Last active
February 28, 2018 11:24
-
-
Save derekjw/8ccab15a78193e18eafcef235216e01d to your computer and use it in GitHub Desktop.
Porting some akka-stream combinators to fs2
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import fs2._ | |
import scala.concurrent.ExecutionContext.Implicits.global | |
Scheduler[IO](2) | |
.flatMap { scheduler => | |
Stream.range(0, 96) | |
.covary[IO] | |
.through(flatMapAsync[IO, Int, Int](16)(n => scheduler.delay[IO, Int](Stream.emit(n), Random.nextInt(1000).millis))) | |
} | |
.onFinalize(IO(println("done"))) | |
.through(groupBy[IO, Int, (Int, Int), Int](_ % 16)(k => _.scan((k, 0))((acc, n) => (acc._1, acc._2 + n)))) | |
.zipWithIndex | |
.observe1(n => IO(println(n))) | |
.compile | |
.drain | |
.unsafeRunSync() | |
def flatMapAsync[F[_]: Effect, I, O](parallelism: Int)(f: I => Stream[F, O])(implicit executionContext: ExecutionContext): Pipe[F, I, O] = | |
_.map(i => f(i)).join(parallelism) | |
def mapEvalAsync[F[_]: Effect, I, O](parallelism: Int)(f: I => F[O])(implicit executionContext: ExecutionContext): Pipe[F, I, O] = | |
flatMapAsync(parallelism)(f.andThen(Stream.eval)) | |
def groupBy[F[_]: Effect, I, O, K](f: I => K, groupBuffer: Int = 16)(subStream: K => Pipe[F, I, O])(implicit executionContext: ExecutionContext): Pipe[F, I, O] = | |
_.noneTerminate | |
.through { | |
evalMapAccumulate[F, Map[K, Queue[F, Option[I]]], Option[I], Option[Stream[F, O]]](Effect[F].pure(Map.empty[K, Queue[F, Option[I]]])) { | |
case (queues, someN@Some(n)) => | |
val partition = f(n) | |
queues.get(partition).map { queue => | |
queue.enqueue1(someN).map(_ => (queues, Option.empty[Stream[F, O]])) | |
}.getOrElse { | |
Queue.bounded[F, Option[I]](groupBuffer).flatMap { queue => | |
queue.enqueue1(someN).map { _ => | |
val result = Option(queue.dequeueAvailable.unNoneTerminate.through(subStream(partition))) | |
(queues + (partition -> queue), result) | |
} | |
} | |
} | |
case (queues, None) => | |
queues.valuesIterator.foldLeft(Effect[F].unit)((e, q) => e.flatMap(_ => q.enqueue1(None))).map(_ => (queues, Option.empty[Stream[F, O]])) | |
} | |
} | |
.map(_._2) | |
.unNone | |
.joinUnbounded | |
def evalMapAccumulate[F[_]: Monad, S, I, O](init: F[S])(f: (S, I) => F[(S, O)]): Pipe[F, I, (S, O)] = { | |
def next(init: S, stream: Stream[F, I]): Pull[F, (S, O), Unit] = { | |
stream.pull.uncons1.flatMap { | |
case Some((i, rest)) => | |
Pull.eval(f(init, i)).flatMap { so => | |
Pull.output1(so) >> next(so._1, rest) | |
} | |
case None => Pull.done | |
} | |
} | |
next(init, _).stream | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment