io.gatling.core.util.RandomDistribution.scala Maven / Gradle / Ivy
The newest version!
/*
* Copyright 2011-2024 GatlingCorp (https://gatling.io)
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package io.gatling.core.util
import java.util.concurrent.ThreadLocalRandom
import scala.annotation.tailrec
import io.gatling.commons.util.Collections._
private[core] object RandomDistribution {
def uniform[T](possibilities: List[T]): RandomDistribution[T] =
new RandomDistribution(possibilities.map(1 -> _), possibilities.size, None)
private val PercentWeightsNormalizingFactor = 1000000 // 100% * 1000000 < Int.MaxValue so no risk of overflowing
def percentWeights[T](possibilities: List[(Double, T)], fallback: T): RandomDistribution[T] = {
val sum = possibilities.sumBy(_._1)
require(sum <= 100.000001, s"Weights sum $sum mustn't be bigger than 100%")
val intendedTotalIs100 = math.abs(sum - 100.0) <= 0.000001
val (_, headChain) :: tail = possibilities
val normalizedTail: List[(Int, T)] = tail.map { case (weight, chain) =>
(weight * PercentWeightsNormalizingFactor).toInt -> chain
} // don't round but truncate so normalized sum doesn't because bigger than original one
val normalizedTailSum = normalizedTail.sumBy(_._1)
val normalizedHeadWeight =
if (intendedTotalIs100) {
100 * PercentWeightsNormalizingFactor - normalizedTailSum
} else {
(sum * PercentWeightsNormalizingFactor).round.toInt - normalizedTailSum
}
val normalizedPossibilities = normalizedHeadWeight -> headChain :: normalizedTail
new RandomDistribution(normalizedPossibilities, 100 * PercentWeightsNormalizingFactor, Some(fallback))
}
}
private[core] class RandomDistribution[T](possibilities: List[(Int, T)], max: Int, fallback: Option[T]) {
// visible for tests
private[util] def next(index: Int): T = {
@tailrec
def nextRec(index: Int, pos: List[(Int, T)]): T = pos match {
case Nil => fallback.getOrElse(throw new UnsupportedOperationException("No fallback is defined"))
case (weight, head) :: tail =>
if (weight > index) {
head
} else {
nextRec(index - weight, tail)
}
}
nextRec(index, possibilities)
}
def next(): T = next(ThreadLocalRandom.current.nextInt(max))
}