As an exercise, I am trying to see if I can take a List[Any] and "cast" it into a case class using shapeless.
A very basic example of what I am trying to achieve:
case class Foo(i: Int, j: String)
val foo: Option[Foo] = fromListToCaseClass[Foo]( List(1:Any, "hi":Any) )
Here is how I am shaping my solution (this can be quite off):
def fromListToCaseClass[CC <: Product](a: List[Any]): Option[CC] = a.toHList[???].map( x => Generic[CC].from(x) )
Here is my reasoning:
I know that you can go from a case class to an HList[T] (CC -> HList[T]); where T is the type of the HList. I also know that you can create an HList from a list (list -> Option[HList]) as long as you know the type of the HList. Finally I know that you can go from an HList to a case class (HList -> CC).
CC -> HList[T]
list -> Option[HList[T]] -> Option[CC]
I am wondering if this makes sense or if I am way off here. Can we make this work? Any other suggestions? Thanks!
This can be done very straightforwardly using shapeless's Generic and FromTraversable type classes,
import scala.collection.GenTraversable
import shapeless._, ops.traversable.FromTraversable
class FromListToCaseClass[T] {
def apply[R <: HList](l: GenTraversable[_])
(implicit gen: Generic.Aux[T, R], tl: FromTraversable[R]): Option[T] =
tl(l).map(gen.from)
}
def fromListToCaseClass[T] = new FromListToCaseClass[T]
(There's some accidental complexity here due to Scala's awkwardness when it comes to mixing explicit and inferred type parameters: we want to specify T explicitly, but have R inferred for us).
Sample REPL session ...
scala> case class Foo(i: Int, j: String)
defined class Foo
scala> fromListToCaseClass[Foo](List(23, "foo"))
res0: Option[Foo] = Some(Foo(23,foo))
scala> fromListToCaseClass[Foo](List(23, false))
res1: Option[Foo] = None
You can do it with shapeless the following way:
import shapeless._
trait Creator[A] { def apply(list:List[Any]): Option[A] }
object Creator {
def as[A](list: List[Any])(implicit c: Creator[A]): Option[A] = c(list)
def instance[A](parse: List[Any] => Option[A]): Creator[A] = new Creator[A] {
def apply(list:List[Any]): Option[A] = parse(list)
}
def arbitraryCreate[A] = instance(list => list.headOption.map(_.asInstanceOf[A]))
implicit val stringCreate = arbitraryCreate[String]
implicit val intCreate = arbitraryCreate[Int]
implicit val hnilCreate = instance(s => if (s.isEmpty) Some(HNil) else None)
implicit def hconsCreate[H: Creator, T <: HList: Creator]: Creator[H :: T] =
instance {
case Nil => None
case list => for {
h <- as[H](list)
t <- as[T](list.tail)
} yield h :: t
}
implicit def caseClassCreate[C, R <: HList](
implicit gen: Generic.Aux[C, R],
rc: Creator[R]): Creator[C] =
instance(s => rc(s).map(gen.from))
}
And
val foo:Option[Foo] = Creator.as[Foo](List(1, "hi"))
Related
Scala compiler detects the following two map functions as duplicates conflicting with each other:
class ADT {
def map[Output <: AnyVal](f: Int => Output): List[Output] = ???
def map[Output >: Null <: AnyRef](f: Int => Output): List[Output] = ???
}
The class type of Output parameter is different. First one limits to AnyVal and second one limits to AnyRef. How can I differentiate them?
The problem is not differentiating AnyVal from AnyRef so much as getting around the fact that both method signatures become the same after erasure.
Here is a neat trick to get around this kind of problem. It is similar to what #som-snytt did, but a bit more generic, as it works for other similar situations as well (e.g. def foo(f: Int => String): String = ??? ; def foo(f: String => Int): Int = ??? etc.):
class ADT {
def map[Output <: AnyVal](f: Int => Output): List[Output] = ???
def map[Output >: Null <: AnyRef](f: Int => Output)(implicit dummy: DummyImplicit): List[Output] = ???
}
The cutest thing is that this works "out of the box". Apparently, a DummyImplicit is a part of standard library, and you always have the thing in scope.
You can have more than two overloads this way too by just adding more dummies to the list.
scala 2.13.0-M5> :pa
// Entering paste mode (ctrl-D to finish)
object X {
def map[Output <: AnyVal](f: Int => Output) = 1
def map[O](f: Int => O)(implicit ev: O <:< AnyRef) = 2
}
// Exiting paste mode, now interpreting.
defined object X
scala 2.13.0-M5> X.map((x: Int) => x*2)
res0: Int = 1
scala 2.13.0-M5> X.map((x: Int) => "")
res1: Int = 2
You could use a typeclass for that map method.
Using your exact example:
trait MyTC[Output]{
def map(f: Int => Output): List[Output]
}
object MyTC{
def apply[A](a : A)(implicit ev : MyTC[A]) : MyTC[A] = ev
implicit def anyRefMyTc[A <: AnyRef] : MyTC[A] = new MyTC[A]{
def map(f: Int => A): List[A] = { println("inside sub-AnyRef"); List.empty }
}
implicit def anyValMyTc[A <: AnyVal] : MyTC[A] = new MyTC[A]{
def map(f: Int => A): List[A] = { println("inside sub-AnyVal"); List.empty }
}
}
import MyTC._
val r1 = Option("Test1")
val r2 = List(5)
val v1 = true
val v2 = 6L
// The functions here are just to prove the point, and don't do anything.
MyTC(r1).map(_ => None)
MyTC(r2).map(_ => List.empty)
MyTC(v1).map(_ => false)
MyTC(v2).map(_ => 10L)
That would print:
inside sub-AnyRef
inside sub-AnyRef
inside sub-AnyVal
inside sub-AnyVal
The advantage of this approach is that, should you then choose to specialise the behaviour further for just some specific type (e.g. say you want to do something specific for Option[String]), you can do that easily:
// This is added to MyTC object
implicit val optMyTc : MyTC[Option[String]] = new MyTC[Option[String]]{
def map(f: Int => Option[String]): List[Option[String]] = { println("inside Option[String]"); List.empty }
}
Then, re-running the code will print:
inside Option[String]
inside sub-AnyRef
inside sub-AnyVal
inside sub-AnyVal
While learning shapeless via this useful guide
I was trying to merge what I have learned from this in this exercice
The result is this code :
import shapeless._
object GenericV3 {
trait CsvEncoder[A] {
def width: Int
def encode(value: A): List[String]
}
def createEncoder[A](cols: Int)(func: A => List[String]): CsvEncoder[A] =
new CsvEncoder[A] {
val width = cols
def encode(value: A): List[String] =
func(value)
}
implicit val stringEncoder: CsvEncoder[String] =
createEncoder(1)(str => List(str))
implicit val intEncoder: CsvEncoder[Int] =
createEncoder(1)(num => List(num.toString))
implicit val booleanEncoder: CsvEncoder[Boolean] =
createEncoder(1)(bool => List(if(bool) "cone" else "glass"))
implicit val doubleEncoder: CsvEncoder[Double] =
createEncoder(1)(d => List(d.toString))
import shapeless.{HList, HNil, ::}
implicit val hnilEncoder: CsvEncoder[HNil] =
createEncoder(0)(hnil => Nil)
implicit def hlistEncoder[H, T <: HList](
implicit
hEncoder: Lazy[CsvEncoder[H]],
tEncoder: CsvEncoder[T]
): CsvEncoder[H :: T] =
createEncoder(hEncoder.value.width + tEncoder.width) {
case h :: t =>
hEncoder.value.encode(h) ++ tEncoder.encode(t)
}
import shapeless.{Coproduct, CNil, :+:, Inl, Inr}
implicit val cnilEncoder: CsvEncoder[CNil] =
createEncoder(0) { cnil =>
throw new Exception("The impossible has happened!")
}
implicit def coproductEncoder[H, T <: Coproduct](
implicit
hEncoder: Lazy[CsvEncoder[H]],
tEncoder: CsvEncoder[T]
): CsvEncoder[H :+: T] =
createEncoder(hEncoder.value.width + tEncoder.width) {
case Inl(h) => hEncoder.value.encode(h) ++ List.fill(tEncoder.width)("")
case Inr(t) => List.fill(hEncoder.value.width)("") ++ tEncoder.encode(t)
}
import shapeless.Generic
implicit def genericEncoder[A, R](
implicit
gen: Generic.Aux[A, R],
lEncoder: Lazy[CsvEncoder[R]]
): CsvEncoder[A] =
createEncoder(lEncoder.value.width) { value =>
lEncoder.value.encode(gen.to(value))
}
def writeCsv[A](values: List[A])(implicit encoder: CsvEncoder[A]): String =
values.map(encoder.encode).map(_.mkString(",")).mkString("\n")
}
when calling GenericV3.writeCsv with a sealed case class that has a case containing the sealed case class itself.
the example from the guide :
sealed trait Tree[A]
final case class Branch[A](left: Tree[A], right: Tree[A]) extends Tree[A]
final case class Leaf[A](value: A) extends Tree[A]
I got a StackOverflowError at runtime.
Exemple
GenericV3.writeCsv[Tree[String]](List(Branch(Leaf("toto"),Leaf("titi"))))
The error is produced because I call hEncoder.value.width before matching Inl and Inr.
I tried to read how it's done in other libs.
And I have seen this which looks the same.
So I think I missed something.
ps : I would like to thank to the author of this guide it's very helpful
I'm working on an API that should enable to build a shapeless Poly1 function dynamically from standard monomorphic functions that operate on types of some coproduct.
The goal is to expose a simple method that receives a function as:
type FooCoproduct = Foo :+: Bar :+: CNil
def addF[E](f: E => E)(implicit ev: Inject[FooCoproduct, E]) = ???
and accumulate these functions in order to build a total Poly1 function covering all types in the coproduct. The evidence ev here is to force that the type paremeter E is a type in the coproduct.
After testing several approaches, including generic derivation of typeclasses, the most promising one has led me to accumulate these monomorphic functions in an HList and try to resolve the one that applies by means of a Selector. This is probably better understood by example:
object CoproductSample extends App {
import shapeless.{ :+:, CNil, Coproduct, HList, HNil, Poly1 }
import shapeless.ops.coproduct.Inject
import shapeless.ops.hlist.Selector
class Builder[A <: Coproduct] {
def accum[B](f: B => B, hl: HList)(implicit ev: Inject[A, B]) = f :: hl
class PolyBuilder[L <: HList](hl: L) extends Poly1 {
implicit def run[T](implicit ev: Selector[L, T => T]) =
at[T](hl.select[T => T])
}
}
type Cop = Int :+: String :+: CNil
val builder = new Builder[Cop]
val hl1 = builder.accum((i: Int) => i + 1, HNil)
val hl2 = builder.accum((s: String) => s + "one", hl1)
object pf extends builder.PolyBuilder(hl2)
val rInt = Coproduct[Cop](10).fold(pf)
val rStr = Coproduct[Cop]("ten").fold(pf)
}
This code doesn't compile with the message:
could not find implicit value for parameter folder:
shapeless.ops.coproduct.Folder[CoproductSample.pf.type, CoproductSample.Cop]
I suppose that I need to provide a Selector[L, T => T] where L is the type of the accumulated HList but I can't come up with the way to do this. On the other hand, I have the feeling that there must be a simpler solution to my problem.
Any help would be appreciated.
Update
After doing some more research I've come up with a solution that almost works. Unfortunately I'm not able to track the result type properly.
object CoproductSample {
import shapeless.{ CNil, Coproduct, HList, HNil, Inl, Inr, Poly2, ::, :+: }
// Accumulates ordinary functions from A => A in an HList
def accum[A, L <: HList](f: A => A, hl: L): (A => A) :: L = f :: hl
// A poly2 function that evaluates some monomorphic function present in
// an HList for certain value that satifies the signature of this function
object PolyEval extends Poly2 {
implicit def hnilCase[A]: Case.Aux[A, HNil, Option[A]] =
at[A, HNil]((a, l) => None)
implicit def hheadCaseSuccess[A, T <: HList]: Case.Aux[A, (A => A) :: T, Option[A]] =
at[A, (A => A) :: T]((a: A, l: (A => A) :: T) => Option(l.head(a)))
implicit def hheadCaseFail[A, H, T <: HList](
implicit tail: Case.Aux[A, T, Option[A]]
): Case.Aux[A, (H => H) :: T, Option[A]] =
at[A, (H => H) :: T]((a: A, l: (H => H) :: T) => PolyEval(a, l.tail))
}
// A poly2 function that uses `PolyEval` for evaluating a value present in
// a coproduct against an HList of monomorphic functions
object PolyEvalCop extends Poly2 {
implicit def cnilCase[A <: CNil, L <: HList]: Case.Aux[A, L, Option[A]] =
at[A, L]((a, l) => sys.error("Impossible!"))
implicit def cconsCase[H, T <: Coproduct, L <: HList](
implicit head: PolyEval.Case.Aux[H, L, Option[H]],
tail: Case[T, L]) // What is the return type here???)
