I am trying to learn how to use FreeMonads to implement interpreters for my services.
Suppose I have
sealed trait ServiceAction[T] extends Product with Serializable
case class ConsumeCommand(cmd: AccruePoints) extends ServiceAction[AccruePointModel]
case class CreateEvent(evt: PointsAccruedEvent) extends ServiceAction[PointsAccruedEvent]
sealed trait LogAction[T] extends Product with Serializable
case class Info(msg: String) extends LogAction[Unit]
case class Error(msg: String) extends LogAction[Unit]
and a Monad of the action
type LogActionF[A] = Free[LogAction, A]
type ServiceActionF[A] = Free[ServiceAction, A]
Next, I define my service like this:
trait PointAccrualService {
def consume(cmd: AccruePoints): ServiceActionF[AccruePointModel] = Free.liftF(ConsumeCommand(cmd))
def emit(evt: PointsAccruedEvent) : ServiceActionF[PointsAccruedEvent] = Free.liftF(CreateEvent(evt))
}
and
trait LogService {
def info(msg: String) : LogActionF[Unit] = Free.liftF(Info(msg))
def error(msg: String) : LogActionF[Unit] = Free.liftF(Error(msg))
}
with an object of each
object LogService extends LogService
object PointAccrualService extends PointAccrualService
My LogServiceInterpreter is like this:
case class LogServiceConsoleInterpreter() extends LogServiceInterpreter {
def apply[A](action: LogActionF[A]): Task[A] = action.foldMap(handler)
protected def handler = new (LogAction ~> Task) {
override def apply[A](fa: LogAction[A]) = fa match {
case Info(m) =>
now(info(m))
case Error(m) =>
now(error(m))
}
}
def info(msg: String): Unit = {
println(s"INFO: $msg")
}
def error(msg: String): Unit = {
println(s"ERROR: $msg")
}
}
Similarly, my PointAccuralServiceInterpreter is like this:
case class PointAccuralServiceInterpreter() {
def apply[A] (action: ServiceActionF[A]) : Task[A] = action.foldMap(handler)
protected def handler = new (ServiceAction ~> Task) {
override def apply[A](fa: ServiceAction[A]): Task[A] = fa match {
case ConsumeCommand(cmd) => {
println("Service ConsumeCommand:" + cmd)
now(cmd)
}
case CreateEvent(evt) => {
println("Service CreateEvent:" + evt)
now(evt)
}
}
}
}
My logic is straightforward, I want to log, and consume my command and then create an event, sort of like an event sourcing:
val ret = for {
_ <- logService.info("Command: " + cmd)
model <- service.consume(cmd)
_ <- logService.info("Model: " + model)
evt <- service.emit(model.toEvent("200", "Event Sent"))
_ <- logService.info("Event:" + evt)
} yield evt
This code doesn't even compile actually.
What should I do from here? I think I am supposed to use Coproduct to chain them and execute this piece of logic by feeding my interpreter.
I found something here
https://groups.google.com/forum/#!topic/scalaz/sHxFsFpE86c
or it's said I can use Shapeless to do so
Folding a list of different types using Shapeless in Scala
They are all too complicated. All I want is, after I define my logic, how do I execute it?
Hope I put enough details here for an answer. I really want to learn this. Thanks
I slightly modified your code to create a self-contained running example. I also added a possible answer to your question, how to execute your program, following Rúnar Bjarnason's ideas, using Scalaz 7.2. (I did not find the or operator for the natural transformations in Scalaz, so I added it here.)
I also added a few stubs to give your actions something to fiddle with and simplified your services to the handlers inside (since I had to create a new service for both languages combined). Furthermore I changed your Task.now{...} to Task{...} to create an asynchronous Task, which is executed on the last line of code.
Here is the full code:
import scala.language.{higherKinds, implicitConversions}
import scalaz._
import scalaz.concurrent.Task
/* Stubs */
case class AccruePoints()
case class AccruePointModel(cmd: AccruePoints) {
def toEvent(code: String, description: String): PointsAccruedEvent = PointsAccruedEvent(code, description)
}
case class PointsAccruedEvent(code: String, description: String)
/* Actions */
sealed trait ServiceAction[T] extends Product with Serializable
case class ConsumeCommand(cmd: AccruePoints) extends ServiceAction[AccruePointModel]
case class CreateEvent(evt: PointsAccruedEvent) extends ServiceAction[PointsAccruedEvent]
sealed trait LogAction[T] extends Product with Serializable
case class Info(msg: String) extends LogAction[Unit]
case class Error(msg: String) extends LogAction[Unit]
/* Handlers */
object PointAccuralServiceHandler extends (ServiceAction ~> Task) {
override def apply[A](fa: ServiceAction[A]): Task[A] = fa match {
case ConsumeCommand(cmd) => {
println("Service ConsumeCommand:" + cmd)
Task(consume(cmd))
}
case CreateEvent(evt) => {
println("Service CreateEvent:" + evt)
Task(evt)
}
}
def consume(cmd: AccruePoints): AccruePointModel =
AccruePointModel(cmd)
}
case object LogServiceConsoleHandler extends (LogAction ~> Task) {
override def apply[A](fa: LogAction[A]): Task[A] = fa match {
case Info(m) =>
Task(info(m))
case Error(m) =>
Task(error(m))
}
def info(msg: String): Unit = {
println(s"INFO: $msg")
}
def error(msg: String): Unit = {
println(s"ERROR: $msg")
}
}
/* Execution */
class Service[F[_]](implicit I1: Inject[ServiceAction, F], I2: Inject[LogAction, F]) {
def consume(cmd: AccruePoints): Free[F, AccruePointModel] = Free.liftF(I1(ConsumeCommand(cmd)))
def emit(evt: PointsAccruedEvent): Free[F, PointsAccruedEvent] = Free.liftF(I1(CreateEvent(evt)))
def info(msg: String): Free[F, Unit] = Free.liftF(I2(Info(msg)))
def error(msg: String): Free[F, Unit] = Free.liftF(I2(Error(msg)))
}
object Service {
implicit def instance[F[_]](implicit I1: Inject[ServiceAction, F], I2: Inject[LogAction, F]) = new Service[F]
}
def prg[F[_]](implicit service: Service[F]) = {
val cmd = AccruePoints()
for {
_ <- service.info("Command: " + cmd)
model <- service.consume(cmd)
_ <- service.info("Model: " + model)
evt <- service.emit(model.toEvent("200", "Event Sent"))
_ <- service.info("Event:" + evt)
} yield evt
}
type App[A] = Coproduct[ServiceAction, LogAction, A]
def or[F[_], G[_], H[_]](f: F ~> H, g: G ~> H) =
new (({type t[x] = Coproduct[F, G, x]})#t ~> H) {
override def apply[A](c: Coproduct[F, G, A]): H[A] = c.run match {
case -\/(fa) => f(fa)
case \/-(ga) => g(ga)
}
}
val app = prg[App]
val ret = app.foldMap(or(PointAccuralServiceHandler, LogServiceConsoleHandler))
ret.unsafePerformSync
Related
I'm trying to implement an application that controls a camera. Camera commands are represented as a stream of CameraAction objects:
sealed trait CameraMessage
case object Record(recordId: String) extends CameraMessage
case object Stop extends CameraMessage
...
val s = Stream[F, CameraMessage]
Let's say I have a test stream that emits "Record" and emits "Stop" 20 seconds later, after another 20 seconds another "Record" message is emitted and so on, the input stream is infinite.
Then the app consumes "Record" it should create an instance of GStreamer pipeline (i.e. it is an effect) and "run" it, on "Stop" it 'stops' the pipeline and closes it. Then on subsequent "Record" the pattern is repeated with new GStreamer pipeline.
The problem is that I need to pass an instance of impure, mutable object between handles of stream events.
FS2 documentation suggest to use chunks to make a stream stateful, so I tried
def record(gStreamerPipeline: String, fileName: String)
(implicit sync: Sync[F]): F[Pipeline] =
{
... create and open pipeline ...
}
def stopRecording(pipe: Pipeline)(implicit sync: Sync[F]): F[Unit] = {
... stop pipeline, release resources ...
