This package provides serialization/deserialization utilities for ArrayBuffer. It is useful for implementing encoder/decoder for you data structure.

Serial Model

Serializer and Deserializer work on Builder model. It has these primitives below:

• i8/u8
• i16/u16
• i32/u32
• i64/u64

And also serializing string as UTF-8 sequence is supported by encUtf8/decUtf8, as a composite builder.

Their features are provided as each monad, monadForCodeM and monadForDecoder respectively. So you can combine them with using these monads and CatT system.

Encoder

An Encoder<T> denotes that a function transforming data T to a Code, which writes bytes to an ArrayBuffer through DataView.

There are utility functions to encode primitive data:

• Basics
  • encUnit - Does nothing
  • encSum - Useful to encode a sum type
• Numerics
  • encI8/encU8
  • encI16Be / encI16Le / encU16Be / encU16Le
  • encI32Be / encI32Le / encU32Be / encU32Le
  • encI64Be / encI64Le / encU64Be / encU64Le
  • encF32Be / encF32Le
  • encF64Be / encF64Le
• String
  • encUtf8
• And internal others exported

Also you can compose and customize encoders by CodeM's monad to create your own encoder like this:

import { type Encoder, encU32Be, encUtf8, monadForCodeM, runCode } from "./serial.js";
import { doT } from "./cat.js";

type MyType = {
  name: string;
  age: number;
};
const myEncoder: Encoder<MyType> = (data: MyType) =>
   doT(monadForCodeM)
     .run(encUtf8(data.name))
     .finishM(() => encU32Be(data.age));
const serial = runCode(myEncoder({ name: "Shigure Ui", age: 17 }));
Copy

Note that an Encoder serializes lazily. You need pass a Code to runCode to evaluate it.

Decoder

A Decoder<T> denotes that a function parsing Code to data T.

There are utility functions to decode primitive data:

• Basics
  • decSum - Useful to decode a sum type
  • failDecoder - Useful to report decoding error
• Numerics
  • decI8/encU8
  • decI16Be / decI16Le / decU16Be / decU16Le
  • decI32Be / decI32Le / decU32Be / decU32Le
  • decI64Be / decI64Le / decU64Be / decU64Le
  • decF32Be / decF32Le
  • decF64Be / decF64Le
• String
  • decUtf8
• And internal others exported

Unlike encoders, they are needed to invoke to get a Decoder by an internal implementation reason.

Also you can compose and customize decoders by Decoder's monad to create your own decoder like this:

import { type Decoder, decU32Be, decUtf8, failDecoder, monadForDecoder, runDecoder } from "./serial.js";
import { doT } from "./cat.js";

type MyType = {
  name: string;
  age: number;
};
const myDecoder: Decoder<MyType> =
  doT(monadForDecoder)
    .addM("name", decUtf8())
    .addM("age", decU32Be())
    .when(
      ({ age }) => age < 0,
      ({ age }) => failDecoder(`expected non-negative age, but got ${age}`),
    )
    .finish(({ name, age }) => ({ name, age }));
const decode = (binary: ArrayBuffer) => {
  const result = runDecoder(myDecoder)(binary);
  // ...
};
Copy

Note that you need to match the process orders of your encoder and decoder, because it reads/writes a binary sequence by written order. If you consider the version of your data, you may use Envelope to append a version.