@hastom/fixed-point

Library for operating with fixed point decimals

It's fully based on native BigInt and does not have any fallbacks (or dependencies)
As fast as pure bigint math

Install

npm i @hastom/fixed-point

Usage

Definition

import { FP, FixedPoint } from '@hastom/fixed-point'

const a = new FixedPoint(1_000n, 3n) // means 1.000, base = 1000, precision = 3
const b = FP(1.212) // means 1.212, base 1212, precision = 3
const c = FP([90_09, 2]) // means 90.09, base 9090, precision = 2

Math

import { FP, FixedPoint } from '@hastom/fixed-point'

const d = a.add(b) // result 2.212, base = 2212, precision = 3
const e = a.sub(c) // result -89.09, base -8909, precision = 3
const f = b.mul(c) // result 109.189, base 109189, precision = 3
const g = c.div(b) // result 74.33, base 7433, precision = 2
const h = e.neg() // result 89.09, base 8909, precision = 3
const i = e.abs() // result 89.09, base 8909, precision = 3

As you can see, by default all maths keeps first arg precision.
This behavior can be modified by extending base FixedPoint class, as well as default precision, for parsing plain numeric types

import { FP, FixedPoint, parseNumeric } from '@hastom/fixed-point'

const BTC: FPParser = (n: FixedPoint | Numeric) => {
  if (n instanceof FixedPoint) {
    return n
  }
  return new _BTC(...parseNumeric(n, 8))
}

class _BTC extends FixedPoint {
  protected parser: FPParser = BTC

  protected precisionResolution: PrecisionResolution = 'max'
}

const a = BTC(1) // means 1.00000000, base 100000000, precision = 8
const b = BTC(0.00000010) // means 0.0000001, base 10, precision = 8
const c = a.add(b) // result 1.00000010, base 100000010, precision = 8
const d = a.div([0.025, 12]) // result 40.000000000000, base 40000000000000, precision = 12