The no-profanity package was born because I was searching for an easy-to-use profanity filter, but found only outdated packages with tons of PR's and issues.
But, I found a useful package I integrated into this package, namely @coffeeandfun/google-profanity-words which contains a list of profanity words.
The bad-words package, which this package aims to replace, has one major flaw. It loops through all the words in the list one by one, instead of incorporating it into a single regex. This makes it very slow. I compared the two packages by looping through 10.000 short strings, of which 1/3rd had a bad word in it. The results are as follows:
My code:
console.time("bad-words");
arr.forEach((item) => {
filter.clean(item);
});
console.timeEnd("bad-words");
console.time("no-profanity");
arr.forEach((item) => {
replaceProfanities(item);
});
console.timeEnd("no-profanity");The result, with 10.000 items:
bad-words: 37.993s
no-profanity: 241.211ms
As you can see this resulted in a massive speed increase. The no-profanity package is about 150 times faster than the bad-words package with this simple test.
Functionality wise they're the same, in fact, no-profanity supports almost the same options as bad-words. It's actually intended to be the replacement package for bad-words.
If you're using bad-words, you can easily migrate to no-profanity. The only thing you need to do is change how you import and create the package, and maybe rename a few properties, though there's backwards compatibilty in place for most of them.
Instead of importing the package like this:
const Filter = require('bad-words');
const filter = new Filter();
// more code
filter.clean('Don't be an ash0le');You can now do this:
const { replaceProfanities } = require('no-profanity');
// more code
replaceProfanities('Don't be an ash0le');When dealing with options, you don't set it at the creation of new Filter(), but instead pass it as an argument to any method, like replaceProfanities().
For example, if you want to replace the bad words with # instead of *, you can do this:
replaceProfanities({ testString: "Don't be an ash0le", options: { replacement: "#" });Pretty much everything is compatible, but either named differently or different to include.
Instead of creating a new filter, you can now just call methods directly.
const { replaceProfanities, isProfane } = require('no-profanity');Instead of calling clean() on the filter, you can now call replaceProfanities().
replaceProfanities("Don't be an ash0le");The metod isProfane() is no longer called from the filter object, but can be called directly.
isProfane("Don't be an ash0le");Instead of calling addWords() on the filter, you can now pass the words as an array to the replaceProfanities() method.
replaceProfanities({ testString: "Don't be an ash0le", options: { includes: ['be'] });Instead of calling removeWords() on the filter, you can now pass the words as an array to the replaceProfanities() method.
replaceProfanities({ testString: "Don't be an ash0le", options: { excludes: ['ash0le'] });EmptyList is still available, but now on the options object.
replaceProfanities({ testString: "Don't be an ash0le", options: { emptyList: true });The placeholder replacement is now called replacement.
replaceProfanities({ testString: "Don't be an ash0le", options: { replacement: "#" });