= at[H :+: T, L]((c, l) =>
c match {
case Inl(h) => PolyEval(h, l)
case Inr(t) => PolyEvalCop(t, l)
})
}
}
Console session:
scala> import shapeless._, CoproductSample._
import shapeless._
import CoproductSample._
scala> case class Foo(i: Int); case class Bar(s: String)
defined class Foo
defined class Bar
scala> val f = (foo: Foo) => foo.copy(i = foo.i * 2)
f: Foo => Foo = <function1>
scala> val g = (bar: Bar) => bar.copy(s = bar.s + "_changed!")
g: Bar => Bar = <function1>
scala> val hl = accum(g, accum(f, HNil))
hl: shapeless.::[Bar => Bar,shapeless.::[Foo => Foo,shapeless.HNil.type]] = <function1> :: <function1> :: HNil
scala> type C = Foo :+: Bar :+: CNil
defined type alias C
scala> PolyEvalCop(Coproduct[C](Foo(10)), hl)
res1: Any = Some(Foo(20))
scala> PolyEvalCop(Coproduct[C](Bar("bar")), hl)
res2: Any = Some(Bar(bar_changed!))
The result type is not properly tracked and it's resolved as Any.
From the signature of addF it looks like you want to map over the Coproduct, i.e. modify it's value and stay on a coproduct, i.e. def add[E](e:E):E, if that were the case this would work:
# {
trait Add[E]{
def add(e:E):E
}
object Add{
def apply[E:Add]:Add[E] = implicitly[Add[E]]
implicit object cnil extends Add[CNil] {
def add(e:CNil) = throw new RuntimeException("Impossible")
}
implicit def coproduct[H, T <: Coproduct](
implicit
addH: Add[H],
addT:Add[T],
basis: ops.coproduct.Basis[H :+: T,T]
):Add[H :+: T] = new Add[H :+: T]{
def add(e: H :+: T) = e match {
case Inl(h) => Coproduct[H :+: T](addH.add(h)) // to stay in the Coproduct
case Inr(t) => addT.add(t).embed[H :+: T] // to stay in the coproduct
}
}
}
}
defined trait Add
defined object Add
# implicit def addString = new Add[String] {
def add(e:String) = e + "-ah"
}
defined function addString
# implicit def addInt = new Add[Int] {
def add(e:Int) = e + 1
}
defined function addInt
# type C = Int :+: String :+: CNil
defined type C
# val i = Coproduct[C](1)
i: C = 1
# Add[C].add(i)
res24: C = 2 // notice that the return type is C
# val s = Coproduct[C]("a")
s: C = a
# Add[C].add(s)
res26: C = a-ah // notice that the return type is C
It obviously works with "plain" types:
# Add[Int].add(1)
res38: Int = 2
The above is equivalent to map; but if you want a fold, i.e. def add[E](e:E):Int, you would just modify these two lines:
case Inl(h) => addH.add(h)
case Inr(t) => addT.add(t)
New to shapeless and I have a question on using polymorphic functions that need some dependencies. I basically have this code and want to pull somePoly object out of the run method:
import shapeless._
object SomeObject {
type SomeType = Int :+: String :+: (String, Int) :+: CNil
def run( someList: List[SomeType], someInt:Int, someWord:String ) = {
object somePoly extends Poly1 {
implicit def doIt = at[Int]( i => i + someInt + someWord.length)
implicit def doIt2 = at[String]( i => i.length + someWord.length)
implicit def doIt3 = at[(String, Int)]( i => i._1.length + someWord.length)
}
someList.map( _.map(somePoly) )
}
}
One way I thought of doing it was like this, but it seems messy:
object TypeContainer {
type SomeType = Int :+: String :+: (String, Int) :+: CNil
}
case class SomePolyWrapper( someList: List[TypeContainer.SomeType], someInt:Int, someWord:String ){
object somePoly extends Poly1 {
implicit def doIt = at[Int]( i => i + someInt + someWord.length)
implicit def doIt2 = at[String]( i => i.length + someWord.length)
implicit def doIt3 = at[(String, Int)]( i => i._1.length + someWord.length)
}
}
object SomeObject {
def run( someList: List[TypeContainer.SomeType], someInt:Int, someWord:String ) = {
val somePolyWrapper = SomePolyWrapper(someList, someInt, someWord)
someList.map( _.map(somePolyWrapper.somePoly) )
}
}
Anyone have any advice?
