what is the format of this data? is it a custom format?

Question

I get this data as an ajax response:

{
    "idArray" = (
        "99516",
        "99518",
        "97344",
        "97345",
        "98425"
    );
    "frame" = {
        "size" = {
            "width" = "8";
            "height" = "8";
        };
        "origin" = {
            "x" = "244";
            "y" = "345";
        };
    };
},

This is just a portion of the Data, but it continues in the same format. I don't have access to the source of the files that generate this data.

Is this a known format or something custom?

Answer 1

Since people tend to throw regular expressions at everything, even things that can not be parsed with regular expressions (ie. non-regular languages): I've written a proof-of-concept parser for this data format:

$input = '{
    "idArray" = (
        "99516",
        "99518",
        "97344",
        "97345",
        "98425"
    );
    "frame" = {
        "size" = {
            "width" = "8";
            "height" = "8";
        };
        "origin" = {
            "x" = "244";
            "y" = "345";
        };
    };
}';

echo json_encode(parse($input));

function parse($input) {
    $tokens = tokenize($input);
    $index = 0;
    $result = parse_value($tokens, $index);
    if ($result[1] !== count($tokens)) {
        throw new Exception("parsing stopped at token " . $result[1] . " but there is more input");
    }
    return $result[0][1];
}

function tokenize($input) {
    $tokens = array();
    $length = strlen($input);
    $pos = 0;
    while($pos < $length) {
        list($token, $pos) = find_token($input, $pos);
        $tokens[] = $token;
    }
    return $tokens;
}

function find_token($input, $pos) {
    $static_tokens = array("=", "{", "}", "(", ")", ";", ",");
    while(preg_match("/\s/mis", substr($input, $pos, 1))) { // eat whitespace
        $pos += 1;
    }
    foreach ($static_tokens as $static_token) {
        if (substr($input, $pos, strlen($static_token)) === $static_token) {
            return array($static_token, $pos + strlen($static_token));
        }
    }
    if (substr($input, $pos, 1) === '"') {
        $length = strlen($input);
        $token_length = 1;
        while ($pos + $token_length < $length) {
            if (substr($input, $pos + $token_length, 1) === '"') {
                return array(array("value", substr($input, $pos + 1, $token_length - 1)), $pos + $token_length + 1);
            }
            $token_length += 1;
        }
    }
    throw new Exception("invalid input at " . $pos . ": `" . substr($input, $pos - 10, 20) . "`");
}

// value is either an object {}, an array (), or a literal ""
function parse_value($tokens, $index) {
    if ($tokens[$index] === "{") {  // object: a list of key-value pairs, glued together by ";"
        $return_value = array();
        $index += 1;
        while ($tokens[$index] !== "}") {
            list($key, $value, $index) = parse_key_value($tokens, $index);
            $return_value[$key] = $value[1];
            if ($tokens[$index] !== ";") {
                throw new Exception("Unexpected: " . print_r($tokens[$index], true));
            }
            $index += 1;
        }
        return array(array("object", $return_value), $index + 1);
    }
    if ($tokens[$index] === "(") {  // array: a list of values, glued together by ",", the last "," is optional
        $return_value = array();
        $index += 1;
        while ($tokens[$index] !== ")") {
            list($value, $index) = parse_value($tokens, $index);
            $return_value[] = $value[1];
            if ($tokens[$index] === ",") {  // last, is optional
                $index += 1;
            } else {
                if ($tokens[$index] !== ")") {
                    throw new Exception("Unexpected: " . print_r($tokens[$index], true));
                }
                return array(array("array", $return_value), $index + 1);
            }
        }
        return array(array("array", $return_value), $index + 1);
    }
    if ($tokens[$index][0] === "value") {
        return array(array("string", $tokens[$index][1]), $index + 1);
    }
    throw new Exception("Unexpected: " . print_r($tokens[$index], true));
}

// find a key (string) followed by '=' followed by a value (any value)
function parse_key_value($tokens, $index) {
    list($key, $index) = parse_value($tokens, $index);
    if ($key[0] !== "string") { // key must be a string
        throw new Exception("Unexpected: " . print_r($key, true));
    }
    if ($tokens[$index] !== "=" ) {
        throw new Exception("'=' expected");
    }
    $index += 1;
    list($value, $index) = parse_value($tokens, $index);
    return array($key[1], $value, $index);
}

The output is:

{"idArray":["99516","99518","97344","97345","98425"],"frame":{"size":{"width":"8","height":"8"},"origin":{"x":"244","y":"345"}}}

Notes

• the original input has a trailing ,. I've removed that character. It will throw an error (more input) if you put it back.

• This parser is naive in the sense that it tokenizes all input before it starts parsing. This is not good for large input.

• I've not added escape detection for strings in the tokenizer. Like: "foo\"bar".

This was a fun exercise. If you have any questions let me know.

Edit: I see this is a JavaScript question. Porting the PHP to JavaScript shouldn't be too hard. The list($foo, $bar) = func() is equivalent to: var res = func(); var foo = res[0]; var bar = res[1];

Answer 2

Try using this function with the response text as a parameter:

function getJsonData(str){
    str = str.replace(/,/g, '')         //remove ,
             .replace(/\(/g, '[')       //replace (
             .replace(/\[/g)', ']')     //replace )
             .replace(/;/g, ',')        //replace ; 
             .replace(/=/g, ':');       //replace :
    return JSON.parse(str);
}

This is an edit made by @SamSal

function getJsonData(str){
    str = str.replace(/\(/g, '[')       //replace (
         .replace(/\)/g, ']')       //replace )
         .replace(/;\n\s+}/g, '}')  //replace ;} with }
         .replace(/;/g, ',')        //replace remaining ; with ,  
         .replace(/=/g, ':');       //replace :
    return JSON.parse(str);
}

Answer 3

Is this a known format or something custom?

It is a custom format, that looks a little bit like JSON without actually being JSON.