Composable folds in Scala
This library defines the Foldl data type (a left fold) which can be combined in the applicative style such that resulting
fold requires only one traversal over the given structure.
Library comes with common folds. You can always define your own fold by providing a step function and initial value.
Library also comes with an extension method on scala's standard collections foldWith. You can use that on standard scala collections as follows,
import fold._
import Fold._
List(1, 2, 3).foldWith(sum[Int])
scala> import fold._
import fold._
scala> import Fold._
import Fold._
scala> Seq(1, 2, 3).foldWith(sum[Int])
res1: Int = 6
scala> import fold._
import fold._
scala> import Fold._
import Fold._
scala> import scalaz._
import scalaz._
scala> type Fold[A] = Foldl[Double, A]
defined type alias Fold
scala> def mean = Apply[Fold].apply2(sum[Double], length[Double, Double])(_ / _)
mean: Fold[Double]
scala> Seq(1.0, 2.0, 3.0).foldWith(mean)
res1: Double = 2.0
Note that combined mean fold traverses List only once!
scala> import fold._
scala> import Fold._
scala> import scalaz.syntax.apply._
import scalaz.syntax.apply._
scala> def mean = (sum[Double] |@| length[Double, Double]) (_ / _)
mean: fold.Foldl[Double,Double]
scala> Seq(1.0, 2.0, 3.0).foldWith(mean)
res3: Double = 2.0
scala> import fold._
import fold._
scala> import Fold._
import Fold._
scala> def mean = sum[Double] / length[Double, Double]
mean: fold.Foldl[Double,Double]
scala> Seq(1.0, 2.0, 3.0).foldWith(mean)
res2: Double = 2.0
(/) function uses Foldl's applicative instance, so again List is traversed only once.
- Gabriel Gonzalez's foldl library
- Phil Freeman's purescript-folds
- Max Rabkin’s Beautiful folding
- Conal Elliott's blogpost
Feedbacks welcome!
Distributed under the Apache 2.0. Copyright © 2016-2018 Amar Potghan.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent