compact-encoding-maybe-enum

Generate compact encoding for defined enum or limited length arbitrary string

why

An example of where this might be useful is encoding the value of a type property, where a set of strings are expected to be used 90% of the time, so encoding them as strings is quite wasteful, but you don't want to prohibit arbitrary strings either.

The trade-off used here is to:

• limit the arbitrary string length to 127 character.
• limit the enum to 128 member.

This way, we only need to store a value between 0 - 127 in the first byte, while the msb indicates whether the string is encoded as an enum or a string .

How it works

0000 0000 => 0111 1111 values encode the index of the enum member.

1000 0000 => 1111 1111 values encode the length of the arbitrary string, after switching the msb to 0 (xor 0x80).

Installation

npm install compact-encoding-maybe-enum

Usage

import c from 'compact-encoding';
import { maybeEnum } from 'compact-encoding-maybe-enum';

const enc = maybeEnum(['foo', 'bar']);

// Encoding decoding a member of enum
const encoded = c.encode(enc, 'bar');
//=> <Buffer 0x01>
const decoded = c.decode(enc, encoded);
//=> 'bar'

// Encoding decoding arbitrary string
const encoded = c.encode(enc, 'x');
//=> <Buffer 81 78>
const decoded = c.decode(enc, encoded);
//=> 'x'

API

const enc = maybeEnum(members)

members should be an array of unique strings (set) with maximum length of 0x80 (128) members.

with compact encoding: c.encode(enc, value)

value should be a string from members, or an arbitrary string with maximum length of 0x7f (127) characters.