@kreablo/k-lexer v1.0.2

K-Lexer

This lexer tokenize a character stream by applying a regular expression composed of fragments on the form

A matching fragment is associated with a token and a context transfer. Thus it is possible to have multiple contexts with different sets of tokens enabled.

API Reference

The class KLexer implements the Lexer interface from typescript-parsec. The parameter to the constructor is an array of lexer context (KLexerContext). The type parameter T is the type of the token kind, typically an enum.

const lexer = new KLexer<T>(contexts: KLexerContext<T>[]);

The first lexer context in the array is used as the initial context.

the parse(input: string) method initates the lexer and returns the first token in the stream. The following token, if any, is obtained by accessing the next property of the token.

The class KLexerContext represent a lexer context.

const context1 = new KLexerContext<T>(
   tokenGenerators: [string, ((character: string) => [T, Context, boolean])][],
   public readonly errormessage: string,
   regexpFlags: string | undefined
);

The parameter tokenGenerators is an array of tuples, each tuple consisting of a regular expression fragment and a function that produce a 3-tuple. The 3-tuple consists of the token kind, the next context, and a boolean flag to indicate if the token should be included in the token stream. The regexp fragments must not contain grouping parenthesis. Non-grouping parenthesis may be used.

The error message is used as a parameter to the exception that is thrown if none of the regular expression fragments matches the current position in the input.

The regexp flags are added to the regular expression in addition to 'ys'. The flags must not conflict with the operation of the lexer. 'i', 'm', 'u' and 'v' is ok to use but not 'g' and 'd'.

Example: double quoted string with variable interpolation:

import { KLexerContext, KLexer } from 'k-lexer';

enum TokenType {
    SCALAR,
    ASSIGNMENT,
    DOUBLE_QUOTE,
    STRING_CONTENT,
    WHITESPACE
};

enum Context {
    TOP,
    DOUBLE_QUOTED_STRING
};

type ctx = [Context, KLexerContext<TokenType>];

const _frag: (regexp: string, tokentype: TokenType, keep: boolean) => (nextContext: Context) => [string, (character: string) => [TokenType, Context, boolean]] =
    (regexp, tokentype, keep) => (nextContext) => [regexp, (_) => [tokentype, nextContext, keep]];

const scalar         = _frag('\\$[a-zA-Z_0-9]+', TokenType.SCALAR, true);
const assignment     = _frag('=', TokenType.ASSIGNMENT, true);
const double_quote   = _frag('"', TokenType.DOUBLE_QUOTE, true);
const string_content = _frag('(?:[^"$\\\\]|\\\\[\\\\$"])+', TokenType.STRING_CONTENT, true);
const whitespace     = _frag('[\\t\\s\\n\\r]+', TokenType.WHITESPACE, false);

const top_context: ctx = [Context.TOP, new KLexerContext(
    [
        scalar(Context.TOP),
        assignment(Context.TOP),
        double_quote(Context.DOUBLE_QUOTED_STRING),
        whitespace(Context.TOP)
    ], 'failed-lexing-position-context', '')
];

const double_quoted_string: ctx = [Context.DOUBLE_QUOTED_STRING, new KLexerContext(
    [
        string_content(Context.DOUBLE_QUOTED_STRING),
        scalar(Context.DOUBLE_QUOTED_STRING),
        double_quote(Context.TOP),
    ], 'failed-double-quoted', '')
];

const lexer = new KLexer([top_context, double_quoted_string]);

let t = lexer.parse('  $hello = "hello, \\"$name\\"" ');

while (t !== undefined) {
    switch (t.kind) {
        case TokenType.SCALAR:
            console.log("SCALAR: " + t.text);
            break;
        case TokenType.ASSIGNMENT:
            console.log("ASSIGNMENT");
            break;
        case TokenType.DOUBLE_QUOTE:
            console.log("DOUBLE_QUOTE");
            break;
        case TokenType.STRING_CONTENT:
            console.log("STRING_CONTENT: '" + t.text + "'");
            break;
        default:
            throw new Error("Invalid token: " + t.kind);
    }
     t = t.next;
}

The output will be:

SCALAR: $hello
ASSIGNMENT
DOUBLE_QUOTE
STRING_CONTENT: 'hello, \"'
SCALAR: $name
STRING_CONTENT: '\"'
DOUBLE_QUOTE