JavaScript Word Frequency

word-frequency.js

function wordFreq(sWords) {
    // make sure input looks right
    if (!sWords) return false;
    if (typeof sWords != 'string') return false;

    // converts to lowercase. trims leading and trailing spaces
    // removes all commas, semicolins, and periods
    // converts string to array of words, split by space
    sWords = sWords.toLowerCase().trim();
    sWords = sWords.replace(/[,;.]/g, '');
    sWords = sWords.split(/[\s\/]+/g);

    // wordFreq() will ignore these words:
    var ignore = [
        'and', 'the', 'to', 'a', 'of', 'for', 'as', 'i', 'with', 'it',
        'is', 'on', 'that', 'this', 'can', 'in', 'be', 'has', 'if'
    ];

    var wordfreqs = []
    for (i in sWords) {
        sWord = sWords[i];
        if (ignore.indexOf(sWord) < 0) {
            // if it's not an ignored word:
            wordindex = findWithAttr(wordfreqs, 'word', sWord);
            if (typeof wordindex == 'undefined') {
                // new word:
                wordfreqs.push({
                    word: sWord,
                    freq: 1
                });
            } else {
                // duplicate word:
                wordfreqs[wordindex].freq++;
            }
        }
    }

    return wordfreqs;
}

function findWithAttr(array, attr, value) {
    for (var i = 0; i < array.length; i += 1) {
        if (array[i][attr] === value) {
            return i;
        }
    }
}

function sortByKey(array, key, order) {
    switch (order) {
        case 'numasc': // numeric, ascending
            mysortfunc = function(x, y) {
                return ((x < y) ? -1 : ((x > y) ? 1 : 0));
            }
            break;
        default:
        case 'numdesc': // numeric, descending
            mysortfunc = function(x, y) {
                return ((x > y) ? -1 : ((x < y) ? 1 : 0))
            }
            break;
    }
    return array.sort(function(a, b) {
        var x = a[key];
        var y = b[key];
        return mysortfunc(x, y);
    });
}

function getTopTenWordsOnPage() {
    var doctext = document.body.innerText;
    wordfreqs = wordFreq(doctext);
    wordfreqs = sortByKey(wordfreqs, 'freq', 'numdesc');
    wordfreqs = wordfreqs.slice(0, 10);
    topten = [];
    for (i in wordfreqs) {
        topten.push([wordfreqs[i].freq, wordfreqs[i].word])
    }
    return topten;
}

topten = getTopTenWordsOnPage();
alert(topten.join('\n'));