hedgehog.predef.package.scala Maven / Gradle / Ivy
package hedgehog
/**
* We have our own FP predef for 2 reasons.
*
* 1. The obvious political reasons. I don't think there are any really good reasons to need more than one
* implementation of this library (if we do our job correctly).
*
* Probably more importantly:
*
* 2. Library dependencies _do_ have a cost. Especially in the JVM world where we insist of relying on binary
* compatibility.
*/
package object predef {
implicit def eitherOps[L, R](e: Either[L, R]): EitherOps[L, R] =
new EitherOps(e)
type State[S, A] = StateT[Identity, S, A]
def State: StateTOpt[Identity] =
new StateTOpt[Identity] {}
def stateT[M[_]]: StateTOpt[M] =
new StateTOpt[M] {}
def some[A](a: A): Option[A] =
Some(a)
@annotation.tailrec
def findMap[A, B](fa: LazyList[A])(f: A => Option[B]): Option[B] = {
fa match {
case LazyList.Nil() =>
None
case LazyList.Cons(h, t) =>
f(h()) match {
case Some(b) =>
Some(b)
case None =>
findMap(t())(f)
}
}
}
/** Performs the action `n` times, returning the list of results. */
def replicateM[M[_], A](n: Int, fa: M[A])(implicit F: Applicative[M]): M[List[A]] =
sequence(List.fill(n)(fa))
/** Strict sequencing in an applicative functor `M` that ignores the value in `fa`. */
def sequence[M[_], A](fa: List[M[A]])(implicit F: Applicative[M]): M[List[A]] =
traverse(fa)(identity)
def traverse[M[_], A, B](fa: List[A])(f: A => M[B])(implicit F: Applicative[M]): M[List[B]] =
fa match {
case Nil =>
F.point(Nil)
case h :: t =>
F.ap(traverse(t)(f))(F.ap(f(h))(F.point((h2 : B) => (t2 : List[B]) => h2 :: t2)))
}
}
© 2015 - 2024 Weber Informatics LLC | Privacy Policy