Calculate length string based on regex using javascript / jquery

Question

Suppose I have an arbitrary regular expression. How I could calculate the length of string required for a match?

Examples (regex => minimum length of matchable string):

1. [0-9]{3},[0-9]{2} => 6
2. [0-9]{4},[0-9]{2} => 7
3. [0-9]{2}.[0-9]{3}.[0-9]{3}/[0-9]{4}-[0-9]{2} => 17
4. [0-9]{3}.[0-9]{3}.[0-9]{3}-[0-9]{2} => 14
5. [0-9]{2}/[A-Z]{2}/[0-9]{4} => 10

I also need a function which take as parameter a regex and a integer number between 1 and the size calculated with the function above (like position(regex, number)), and return what the type of the character in that position (number, letter or symbol).

Examples:

• Example 1: Position 3 is a "number"
• Example 2: Position 3 is a "symbol"
• Example 5: Position 4 is a "letter"

UPDATE

the objective here is implement this:

function size_of(regex) {
    //
}

function type_of(regex, posicao) {
    //
}

function generate_string(tamanho) {
    //
}

$(document).on('.valida', 'focus', function(){
    var regex = $(this).attr('pattern');

    var counter = 0;
    var tam = size_of(regex);
    var str = generate_string(tam);

    $(this).val(str);
    $(this).keypress(function(event){
        var tecla = e.which;

        if(typeof tecla == type_of(regex, counter)){
            str = str + tecla;
            counter++;
        }

        $(this).val(str);
    });
});

UPDATE 2

some examples that would be useful:

1-> calculate the lengh: http://js.do/code/38693 (just need be more generic).

UPDATE 3 - FINAL CODE

the final code for the script above is that:

jsfiddle

http://jsfiddle.net/klebermo/f8U4c/78/

code

function parse(regexString){
    var regex = /((?!\[|\{).(?!\]|\}))|(?:\[([^\]]+)\]\{(\d+)\})/g,
        match,
        model = [];
    while (match = regex.exec(regexString)) {
        if(typeof match[1] == 'undefined'){
            for(var i=0;i<match[3];i++){
                model.push(match[2]);
            }
        }else{
            model.push(match[1]);
        }
    }
    return model;
}

function replaceAt(s, n, t) {
    return s.substring(0, n) + t + s.substring(n + 1);
}

function size_of(regex) {
    var parsedRegexp = parse(regex);
    return parsedRegexp.length;
}

function type_of(regex, posicao) {
    var parsedRegexp = parse(regex);
    var pos = parsedRegexp[posicao];

    if(pos == '0-9')
        return 'number';

    if(pos == 'A-Z' || pos == 'a-z')
        return 'string';

    return pos;
}

function generate_string(regex, tamanho) {
    var str = '';

    for(var i=0; i<tamanho; i++) {
        var type = type_of(regex, i);
        if(type == 'number' || type == 'string')
            str = str + '_';
        else
            str = str + type;
    }

    return str;
}

var counter;
var tam;
var str;
var regex;

$('.valida').each(function(){

    $(this).on('focus', function(e){
        regex = $(this).attr('pattern');

        counter = 0;
        tam = size_of(regex);
        str = generate_string(regex, tam);

        $(this).val(str);
    });

    $(this).on('keypress', function(e){
        e.preventDefault();

        var tecla = e.which;

        if(tecla >= 48 && tecla <= 57)
            var tecla2 = tecla - 48;
        else
            var tecla2 = String.fromCharCode(tecla);

        result = $("<div>");
        result.append( "tecla = "+tecla+"<br>" );

        var t = type_of(regex, counter);

        if(counter < tam) {
            if(t != 'number' && t != 'string') {
                str = replaceAt(str, counter, t);
                counter++;
            }

            t = type_of(regex, counter);

            if(typeof tecla2 == t) {
                result.append( "tecla2 = "+tecla2+"<br>" );
                str = replaceAt(str, counter, tecla2);
                counter++;
            }
        }

        result.append( "counter = "+counter+"<br>" );
        $("#result").empty().append(result);

        $(this).val(str);
    });

});

Answer 1

I've made a little parser for simple regex like the ones you're using.

It basically creates an array for each expected character with the type of character (0-9, A-Z) or the character itself.

function parse(regexString){
    var regex = /((?!\[|\{).(?!\]|\}))|(?:\[([^\]]+)\]\{(\d+)\})/g,
        match,
        model = [];
    while (match = regex.exec(regexString)) {
        if(typeof match[1] == 'undefined'){
            for(var i=0;i<match[3];i++){
                model.push(match[2]);
            }
        }else{
            model.push(match[1]);
        }
    }
    return model;
}

And jsfiddle to demo.

About the regex used inside the parse method, a debuggex schema will explain it better than i could do :

((?!\[|\{).(?!\]|\}))|(?:\[([^\]]+)\]\{(\d+)\})

Also, you can get total number of characters through :

myresult.length;

And the type of the n-th character through :

myresult[n];

Answer 2

I believe a generic solution to this problem would involve implementing a function that generates a finite state automaton object corresponding to each regex.

This SO post seems related to the question at hand.

Also check out this link: (C# Code to generate strings that match a regex)

Answer 3

In your specific case, you could try this code (demo):

var basicRegexLength = function(regex){
    var i;
    regex = regex.replace(/(\[0-9\]|\[A-Z\])/gi, '');
    for (i = 1; i < 10; i++) {
        regex = regex.replace( new RegExp('\\{' + i + '\\}', 'g'), Array(i+1).join('.') );
    }
    return regex.length;
};