Skip to content
This repository has been archived by the owner on Apr 25, 2020. It is now read-only.

loverdos/scalapipes

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

What

Scala with forward pipe operator, as in F# and OCaml.

The original use-case is how to handle collections, that is why the parent package of all the code is com.ckkloverdos.collection.pipes.

To get an idea of what you can do:

import com.ckkloverdos.pipes.collection.pipes._

val letters = Seq('a', 'b', 'c', 'd', 'a')
val As = letters |> 
         PSeq.filter(ch => ch == 'a')
// _ works too:
val As_ = letters |>
          PSeq.filter(_ == 'a')

val numbers = IndexedSeq(1, 2, 3, 4, 5, 6)
val lessThanFour = numbers |>
                   PSeq.filter(_ < 4)
                   
val keys = Map(1 -> "keyA", 2 -> "keyB", 3 -> "keyC")
val keysPlus = keys |>
    PMap.map { case (k, v) => (k + 1, v) } |>
    PMap.ofIterable

Note that due to Scala's syntax quirkiness, |> has to be at the end of the line, so that this won't work the way you would write it in F#:

// !! Does not compile
val lessThanFour = numbers
                   |> PSeq.filter(_ < 4)

Why

First, I did it for fun, while learning F# and it was a small design challenge to translate it into Scala. The pipe operator is mainstream now. It has been since Unix time. Ocaml will have it builtin in its next version (I did not check, this could be live now).

I discovered the code to be more readable this way than the traditional OO dot chaining. Why? Because the flow is still natural (you start from the original data and you keep transforming them) and the intent is crystal clear: at each step you can actually understand both the flow of data and the flow of their respective types. I have not seen this noted elsewhere. This observation struck me as lightning and I think is of paramount importance for code readability.

Currently, I am using this at a couple of projects, trying to see how it scales.

Details

This is the pipe operator in OCaml and F#:

let (|>) x f = f x

and this is the translation I gave it for Scala:

implicit final class MLPipe[T](val x: T) extends AnyVal {
    def |>[B](f: (T) => B) = f(x)
}

So, any value is lifted to an object having the pipe operator built-in.

The implementation is deliberately kept simple, for now at least.

Type inference as served from the compiler may not be your best friend at all times.

Also, I could do away with some code duplication, although this does not bother me very much right now.

Overall, the fun factor of doing this has been enormous.

Motto

Less is more. Use both with a pipe.

About

Scala with forward pipe operator, as in F# and OCaml.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages