# quantile.js

``````import { curry, sort, length, filter} from "ramda"

import isNumeric from "./internal/isNumeric"

/**
* Quantile of a value in a series of series using midpoint interpolation. see https://numpy.org/doc/stable/reference/generated/numpy.percentile.html
*
* @func
* @memberOf Z
* @category Analysis
* @param {Number} quantile to calculate as represented by a number from 0 to 1
* @param {Array} array of values to calculate percentile from
* @return {Number}
* @example
*
* const series = [7, 2, 30, 56, 75, 150]
* Z.quantile(.95, series)
* // 112.5
*/
const quantile = curry((q, arr ) => {

const sortedArr = sort(diff, filter(isNumeric, arr))

// if requesting 0 or 1 do the easy thing.
if (q == 0) {
return sortedArr[0];
}
else if (q == 1) {
return sortedArr[length(sortedArr) - 1];
}
else if ( q > 1 || q < 0) {
throw new RangeError("Quantile must be value between 0 and 1")
}

const i = q * length(sortedArr)
var low_index = sortedArr[Math.floor(i - 1)]
var high_index = sortedArr[Math.floor(i)]
return (low_index + high_index)/2

})

const diff = function(a, b) { return a - b; };

export default quantile
``````