Just Another Lexer is a jison-compatible lexical analyzer with streaming capability.

Create an instance of Jalex and add rules to watch for on the input.

var Jalex = require('jalex');
var lex = new Jalex();

// match tokens
lex.addRule(/[A-Za-z_]\w+/, (match) => {
    return 'ID';
});

// ignore whitespace
lex.addRule(/\s+/, () => { });

// interact using Node Streaming interface
lex.pipe(process.stdout);
lex.write("  token");

Jalex will execute the handler for the longest matching rule.

The addRule function accepts a regular expression and a handler function. Both arguments are required.

The regular expression matches are spread to the arguments of the handler. So the first argument is the entire match, the second argument is the first subgroup, and so forth.

lex.addRule(/\s+(\w+)/, (match, name) => {
    console.log(name); // "word"
    return 'NAME';
});
lex.write("   word");

Jalex can be used as a custom lexer for Jison.

var Jalex = require('jalex');
var lex = new Jalex();

lex.addRule(/[A-Za-z_]\w+/, (match) => {
    return 'ID';
});
lex.addRule(/\s+/, () => { });

var Parser = require('jison').Parser;
var parser = new Parser({
    bnf: {
        S: ["ID EOF", "return 'token';"]
    }
});

// set Jalex as the lexer for Jison
parser.lexer = lex;
var result = parser.parse("   token");

The this argument of the addRule callback will be the current instance of the lexer.

The yytext property can be set on the lexer to forward a value to Jison.

lex.addRule(/\w+/, function(match) {
    this.yytext = match;
    return 'ID';
});

var parser = new Parser({
    bnf: {
        S: ["ID EOF", "return yytext;"]
    }
});

Be sure to set yytext via the this argument as Jison creates a cloned instance of Jalex during parsing.

As the lexer consumes input, the yyloc object on the Jalex instance will be updated to indicate the position of the token. It has the following values to assist with source mapping:

• first_index the index in the input file where the current match starts
• last_index the index in the input file where the current match ends
• first_line the line on which the current match begins
• last_line the line on which the current match ends
• first_column the index on the first line where the current match begins
• last_column the index on the last line where the current match ends

The yyloc values should not be altered as the values are incremental rather than absolute.

Jalex responds to and returns for some special tokens: EOF and REJECT

Jalex returns the EOF token when it reaches the end of the input.

If the longest matching rule returns a REJECT token then Jalex will use the next longest matching rule and continue in that fashion until a non-REJECT token is returned by a handler, nothing is returned (empty rule), or no more rules match the input. In the last case lexing continues at the next input character.

var Jalex = require("../jalex");
var lex = new Jalex();

lex.addRule(/\s+/, () => { });
lex.addRule(/\w+/, function(id) {
    return 'ID';
});

// look for 'simple example' but reject to allow 'simple' and 'example'
// to match as IDs
lex.addRule(/simple example/, () => {
    console.log("\n\'simple example\' found.");
    return 'REJECT';
});

lex.pipe(process.stdout);
lex.write("a simple example word capture");