}
def effectPipe(pipelineDef: String)(implicit L: Logger[F]):
Pipe[F, CameraMessage, F[Unit]] = {
type CameraSessionHandle = Pipeline
type CameraStream = Stream[F, CameraSessionHandle]
s: Stream[F, CameraMessage] =>
s.scanChunks(Stream[F, CameraSessionHandle]()) {
case (s: CameraStream, c: Chunk[CameraMessage]) =>
c.last match {
case Some(Record(fileName)) =>
(Stream.bracket(record(pipelineDef, fileName))(p => stopRecording(p)), Chunk.empty)
case Some(StopRecording) =>
(Stream.empty, Chunk(s.compile.drain))
case _ =>
(s, Chunk.empty)
}
}
}
The problem with this code that actual recording does not happen on 'Record' event but rather then the effect of the whole chunk is evaluated, i.e. when 'StopRecording' message arrives the camera is turned on and then immediately turned off again.
How can I pass a "state" without chunking? Or is there any other way to achieve the result I need?
This may be similar to
FS2 Stream with StateT[IO, _, _], periodically dumping state
but the difference is that the state in my case is not a pure data structure but a resource.
I eventually was able so solve it using Mutable Reference pattern as described in https://typelevel.org/blog/2018/06/07/shared-state-in-fp.html
Here is the code:
import cats.effect._
import cats.syntax.all._
import fs2.Stream
import scala.concurrent.{ExecutionContext, ExecutionContextExecutor}
import scala.language.higherKinds
class FRef[F[_], T](implicit sync: Sync[F]) {
private var state: T = _
def set(n: T): F[Unit] = sync.delay(this.state = n)
def get: F[T] = sync.pure(state)
}
object FRef {
def apply[F[_], T](implicit sync: Sync[F]): F[FRef[F, T]] = sync.delay { new FRef() }
}
class CameraImpl(id: String) extends Camera {
override def record(): Unit = {
println(s"Recording $id")
}
override def stop(): Unit = {
println(s"Stopping $id")
}
override def free(): Unit = {
Thread.sleep(500)
println(s"Freeing $id")
}
}
object Camera {
def apply(id: String) = new CameraImpl(id)
}
trait Camera {
def record(): Unit
def stop(): Unit
def free(): Unit
}
sealed trait CameraMessage
case class Record(recordId: String) extends CameraMessage
case object StopRecording extends CameraMessage
class Streamer[F[_]](implicit sync: Sync[F]) {
def record(id: String): F[Camera] = {
sync.delay {
val r = Camera(id)
r.record()
r
}
}
def stopRecording(pipe: Camera): F[Unit] = {
sync.delay {
pipe.stop()
pipe.free()
}
}
def effectPipe(state: FRef[F, Option[Camera]])(
implicit sync: Sync[F]): Stream[F, CameraMessage] => Stream[F, Unit] = {
type CameraStream = Stream[F, Camera]
s: Stream[F, CameraMessage] =>
s.evalMap {
case Record(fileName) =>
for {
r <- record(fileName)
_ <- state.set(Some(r))
} yield ()
case StopRecording =>
for {
s <- state.get
_ <- stopRecording(s.get)
_ <- state.set(None)
} yield ()
}
}
}
object FS2Problem extends IOApp {
import scala.concurrent.duration._
override def run(args: List[String]): IO[ExitCode] = {
implicit val ec: ExecutionContextExecutor = ExecutionContext.global
val streamer = new Streamer[IO]
val s = Stream.awakeEvery[IO](5.seconds).take(10).zipWithIndex.map {
case (_, idx) =>
idx % 2 match {
case 0 =>
Record(s"Record $idx")
case _ =>
StopRecording
}
}
val ss = for {
streamerState <- Stream.eval(FRef[IO, Option[Camera]])
s <- s.through(streamer.effectPipe(streamerState))
} yield ()
ss.compile.drain.map(_ => ExitCode.Success)
}
}
I am trying to compose futures with for-comprehension and EitherT, but I am having trouble due the return types. Please can someone explain why this does not compile and how can I make it compile changing the for-comprehension?
