Significantly improved isValid performance.

Older implementation relied on iterating over each character of id(M)
and every time searching for that character in alphabet(N).
O(M * N), which seemed not too fast.

Rewritten it to use regular expressions, which boosted performance.

Here is the benchmark - https://jsperf.com/regex-vs-indexof-in-loop-2312
Used 1000 different generated ids, half of them were incorrect.

On average, the current algorithm would score ~1.3k op/sec
On the other hand, the new one makes around ~12.5k op/sec

Which I've found significant enough to make a PR.

lib/is-valid.js

@@ -6,14 +6,8 @@ function isShortId(id) {
         return false;
     }
 
-    var characters = alphabet.characters();
-    var len = id.length;
-    for(var i = 0; i < len;i++) {
-        if (characters.indexOf(id[i]) === -1) {
-            return false;
-        }
-    }
-    return true;
+    var nonAlphabetic = new RegExp('[^' + alphabet.characters() + ']+', 'g');
+    return !nonAlphabetic.test(id);
 }
 
 module.exports = isShortId;

test/is-valid.test.js

@@ -16,6 +16,7 @@ describe('testing is-valid', function() {
 
     it('should find these invalid', function(done) {
         expect(isValid('i have spaces')).to.equal(false);
+        expect(isValid('i have \n breaks \n of \n the \n lines')).to.equal(false);
         expect(isValid('abc')).to.equal(false);
         expect(isValid(1234)).to.equal(false);
         expect(isValid()).to.equal(false);