ] unleet("1337 $P33K"); // => [ 'ieet zpeek', 'ieet speek', 'leet zpeek', 'leet speek' ] unleet("0rg@ni℠"); // => [ 'organizm', 'organism' ]
Found some text that doesn't unleet properly? Open an issue.
Pull requests with more bizarre Unicode-to-letter mappings are more than welcome!