state-based-json-parser

A parser for parsing the next JSON value in a provided string. In contrast with regular parsers it is allowed to have extraneous characters following the JSON value. Or to put it differently: only as much valid JSON is parsed as is possible.

	var JSONParser = require("state-based-json-parser");

        // Parse a string containing a number and a literal
        // JSON.parse() throws an error on this string
        var inputString = "12.34, true";
        var parser = new JSONParser();
        var parseResult = parser.parse(inputString);
        console.log(parseResult);       // { value: 12.34, index: 5 }

        // Parse remaining string (skipping "," by adding 1 to the index)
        parseResult = parser.parse(inputString, parseResult.index + 1);
        console.log(parseResult);       // { value: true, index: 11 }

        // Parse a string containing an array with a number and a literal
        parseResult = parser.parse("[ 12.34, true ]");
        console.log(parseResult);       // { value: [ 12.34, true ], index: 15 }

Check the resulting field index to see which part was parsed correctly.

The field index can also be used to continu parsing as shown in the above example. The JSONParser.parse() function can therefore be called with an additional parameter from to specify from which position the parsing should start. This can be useful for repeatedly parsing the next JSON value in a (comma separated) list of JSON values without having to slice the string to be parsed.

The parser uses a state machine to keep track of the parsing process. No regular expressions or calls to eval are used to parse the string.

The JSON specification is followed very strictly. This means that for example "00" will be parsed for only the first character (result is { value: 0, index: 1 }) . The second zero is not parsed, since the JSON specification does not allow two consecutive zero's at the start of a number.

Also Unicode surrogates are checked for correctness. This means that if a low surrogate character is found, a high surrogate must follow. Otherwise the parser will return the error code (string) "MISSING_HIGH_SURROGATE".

In case a valid JSON value is present the result of the invocation of JSONParser.parse(<string>) has the following structure:

	{
		value: <value>,    // The JSON value parsed
		index: <index>     // The index inside the string where parsing stopped (further parsing failed)
	}

In case of an invalid JSON value, the following structure is the result:

	{
		value: undefined,
		index: <index>,           // The index inside the string where parsing failed
		errorCode: <errorCode>    // The error code (a string, see explanation below)
	}

	// The error code will be one of:
	//     "MISSING_VALUE"              If no valid value is present (partial parsing might have taken place, check {index} for location of parse failure).
	//     "MISSING_MEMBER"
	//     "INVALID_MEMBER_NAME"
	//     "MISSING_COLON"
	//     "INVALID_OBJECT"
	//     "INVALID_ARRAY"
	//     "INVALID_STRING"
	//     "INVALID_ESCAPE_CHAR"
	//     "INVALID_UNICODE_HEX_STRING"
	//     "MISSING_HIGH_SURROGATE"     If a unicode low surrogate (0xd800 - 0xdbff) is found, the following unicode should be a high surrogate (0xdc00 - 0xdfff).
	//     "INVALID_NUMBER"
	//     "INVALID_NUMBER_FRACTION"
	//     "INVALID_NUMBER_EXPONENT"

license

BSD-3-Clause