Less painful binary encoding / decoding
Allows declarative definition for complex structures (like in C/Rust) via composable primitives, which enables encoding these structures into bytes and parsing back.
Comes with an separate debugger package.
npm install micro-packed
import * as P from 'micro-packed';
let other = P.struct({ a: U16BE, b: U16LE });
let s = P.struct({
field1: P.U32BE,
// strings, bytes, prefix and array first arg is length. It can be:
// - dynamic: via CoderType<number>. First U8 in this case if length of elemnt
field2: P.string(P.U8),
// - fixed length, reads 32 bytes, no prefix
field3: P.bytes(32),
// NOTE: array uses prefix as element count, not byte count!
field4: P.array(P.U16BE, P.struct({subField1: P.U64BE, subField2: P.string(10) }))
// - string to access previous fields in structure
field5: P.array('field1', P.U8), // Array of size 'field1' of U8
// Use sub-structure
field6: other,
field7: P.string(null),
// - null -- read until buffer exhausted
field8: P.array(null, P.U64BE),
})
Probably most powerful building block
import * as P from 'micro-packed';
let a1 = P.array(P.U16BE, child); // Dynamic size array (prefixed with P.U16BE number of array length)
let a2 = P.array(4, child); // Fixed size array
let a3 = P.array(null, child); // Unknown size array, will parse until end of buffer
let a4 = P.array(new Uint8Array([0]), child); // zero-terminated array (NOTE: terminator can be any buffer)
Same as array of bytes, should be a bit faster than generic implementation, also returns Uint8Array, instead of array of ints
import * as P from 'micro-packed';
// same as
let bytes = (len) => P.array(len, P.U8);
const b1 = bytes(P.U16BE);
const b2 = bytes(P.U16BE, true); // bytes in little-endian order
Same as bytes, but returns utf8 decoded string
import * as P from 'micro-packed';
const s = P.string(P.U16BE);
s.decode(new Uint8Array([116, 101, 115, 116])); // -> test
const s2 = P.cstring; // NUL-terminated strings
s.decode(new Uint8Array([116, 101, 115, 116, 0])); // -> test
Same as struct, but without fields names
import * as P from 'micro-packed';
let s = P.tuple([P.U32BE, P.U8, P.bytes(32), ...])
Like enum in C (but without iota).
Allows to map encoded values to string
import * as P from 'micro-packed';
let s = P.map(P.U8, {
name1: 1,
name2: 2,
});
s.decode(new Uint8Array([0x01])); // 'name1'
s.decode(new Uint8Array([0x02])); // 'name2'
s.decode(new Uint8Array([0x00])); // Error!
Like enum in Rust.
Allows to choice stucture based on some value. Depending on value of first byte, it will be decoded as array, string or number.
import * as P from 'micro-packed';
let s = P.tag(P.U8, {
0x1: P.array(u16, ...),
0x2: P.string(u16, ...),
0x3: P.U32BE,
})
Encodes some constant value into bytes and checks if it is the same on decoding.
import * as P from 'micro-packed';
let s = P.magic(U8, 123);
s.encode(); // Uint8Array([123])
s.decode(new Uint8Array([123])); // ok
s.decode(new Uint8Array([124])); // error!
Allows to parse bit-level elements:
import * as P from 'micro-packed';
// NOTE: structure should parse whole amount of bytes before it can start parsing byte-level elements.
let s = P.struct({ magic: P.bits(1), version: P.bits(1), tag: P.bits(4), len: P.bits(2) });
Encodes element as offset into real bytes
import * as P from 'micro-packed';
const s = P.pointer(P.U8, P.U8);
s.encode(123); // new Uint8Array([1, 123]), first byte is offset position of real value
Allows to pad value with zero bytes. Optional argument allows to generate padding value based on position.
import * as P from 'micro-packed';
P.padLeft(3, U8).encode(123); // Uint8Array([0, 0, 123])
P.padRight(3, U8).encode(123); // Uint8Array([123, 0, 0])
Decodes as true if the value is the same.
import * as P from 'micro-packed';
const s = P.flag(new Uint8Array([1, 2, 3]));
Decodes / encodes struct only when flag/bool value (described as path in structure) is true (conditional encoding).
import * as P from 'micro-packed';
const s = P.struct({ f: P.flag(new Uint8Array([0x0, 0x1])), f2: P.flagged('f', P.U32BE) });
Decodes/encodes value only if prefixed flag is true (or encodes default value).
import * as P from 'micro-packed';
const s = P.optional(P.bool, P.U32BE, 123);
Allows definition of circular structures
import * as P from 'micro-packed';
type Tree = { name: string; childs: Tree[] };
const tree = P.struct({
name: P.cstring,
childs: P.array(
P.U16BE,
P.lazy((): P.CoderType<Tree> => tree)
),
});
Converts array (key, value) tuples to dict/object/hashmap:
import * as P from 'micro-packed';
const dict: P.CoderType<Record<string, number>> = P.apply(
P.array(P.U16BE, P.tuple([P.cstring, P.U32LE] as const)),
P.coders.dict()
);
Validation of value before encoding and after decoding:
import * as P from 'micro-packed';
const val = (n: number) => {
if (n > 10) throw new Error(`${n} > 10`);
return n;
};
const RangedInt = P.validate(P.U32LE, val); // will check in both encoding and decoding
Easy debug (via console.log), just wrap specific coder for it:
import * as P from 'micro-packed';
const debugInt = P.debug(P.U32LE); // Will print info to console
There is: bool, U8, U[16|32|64|128|256][le|be] Other numeric types can be created via
import * as P from 'micro-packed';
const U32LE = P.int(4, true); // up to 6 bytes (48 bits)
const I256LE = P.bigint(32, true, true); // no limits
MIT (c) Paul Miller (https://paulmillr.com), see LICENSE file.