src/core/priors.js
Prior distributions as unit-cube transforms (the form nested sampling needs): each maps u ~ U(0,1) to a draw from the prior via its inverse CDF.
Import: import { … } from './src/core/priors.js' — no DOM, runs in Node and the browser.
PRIOR_TYPES
js
const PRIOR_TYPESThe fitting UI offers these two; makeTransform also understands 'truncnormal', 'loguniform' and 'categorical' ({ values: [...], weights?: [...] }, a discrete parameter such as a period alias) for programmatic use.
makeTransform
js
function makeTransform(pr)text
Build u -> x for one prior spec { type, a, b, lo, hi }.validatePrior
js
function validatePrior(pr)Returns an error string, or null if the prior is usable.
priorPdf
js
function priorPdf(pr, x)Prior density (unnormalised is fine) for drawing the prior curve in the UI.