Scans text for bigram, trigram...ngram frequency

ngram_scanner.js

const Tuple = (k, v) => ({
    key: k,
    val: v
})

const Ngram = n =>
({
    n,
    analyze: txt => {
        const cleanText = txt.toLowerCase().replace(/[^0-9a-z]/gi,"")
        let ngrams = []
        for (let i = 0; i <= txt.length - n; i++) {
            let thisNgram = cleanText.substring(i, i + n)
            const curNgram = ngrams.filter(n => n.key == thisNgram)
            
            if (curNgram.length > 0) {
                curNgram.map(n => n.val++)
            } else { 
                if (thisNgram.length > 0) {
                    newNgram = Tuple(thisNgram, 1)
                    ngrams.push(newNgram)
                }
            }
        }
        return ngrams.sort((a,b) => {
            if (a.val < b.val) {
                return 1
            } else if (a.val > b.val) {
                return -1
            }
            else {
                return 0
            }
        })
    }
})

const txt = [
    ["First", "A Partridge in a Pear Tree."],
    ["Second", "Two Turtle Doves, and"],
    ["Third", "Three French Hens,"],
    ["Fourth","Four Calling Birds,"],
    ["Fifth","Five Gold Rings,"],
    ["Sixth","Six Geese-a-Laying,"],
    ["Seventh", "Seven Swans-a-Swimming,"],
    ["Eighth","Eight Maids-a-Milking,"],
    ["Ninth","Nine Ladies Dancing,"],
    ["Tenth","Ten Lords-a-Leaping,"],
    ["Eleventh","Eleven Pipers Piping,"],
    ["Twelfth", "Twelve Drummers Drumming,"],
].join("")

const ns = Ngram(3).analyze(txt).filter(occurrence => occurrence.val > 2)
ns.forEach(n => console.log(`You can save ${(n.val * n.key.length) - n.val} bytes by substituting ${n.key}`))

This is very dirty, and will choke on large texts.
Should rewrite this sometime so it uses generators and promises and can perform in O(log n) time.