Replacing and trimming messages

Question

I'm trying to toLowerCase, replace and then trim that replace so something like Hello [- the re _- wo rl d. will turn into a string with no spaces or punctuation like hellothereworld

I'm using this

    let msg = message.content.toLowerCase()
    let originalMessage = msg.split(" ");
    let removePunctuation = originalMessage.toString().replace(/[.,\/#!$%\^&\*;:{}=\-_`~()]/g,"");
    let checkMessage = removePunctuation.toString().trim();

then this to filter the message:

for (let i = 0; i < checkMessage.length; i++) {
    if(allowed.includes(checkMessage[i])) {
        continue;
    }
    if(curse.includes(checkMessage[i])) {
        const filterViolation = new Discord.RichEmbed()
            .setColor('#ff0000')
            .setAuthor(message.author.username, message.author.avatarURL)
            .setTitle('Filter Violation')
            .setDescription(`${rule} **Profanity.** \n${ifcont}\n${ifmistake}`)
            .setTimestamp()
            .setFooter(copyright);
        message.delete()
        message.author.send(filterViolation)
        message.channel.send(msgvio).then(msg => {msg.delete(15000)})
        logger.write(logviolation)
    }
//More code
}

^{Thanks to Tarazed for the help with the filter in general.}

But when I type any message that goes against the filter, nothing happens and no errors are thrown in the console. Any ideas on what I did wrong?

New Code:

let checkMessage = message.content.toLowerCase().replace(/[^\w ]/g,"");

    console.log(checkMessage);

    // let checkMessage = msg.split(" ");
    for (let i = 0; i < checkMessage.length; i++) {
    if(checkMessage.includes(allowed[i])) {
        continue;
    }
//--------------------------------------- CURSE
    if(checkMessage.includes(curse[i])) {
        const filterViolation = new Discord.RichEmbed()
            .setColor('#ff0000')
            .setAuthor(message.author.username, message.author.avatarURL)
            .setTitle('Filter Violation')
            .setDescription(`${rule} **Profanity.** \n${ifcont}\n${ifmistake}`)
            .setTimestamp()
            .setFooter(copyright);
        message.delete()
        message.author.send(filterViolation)
        message.channel.send(msgvio).then(msg => {msg.delete(15000)})
        logger.write(logviolation)
    }
}

It logs the check message correctly console.log(checkMessage); but it doesn't go through the filter, neither allowed or unallowed.

It logs the message correctly but the word violates the filter, but nothing is done.

New Code 2:

if(curse.some(word => checkMessage.includes(word) && !allowed.some(allow => allow.includes(word) && checkMessage.includes(allow)))) {
        const filterViolation = new Discord.RichEmbed()
            .setColor('#ff0000')
            .setAuthor(message.author.username, message.author.avatarURL)
            .setTitle('Filter Violation')
            .setDescription(`${rule} **Profanity.** \n${ifcont}\n${ifmistake}`)
            .setTimestamp()
            .setFooter(copyright);
        message.delete()
        message.author.send(filterViolation)
        message.channel.send(msgvio).then(msg => {msg.delete(15000)})
        logger.write(logviolation)
        return;
    }

Answer 1

Your current problem is that you're iterating over literally just a string like "test" and checking if each letter is a curse word.

That is, checkMessage[0] will be 't', checkMessage[1] will be 'e', etc. I'm guessing no single character will match anything in your curse or even allowed array.

You can get rid of the loop entirely and simply check if(curse.includes(checkMessage)), the entire message... but be wary this may easily return false positives. Heaven forbid someone sends a message like "it was a nice pic until i looked closer" and pic until put together triggers a certain c-word in your filter.

Anyway, I'd like to also point out that your code for stripping out spaces and punctuation does some rather strange things. I'll comment what I expect to be happening at each stage. Assume an input message of " Hello - world. "

let originalMessage = msg                    //" hello - world. "
                       .split(" ");          //["", "hello", "-", "world.", ""]
let removePunctuation = originalMessage
                         .toString()         //",hello,-,world.," (why even split?)
                         .replace(/[.,\/#!$%\^&\*;:{}=\-_`~()]/g,""); //"helloworld"
let checkMessage = removePunctuation
                    .toString() //"helloworld" (does nothing ever)
                    .trim();    //"helloworld" (does nothing ever)

You can achieve the same thing very easily by using the regex class \W which matches all non-word characters:

let checkMessage = message.content.toLowerCase().replace(/\W/g,""); //"helloworld"

---

It looks like there's still issues related to the loop. I suggest using Array#some to test if the message contains any curse words:

let checkMessage = message.content.toLowerCase().replace(/[^\w ]/g,"");

if(curse.some(word =>                //search through the curse words
     checkMessage.includes(word) &&  //if message has this curse word
     !allowed.some(allow =>          //and there's no allowed word which:
       allow.includes(word) &&       //1. contains this curse word
       checkMessage.includes(allow)  //2. is in the message
     )
  )) {
  //then send violation
}

Answer 2

Since checkMessage is a string, your for loop is looping over and testing individual characters.

For example:

let checkMessage = "hellothereworld";
for (let i = 0; i < checkMessage.length; i++) {
    console.log(checkMessage[i]);
}

Perhaps you'd rather keep the spaces in your string and operate on individual words:

let originalMessage = "Hello [-the world"


// Use regex that keeps spaces
let removePunctuation = originalMessage.toString().replace(/[^\w ]/g,"");
let checkMessage = removePunctuation.toString().trim();

// Split into words
let messageArray = checkMessage.split(' ');

// Loop over words
for (let i = 0; i < messageArray.length; i++) {
    console.log(messageArray[i]);
}