@ansi-tools/parser

Parser for ANSI escape sequences.

Features

• ✅ Supports CSI, OSC, DCS, ESC, APC, SOS, PM, etc.
• ✅ Handles 7-bit (\x1b or \u001b) and 8-bit (\u009b) introducers
• ✅ Handles octal (\033) and shorthand \e introducers (only escaped)
• ✅ Multiple string terminators (\x1b\\, \x07)
• ✅ Zero dependencies
• ✅ Separate optimized modules for raw and escaped input

Used by ansi.tools.

Installation

npm install @ansi-tools/parser

Usage

import { parse } from "@ansi-tools/parser";

const input = "\x1b[31mHello\x1b[0m World";

for (const code of parse(input)) {
  console.log(code);
}

There is a difference between escaped and unescaped input. Only with an escaped input string the raw input and the positions can be preserved in the tokens and control codes. See the example below for the default and the /escaped import.

The default and unescaped tokenization is roughly ~30% faster. Use this if you just need the control codes.

Examples

Default (raw/unescaped)

import { parse } from "@ansi-tools/parser";

parse(`\x1b[31mHello\x1b[0m`);

// result:
[
  {
    type: "CSI",
    pos: 0,
    raw: "\u001b[31m",
    command: "m",
    params: ["31"],
  },
  {
    type: "TEXT",
    pos: 5,
    raw: "Hello",
  },
  {
    type: "CSI",
    pos: 10,
    raw: "\u001b[0m",
    command: "m",
    params: ["0"],
  },
];

Escaped

import { parse } from "@ansi-tools/parser/escaped";

parse(String.raw`\x1b[31mHello\x1b[0m`);

// result:
[
  {
    type: "CSI",
    pos: 0,
    raw: "\\x1b[31m",
    command: "m",
    params: ["31"],
  },
  {
    type: "TEXT",
    pos: 8,
    raw: "Hello",
  },
  {
    type: "CSI",
    pos: 13,
    raw: "\\x1b[0m",
    command: "m",
    params: ["0"],
  },
];

Tokenizer & generators

The tokenizer and generators are also available, for both the default and the /escaped versions.

tokenize

import { tokenize } from "@ansi-tools/parser";

const input = "\x1b[31m";

for (const token of tokenize(input)) {
  console.log(token);
}

Generators

import { tokenizer, parser } from "@ansi-tools/parser";

const input = "\x1b[31mHello\x1b[0m";

const tokens = tokenizer(input);

const codes = parser(tokens);

for (const code of codes) {
  console.log(code);
}

Type Definitions

function tokenize(input: string): TOKEN[];
function parse(input: string): CODE[];

function* tokenizer(input: string): Generator<TOKEN>;
function* parser(tokens: Generator<TOKEN>): Generator<CODE>;

type TOKEN = {
  type: "INTRODUCER" | "DATA" | "FINAL" | "TEXT";
  pos: number;
  raw: string;
  code?: string;
  intermediate?: string;
};

type CONTROL_CODE = {
  type: "CSI" | "DCS" | "DEC" | "ESC" | "OSC" | "SGR" | "STRING" | "PRIVATE";
  command: string;
  raw: string;
  params: string[];
  pos: number;
};

type TEXT = {
  type: "TEXT";
  raw: string;
  pos: number;
};

type CODE = CONTROL_CODE | TEXT;

License

ISC