scala 中参数化方法的类型覆盖

Question

我尝试在 scala 中重写我的参数化方法，

我有一个像这样的抽象类：

abstract class Ranking[G <: AbstractGenome, MG <: MultiGoalLike] {
  def operate(individuals :IndexedSeq [IndividualMG[G,MG]]):IndexedSeq [IndividualMG[G,MG]
}

我想要两种类型的排名，需要 IndividualMG 的装饰版本：IndividualMG[G,MG] 与 IDistance< /code> 和 IndividualMG[G,MG] 与 IRank。

我的 IndividualMG 类签名：

class IndividualMG[G <: AbstractGenome,MG <: MultiGoalLike] (val genome: G,val multiGoal:MG) 

我的两个排名类：

class Ranking1 [G <: AbstractGenome, MG <: MultiGoalLike] extends Ranking[G,MG] {
override def operate(individuals :IndexedSeq [IndividualMG [G,MG] with IRanking])
:IndexedSeq [IndividualMG [G,MG] with IRanking]= {
return ...
}

class Ranking2 [G <: AbstractGenome, MG <: MultiGoalLike] extends Ranking[G,MG] {
override def operate(individuals :IndexedSeq [IndividualMG [G,MG] with IDistance])
:IndexedSeq [IndividualMG [G,MG] with IDistance] = {
return ...
}

我有一个错误，这是逻辑，因为当我尝试覆盖时类型不同，但我如何验证继承类型 [I ：IndividualMG [G，MG]]对于我在类Ranking1和Ranking2中的两个排名运算符？

谢谢堆栈， SR。

Answer 1

怎么样

abstract class Ranking[
    -I <: IndividualMG[G, MG], 
    G <: AbstractGenome, 
    MG <: MultiGoalLike] {
  def operate(individuals :IndexedSeq[I])
}

class Ranking1[
   G <: AbstractGenome, 
   MG <: MultiGoalLike] 
  extends Ranking[IndividualMG[G,MG] with IDistance, G, MG] ...

Answer 2

我认为您正在寻找路径依赖类型。

abstract class AbstractGenome
class ConcreteGenome extends AbstractGenome
class MultiGoalLike
trait IRanking

abstract class Ranking[G <: AbstractGenome, MG <: MultiGoalLike] {
  type RankType <: IndividualMG[G,MG]
  def operate(individuals :IndexedSeq [RankType])
}

class IndividualMG[G <: AbstractGenome,MG <: MultiGoalLike] (val genome: G,
                                                          val multiGoal:MG)

class Ranking1 [G <: AbstractGenome, MG <: MultiGoalLike] extends Ranking[G,MG] {
  type RankType = IndividualMG [G,MG] with IRanking
  override def operate(individuals :IndexedSeq [RankType])= println("it worked!")
}

val cg = new ConcreteGenome
val mgl = new MultiGoalLike

val r1 = new Ranking1[ConcreteGenome, MultiGoalLike]

val i1 = new IndividualMG(cg, mgl)
val i2 = new IndividualMG(cg, mgl) with IRanking

r1.operate(Vector(i1)) // error
r1.operate(Vector(i2)) // OK