import scala.concurrent.Future
import scala.concurrent.ExecutionContext.Implicits.global
import cats.data.EitherT
import cats.implicits._
object CatsApp extends App {
case class L1(a: String)
case class L2(a: String)
case class L3(a: String)
case class R1(num: Int)
case class R2(num: Int)
case class R3(num: Int)
def func1: Future[Either[L1, R1]] = {
if (true) Future(Right(R1(1)))
else Future(Left(L1("A")))
}
def func2: Future[Either[L2, R2]] = {
if (true) Future(Right(R2(1)))
else Future(Left(L2("A")))
}
def func3(a: R1, b: R2): Future[Either[L3, R3]] = {
if (true) Future(Right(R3(a.num + b.num)))
else Future(Left(L3("A")))
}
def comp = {
for {
f1 <- EitherT(func1)
f2 <- EitherT(func2)
f3 <- EitherT(func3(f1, f2))
} yield f3
}
}
In a for-comprehension the type and the bias of the first step in the chain determines the type of all the rest of the steps in the chain have to be. Because Either is right-biased, we can only change the right type between the steps of the for-comprehension, as advised by #Krzysztof. For example,
val e1: Either[String, Int] = Right(42)
val e2: Either[String, Char] = Right('A')
for {
num <- e1
char <- e2
} yield "I compile despite having different Rights"
In your case the type of first step EitherT(func1) is EitherT[Future, L1, R1], hence the next steps EitherT(func2) and EitherT(func3(f1, f2)) must have the following shape of the type
EitherT[Future, L1, X]
where only X can vary. One way to make your for-comprehension happy is to create algebraic data types out of Ls like so
sealed abstract class L(val a: String)
final case class L1(s: String) extends L(s)
final case class L2(s: String) extends L(s)
final case class L3(s: String) extends L(s)
Here is a working example
object CatsApp extends App {
sealed abstract class L(val a: String)
final case class L1(s: String) extends L(s)
final case class L2(s: String) extends L(s)
final case class L3(s: String) extends L(s)
case class R1(num: Int)
case class R2(num: Int)
case class R3(num: Int)
def func1: Future[Either[L, R1]] = {
if (true) Future(Right(R1(1)))
else Future(Left(L1("A")))
}
def func2: Future[Either[L, R2]] = {
if (true) Future(Right(R2(1)))
else Future(Left(L2("A")))
}
def func3(a: R1, b: R2): Future[Either[L, R3]] = {
if (true) Future(Right(R3(a.num + b.num)))
else Future(Left(L3("A")))
}
def comp: EitherT[Future, L, R3] = {
for {
f1 <- EitherT(func1)
f2 <- EitherT(func2)
f3 <- EitherT(func3(f1, f2))
} yield f3
}
comp.value.andThen(v => println(v))
}
which outputs
Success(Right(R3(2)))
I have a type class and a few instances:
trait TC[T] { def doThings(x: T): Unit }
implicit val tcA = new TC[A] { /* ... */}
implicit val tcB = new TC[B] { /* ... */}
implicit val tcC = new TC[C] { /* ... */}
/* ... */
In my call site, I have input as Any, and I need to check if there is an implicit for the input actual type:
def process(in: Any) = in match {
case x: A => implicitly[TC[A]].doThings(x)
case x: B => implicitly[TC[B]].doThings(x)
case x: C => implicitly[TC[C]].doThings(x)
//...
}
This seems tedious and unnecessary, as I have to list all the classes that have this type class instance. Can I achieve this by something like:
def process(in: Any) = in match {
case x: T : TC => implicitly[TC[T]].doThings(x) //This does not work
}
Edit: input is an Any (an Object from a Java library). Cannot use generic or context bound on the input.
If you really want to do what you have mentioned in your question, you can write it as below, but if you just want to call doThings by finding an implicit instance of appropriate TC - refer João Guitana answer
object Main extends App {
class A
class B
class C
trait TC[T] { def doThings(x: T): Unit }
implicit val tcA = new TC[A] {
override def doThings(x: A): Unit = println("From A")
}
implicit val tcB = new TC[B] {
override def doThings(x: B): Unit = println("From B")
}
implicit val tcC = new TC[C] {
override def doThings(x: C): Unit = println("From C")
}
def process[T: ClassTag](in: T) = in match {
case x: A => implicitly[TC[A]].doThings(x)
case x: B => implicitly[TC[B]].doThings(x)
case x: C => implicitly[TC[C]].doThings(x)
}
process(new A())
process(new B())
process(new C())
}
/* === Output ====
From A
From B
From C
*/
You need to ask for an implicit TC, Any won't work. As follows:
trait TC[T] { def doThings(x: T): Unit }
implicit def tcS: TC[String] = new TC[String] {
override def doThings(x: String): Unit = println("string")
}
implicit def tcI: TC[Int] = new TC[Int] {
override def doThings(x: Int): Unit = println("int")
}
def process[T : TC](x: T): Unit = implicitly[TC[T]].doThings(x)
process("")
process(1)
// process(4L) wont compile
Try it out!
I'm learning about the Free monads, and I've put together a simple example in Scala where I use them to define two domain specific languages.
The first monad deals with the side effects of a repository. I have implemented an interpreter that uses the state monad to manage the state, but in a real program I'd use a database.
The second monad deals with IO.
import cats.data.State
import cats.{Id, ~>}
import cats.free.Free
import cats.free.Free.liftF
final case class Todo(title: String, body: String)
def represent(todo: Todo) = s"${todo.title}: ${todo.body}"
sealed trait CRUDActionA[T]
final case class Find(key: String) extends CRUDActionA[Option[Todo]]
final case class Add(data: Todo) extends CRUDActionA[Unit]
type CRUDAction[T] = Free[CRUDActionA, T]
def find(key: String): CRUDAction[Option[Todo]] = liftF[CRUDActionA, Option[Todo]](Find(key))
def add(data: Todo): CRUDAction[Unit] = liftF[CRUDActionA, Unit](Add(data))
type TodosState[A] = State[List[Todo], A]
val repository: CRUDActionA ~> TodosState = new (CRUDActionA ~> TodosState) {
def apply[T](fa: CRUDActionA[T]): TodosState[T] = fa match {
case Add(todo) => State.modify(todos => todos :+ todo)
case Find(title) => State.inspect(todos => todos find (_.title == title))
}
}
sealed trait IOActionA[T]
final case class Out(str: String) extends IOActionA[Unit]
type IOAction[T] = Free[IOActionA, T]
def out(str: String): IOAction[Unit] = liftF[IOActionA, Unit](Out(str))
val io: IOActionA ~> Id = new (IOActionA ~> Id) {
override def apply[A](fa: IOActionA[A]): Id[A] = fa match {
case Out(todo) => println(todo)
}
}
Then, I can put together these two "programs"
def addNewTodo: Free[CRUDActionA, Option[Todo]] = for {
_ <- add(Todo(title = "Must do", body = "Must do something"))
todo <- find("Must do")
} yield todo
def outProgram(todo: Todo): IOAction[Unit] = for {
_ <- out(represent(todo))
} yield ()
And run them doing
val (_, mayBeTodo) = (addNewTodo foldMap repository run List()).value
outProgram(mayBeTodo.get).foldMap(io)
I understand this is far from ideal, and I'd like to write a program as and an interpreter that supports:
def fullProgram = for {
_ <- add(Todo(title = "Must do", body = "Must do something"))
todo <- find("Must do") // This is an option!!!
_ <- out(represent(todo)) // But represent expects a Todo
} yield ()
So the questions are:
How can I stack the two monads together into a "fullProgram"
How can I compose the two interpreters into a new interpreter?
How do I deal with the Option[Todo] returned by find, and then passed to
represent
Answer to questions 1 & 2:
type TodoApp[A] = Coproduct[IOActionA, CRUDActionA, A]
class CRUDActions[F[_]](implicit I: Inject[CRUDActionA, F]) {
def find(key: String): Free[F, Option[Todo]] = Free.inject[CRUDActionA, F](Find(key))
def add(data: Todo): Free[F, Unit] = Free.inject[CRUDActionA, F](Add(data))
}
object CRUDActions {
implicit def crudActions[F[_]](implicit I: Inject[CRUDActionA, F]): CRUDActions[F] = new CRUDActions[F]
}
class IOActions[F[_]](implicit I: Inject[IOActionA, F]) {
def out(str: String): Free[F, Unit] = Free.inject[IOActionA, F](Out(str))
}
object IOActions {
implicit def ioActions[F[_]](implicit I: Inject[IOActionA, F]): IOActions[F] = new IOActions[F]
}
def fullProgram(implicit C : CRUDActions[TodoApp], I : IOActions[TodoApp]): Free[TodoApp, Unit] = {
for {
_ <- C.add(Todo(title = "Must do", body = "Must do something"))
todo <- C.find("Must do")
_ <- I.out(represent(todo.get))
} yield ()
}
object ConsoleCatsInterpreter extends (IOActionA ~> Id) {
def apply[A](i: IOActionA[A]) = i match {
case Out(prompt) => println(prompt).asInstanceOf[A]
}
}
object MutableListCrudInterpreter extends (CRUDActionA ~> Id) {
val data = new ListBuffer[Todo]
override def apply[A](fa: CRUDActionA[A]): Id[A] = fa match {
case Add(todo) => data.append(todo).asInstanceOf[A]
case Find(title) => data.find( _.title == title).asInstanceOf[A]
}
}
val interpreter: TodoApp ~> Id = ConsoleCatsInterpreter or MutableListCrudInterpreter
fullProgram.foldMap(interpreter)
Here is the code:
def transform1(f: String => String): Unit = {
val s = getString
f.andThen(putString)(s)
}
def transform2(f: String => Option[String]): Unit = {
val s = getString
f(s).foreach(putString(_))
}
How do you express these two ideas in one single function?
Method overloading does not work and seems discouraged by the community.
I didn't understand that why anyone may want this but here is a way to do it:
def transform(f: Either[(String => String), (String => Option[String])]: Unit = f match {
case Left(f) => // do transform1 here
case Right(f) => //do transform2 here
}
As I said at the begining you probably shouldn't want to do this; perhaps you should directly ask what you want.
The pattern to avoid overloading is to convert disparate arguments to a common, specific type. There could be any number of such conversions.
Not sure this is the most compelling example, however.
object X {
trait MapFlat[-A, +B] { def apply(x: A): B }
implicit class mapper[A](val f: A => A) extends MapFlat[A, A] {
override def apply(x: A) = {
val res = f(x)
println(res)
res
}
}
implicit class flatmapper[A](val f: A => Option[A]) extends MapFlat[A, Option[A]] {
override def apply(x: A) = {
val res = f(x)
res foreach println
res
}
}
def f[B](g: MapFlat[String, B]) = {
g("abc")
}
}
object Test extends App {
import X._
f((s: String) => s)
f((s: String) => Some(s))
}
One way to do it will be type classes, here's a sample -
trait Transformer[T] {
def transform(foo: String => T)
}
object Transformer {
implicit object StringTransformer extends Transformer[String] {
override def transform(foo: (String) => String): Unit = ??? // Your logic here
}
implicit object OptStringTransformer extends Transformer[Option[String]] {
override def transform(foo: (String) => Option[String]): Unit = ??? // Your logic here
}
}
class SampleClass {
def theOneTransformYouWant[T: Transformer](f: String => T) = {
implicitly[Transformer[T]].transform(f)
}
def canUseBothWays(): Unit = {
theOneTransformYouWant((s: String) => s)
theOneTransformYouWant((s: String) => Some(s))
}
}
Another way would be the magnet pattern
http://spray.io/blog/2012-12-13-the-magnet-pattern/
sealed trait TransformationMagnet {
def apply(): Unit
}
object TransformationMagnet {
implicit def fromString(f: String => String): TransformationMagnet =
new TransformationMagnet {
def apply(): Unit = ??? // Your code goes here
}
implicit def fromOptString(f: String => Option[String]): TransformationMagnet =
new TransformationMagnet {
def apply(): Unit = ??? // your code goes here
}
}
class SampleClass {
def theOneTransformYouWant(f: TransformationMagnet) = {
???
}
def hereWeUseItInBothWays(): Unit = {
theOneTransformYouWant((s: String) => s)
theOneTransformYouWant((s: String) => Some(s))
}
}
add a new parameter on the description typeOfTransform
add a conditional inside the function
if (typeOfTransform == type1){
//functionality1
}else {
//functionality2
}
Just for completeness, you can actually overload methods like this by adding implicit arguments which will always be available:
def transform(f: String => Option[String]): Unit = ...
def transform(f: String => String)(implicit d: DummyImplicit): Unit = ...