Highlights

• Small (~ 1.1 KB)
• Minimal runtime overhead
• Static type inference
• Single-pass data validation

Example

import { email, rangeLength } from "@the-minimal/validator";
import { encodeObject, encodeAscii8, encodeUint8, encodeTap } from "@the-minimal/protocol";

const encodeUserData = encodeObject([
  {
    key: "email",
    type: encodeTap(
      encodeAscii8,
      email
    )
  },
  {
    key: "age",
    type: encodeTap(
      encodeUint8,
      rangeValue(0, 120)
    )
  },
]);

const array = new Uint8Array(128);
const view = new DataView(array.buffer);
const state = {
  a: array,
  v: view,
  o: 0
};

encodeUserLogin(state, {
  email: "user@example.com",
  age: 26
});

return array.subarray(0, state.o);

State

Every encoder and decoder accepts State which looks like this:

type State = {
  // array
  a: Uint8Array;
  // view
  v: DataView;
  // offset
  o: number;
};

This library doesn't come with its own memory allocator.

It's up to you and your specific use-case what kind of memory allocator you use.

Modularity

This library is basically just a tiny opinionated wrapper around State.

We use either Uint8Array or DataView to modify or read buffer and then we bump the offset.

All functions are standalone encoders and decoders which take State as a parameter.

This makes it very easy to extend the library to your liking and compose our encoders and decoders with your custom ones.

API

Bool

Encoded as Uint8 with values 0 or 1.

encodeBool(state, true);
decodeBool(state);

Uint8

encodeUint8(state, 64);
decodeUint8(state);

Uint16

encodeUint16(state, 1024);
decodeUint16(state);

Uint32

encodeUint32(state, 128_000);
decodeUint32(state);

Int8

encodeInt8(state, 64);
decodeInt8(state);

Int16

encodeInt16(state, 1024);
decodeInt16(state);

Int32

encodeInt32(state, 128_000);
decodeInt32(state);

Float32

encodeFloat32(state, 3.16);
decodeFloat32(state);

Float64

encodeFloat64(state, 420_000.69);
decodeFloat64(state);

Ascii8

ASCII string of maximum length of 256 bytes (256 characters).

encodeAscii8(state, "Hello, World!");
decodeAscii8(state);

Ascii16

ASCII string of maximum length of 65536 bytes (65536 characters).

encodeAscii16(state, "Lorem ipsum dolor sit amet, ..");
decodeAscii16(state);

Unicode8

Unicode string of maximum length of 256 bytes (64-128 characters).

encodeUnicode8(state, "Dobré ráno, světe!");
decodeUnicode8(state);

Unicode16

ASCII string of maximum length of 65536 bytes (16384-32768 characters).

encodeUnicode16(state, "Oprávněné aniž i odstoupil o snadno osoby ..");
decodeUnicode16(state);

Array8

Array of maximum length of 256.

encodeArray8(encodeUint8)(state, [1, 16, 4, 8, 7]);
decodeArray8(encodeUint8)(state);

Array16

Array of maximum length of 65536.

encodeArray16(encodeUint8)(state, [1, 16, 4, 8, 7, ..]);
decodeArray16(encodeUint8)(state);

Object

All keys have to be defined, otherwise the buffer will be misaligned.

encodeObject([
  { key: "email", type: encodeAscii8 },
  { key: "password", type: encodeAscii8 },
])(state, {
  email: "yamiteru@icloud.com",
  password: "Test123456"
});

decodeObject([
  { key: "email", type: decodeAscii8 },
  { key: "password", type: decodeAscii8 },
])(state);

Tuple

The maximum length of tuple is 256.

encodeTuple([
  encodeUint8,
  encodeUint8,
  encodeAscii
])(state, [185, 90, "yamiteru"]);

decodeTuple([
  decodeUint8,
  decodeUint8,
  decodeAscii
])(state);

Enum

Maximum length of enum options is 256. Options are represented as uint8 indexes.

encodeEnum(["ADMIN", "USER"])(state, "USER");
decodeEnum(["ADMIN", "USER"])(state);

Nullable

Value in buffer is prefixed with uint8 0 if value is not null or 1 if value is null.

encodeNullable(encodeUint8)(state, 2);
decodeNullable(decodeUint8)(state);

Tap

This function is not encoded into buffer.

It's used for intercepting values.

Most notable use-case would be data validation while encoding or decoding.

The tap is executed before encoding and after decoding.

encodeTap(encodeUint8, () => { /* .. */ })(state, 2);
decodeTap(decodeUint8, () => { /* .. */ })(state);