X-Git-Url: https://git.piment-noir.org/?a=blobdiff_plain;f=src%2Futils.ts;h=efd1f124084776667bd60495f202cec32c7e04a9;hb=0676e5d32c8f8ee5c5b7f5998f4973b0a276460f;hp=d85c5ba58b4988fad867a9b35f9dca5bb5bb2b45;hpb=78099a150dc54d7adab495195fa5f133fd54e114;p=poolifier.git diff --git a/src/utils.ts b/src/utils.ts index d85c5ba5..efd1f124 100644 --- a/src/utils.ts +++ b/src/utils.ts @@ -1,3 +1,16 @@ +import * as os from 'node:os' +import type { + MeasurementStatisticsRequirements, + WorkerChoiceStrategyOptions +} from './pools/selection-strategies/selection-strategies-types' +import type { KillBehavior } from './worker/worker-options' +import type { MeasurementStatistics } from './pools/worker' + +/** + * Default task name. + */ +export const DEFAULT_TASK_NAME = 'default' + /** * An intentional empty function. */ @@ -6,19 +19,166 @@ export const EMPTY_FUNCTION: () => void = Object.freeze(() => { }) /** - * Returns the median of the given data set. + * Default worker choice strategy options. + */ +export const DEFAULT_WORKER_CHOICE_STRATEGY_OPTIONS: WorkerChoiceStrategyOptions = + { + choiceRetries: 6, + runTime: { median: false }, + waitTime: { median: false }, + elu: { median: false } + } + +/** + * Default measurement statistics requirements. + */ +export const DEFAULT_MEASUREMENT_STATISTICS_REQUIREMENTS: MeasurementStatisticsRequirements = + { + aggregate: false, + average: false, + median: false + } + +/** + * Returns safe host OS optimized estimate of the default amount of parallelism a pool should use. + * Always returns a value greater than zero. + * + * @returns The host OS optimized maximum pool size. + */ +export const availableParallelism = (): number => { + let availableParallelism = 1 + try { + availableParallelism = os.availableParallelism() + } catch { + const numberOfCpus = os.cpus() + if (Array.isArray(numberOfCpus) && numberOfCpus.length > 0) { + availableParallelism = numberOfCpus.length + } + } + return availableParallelism +} + +// /** +// * Computes the retry delay in milliseconds using an exponential back off algorithm. +// * +// * @param retryNumber - The number of retries that have already been attempted +// * @param maxDelayRatio - The maximum ratio of the delay that can be randomized +// * @returns Delay in milliseconds +// */ +// export const exponentialDelay = ( +// retryNumber = 0, +// maxDelayRatio = 0.2 +// ): number => { +// const delay = Math.pow(2, retryNumber) * 100 +// const randomSum = delay * maxDelayRatio * Math.random() // 0-(maxDelayRatio*100)% of the delay +// return delay + randomSum +// } + +/** + * Computes the median of the given data set. * * @param dataSet - Data set. * @returns The median of the given data set. */ export const median = (dataSet: number[]): number => { + if (Array.isArray(dataSet) && dataSet.length === 0) { + return 0 + } if (Array.isArray(dataSet) && dataSet.length === 1) { return dataSet[0] } const sortedDataSet = dataSet.slice().sort((a, b) => a - b) - const middleIndex = Math.floor(sortedDataSet.length / 2) - if (sortedDataSet.length % 2 === 0) { - return sortedDataSet[middleIndex / 2] + return ( + (sortedDataSet[(sortedDataSet.length - 1) >> 1] + + sortedDataSet[sortedDataSet.length >> 1]) / + 2 + ) +} + +/** + * Rounds the given number to the given scale. + * The rounding is done using the "round half away from zero" method. + * + * @param num - The number to round. + * @param scale - The scale to round to. + * @returns The rounded number. + */ +export const round = (num: number, scale = 2): number => { + const rounder = Math.pow(10, scale) + return Math.round(num * rounder * (1 + Number.EPSILON)) / rounder +} + +/** + * Is the given object a plain object? + * + * @param obj - The object to check. + * @returns `true` if the given object is a plain object, `false` otherwise. + */ +export const isPlainObject = (obj: unknown): boolean => + typeof obj === 'object' && + obj !== null && + obj?.constructor === Object && + Object.prototype.toString.call(obj) === '[object Object]' + +/** + * Detects whether the given value is a kill behavior or not. + * + * @typeParam KB - Which specific KillBehavior type to test against. + * @param killBehavior - Which kind of kill behavior to detect. + * @param value - Any value. + * @returns `true` if `value` was strictly equals to `killBehavior`, otherwise `false`. + */ +export const isKillBehavior = ( + killBehavior: KB, + value: unknown +): value is KB => { + return value === killBehavior +} + +/** + * Detects whether the given value is an asynchronous function or not. + * + * @param fn - Any value. + * @returns `true` if `fn` was an asynchronous function, otherwise `false`. + */ +export const isAsyncFunction = ( + fn: unknown +): fn is (...args: unknown[]) => Promise => { + return typeof fn === 'function' && fn.constructor.name === 'AsyncFunction' +} + +/** + * Updates the given measurement statistics. + * + * @param measurementStatistics - The measurement statistics to update. + * @param measurementRequirements - The measurement statistics requirements. + * @param measurementValue - The measurement value. + * @param numberOfMeasurements - The number of measurements. + */ +export const updateMeasurementStatistics = ( + measurementStatistics: MeasurementStatistics, + measurementRequirements: MeasurementStatisticsRequirements, + measurementValue: number, + numberOfMeasurements: number +): void => { + if (measurementRequirements.aggregate) { + measurementStatistics.aggregate = + (measurementStatistics.aggregate ?? 0) + measurementValue + measurementStatistics.minimum = Math.min( + measurementValue, + measurementStatistics.minimum ?? Infinity + ) + measurementStatistics.maximum = Math.max( + measurementValue, + measurementStatistics.maximum ?? -Infinity + ) + if (measurementRequirements.average && numberOfMeasurements !== 0) { + measurementStatistics.average = + measurementStatistics.aggregate / numberOfMeasurements + } + if (measurementRequirements.median && measurementValue != null) { + measurementStatistics.history.push(measurementValue) + measurementStatistics.median = median(measurementStatistics.history) + } } - return (sortedDataSet[middleIndex - 1] + sortedDataSet[middleIndex]) / 2 }