Shorten Javascript Function

Question

I wrote myself a function to turn a string into an abbreviation and it's currently fairly long and it's case-sensitive.

I need a way shorten it so it works 100% of the time. Currently it screws up if one of the splitting words has a capital, if a word ends in a splitting word.

My splitting words are basically the words I'm removing (as most companies and such don't include them). They include:

• and
• of
• the
• for
• to

Also, the way I'm removing them is using split and join (str.split('and ').join('')) which to me doesn't seem like the easiest way.

Other than these issues, it works fine. Could anyone help me shrink the function and fix the issues? Thanks.

Function:

String.prototype.toAbbrev = function () {
    var s = [];
    var a = this.split('and ').join('').split('of ').join('').split('the').join('').split('for ').join('').split('to ').join('').split(' ');
    for (var i = 1; i < a.length + 1; i++) {
        s.push(a[i - 1].charAt(0).toUpperCase());
    }

    return s.join('.');
}

Outputs on Tested Companies

The National Aeronautics and Space Administration           ->    N.A.S.A
The National Roads and Motorists' Association               ->    N.R.M.A
Royal Society for the Prevention of Cruelty to Animals      ->    R.S.P.C.A

Answer 1

I think an approach like this might work better:

var toAbbrev = function(str){
    return str.replace(/\b(?:and|of|the|for|to)(?: |$)/gi,''). // remove all occurances of ignored words
               split(' ').                                     // split into words by spaces
               map(function(x){                          
                   return x.charAt(0).toUpperCase();           // change each word into its first letter capitalized
               }).
               join('.');                                      // join with periods
};

and here's a breakdown of the regular expression:

/
    \b                    // word boundary
    (?:and|of|the|for|to) // non-capturing group. matches and/of/the/for/to
    (?: |$)               // non-capturing group. matches space or end of string
/gi                       // flags: g = global (match all), i = case-insensitive

And here's an alternative method that has a less complicated regular expression:

var toAbbrev = function(str){
    return str.split(' '). // split into words
               filter(function(x){
                   return !/^(?:and|of|the|for|to)$/i.test(x); // filter out excluded words
               }).
               map(function(x){
                    return x.charAt(0).toUpperCase(); // convert to first letter, captialized
               }).
               join('.'); // join with periods
};

And regex breakdown:

/
    ^                     // start of string
    (?:and|of|the|for|to) // non-capturing group. matches and/of/the/for/to
    $                     // end of string
/i                        // flags: i = case-insensitive

Answer 2

An even shorter one:

str.replace(/(and|of|the|for|to)( |$)/gi, "").replace(/(.).+?(\s|$)/g, "$1.");

To make sure it is capitalized you can do .toUpperCase at the end.

(.)     //selects the first character
.+      //matches the rest of the characters
  ?     //? indicates a lazy match
(\s|$)  //match a space or the end

$1.     //means "the first selected match plus a dot"

---

Let's make it into one Regex!

str.replace(/((and|of|the|for|to) )*(.).+?(\s|$)/ig, "$3.");

"Royal Society for the Prevention of Cruelty to Animals"
    .replace(/((and|of|the|for|to) )*(.).+?(\s|$)/ig, "$3.");
//R.S.P.C.A

"Josie and the Pussycats"
    .replace(/((and|of|the|for|to) )*(.).+?(\s|$)/ig, "$3.");
//J.P.

This should, in theory, cover all legit names. For names with preposition(s) at the end, you can technically do this:

.replace(/((and|of|the|for|to) )*(.).+?(\s|$)((and|of|the|for|to) ?)*/ig, "$3.")

But this is clearly longer than the one with two replaces and this defeats its purpose.