The limitations of Scala's implicit resolution system mean the Poly definition needs to be a stable identifier, which makes this kind of thing more painful than it should be. As I mentioned on Gitter, there are a couple of workarounds that I know of (there may be others).
One approach would be to make the Poly1 a PolyN, where the extra arguments are for the someInt and someWord values. If you were mapping over an HList, you'd then use mapConst and zip to make the input HList have the right shape. I've never done this for a coproduct, but something similar is likely to work.
Another approach is to use a custom type class. In your case that might look something like this:
import shapeless._
trait IntFolder[C <: Coproduct] {
def apply(i: Int, w: String)(c: C): Int
}
object IntFolder {
implicit val cnilIntFolder: IntFolder[CNil] = new IntFolder[CNil] {
def apply(i: Int, w: String)(c: CNil): Int = sys.error("Impossible")
}
def instance[H, T <: Coproduct](f: (H, Int, String) => Int)(implicit
tif: IntFolder[T]
): IntFolder[H :+: T] = new IntFolder[H :+: T] {
def apply(i: Int, w: String)(c: H :+: T): Int = c match {
case Inl(h) => f(h, i, w)
case Inr(t) => tif(i, w)(t)
}
}
implicit def iif[T <: Coproduct: IntFolder]: IntFolder[Int :+: T] =
instance((h, i, w) => h + i + w.length)
implicit def sif[T <: Coproduct: IntFolder]: IntFolder[String :+: T] =
instance((h, i, w) => h.length + i + w.length)
implicit def pif[T <: Coproduct: IntFolder]: IntFolder[(String, Int) :+: T] =
instance((h, i, w) => h._1.length + i + w.length)
}
And then you could write a more generic version of your run:
def run[C <: Coproduct](
someList: List[C],
someInt: Int,
someWord: String
)(implicit cif: IntFolder[C]): List[Int] = someList.map(cif(someInt, someWord))
And use it like this:
scala> run(List(Coproduct[SomeType](1)), 10, "foo")
res0: List[Int] = List(14)
scala> run(List(Coproduct[SomeType](("bar", 1))), 10, "foo")
res1: List[Int] = List(16)
The specificity of the operation makes this approach look a little weird, but if I really needed to do something like this for different coproducts, this is probably the solution I'd choose.
A type level foldRight works fine (getLabelWithValues), and a follow-on type level map (getValues) also works fine. If I combine both in one method (getValuesFull), it doesn't work any more though. What is the missing piece?
The full source (with sbt ready to ~run with implicit debug output) is here: https://github.com/mpollmeier/shapeless-playground/tree/8170a5b
case class Label[A](name: String)
case class LabelWithValue[A](label: Label[A], value: A)
val label1 = Label[Int]("a")
val labels = label1 :: HNil
object combineLabelWithValue extends Poly2 {
implicit def atLabel[A, B <: HList] = at[Label[A], (B, Map[String, Any])] {
case (label, (acc, values)) ⇒
(LabelWithValue(label, values(label.name).asInstanceOf[A]) :: acc, values)
}
}
object GetLabelValue extends (LabelWithValue ~> Id) {
def apply[B](labelWithValue: LabelWithValue[B]) = labelWithValue.value
}
val labelsWithValues: LabelWithValue[Int] :: HNil = getLabelWithValues(labels)
// manually mapping it works fine:
val valuesManual: Int :: HNil = labelsWithValues.map(GetLabelValue)
// using a second function with Mapper works fine:
val valuesSecondFn: Int :: HNil = getValues(labelsWithValues)
// error: could not find implicit value for parameter mapper: shapeless.ops.hlist.Mapper.Aux[Main.GetLabelValue.type,WithValues,Values]
// val valuesFull: Int :: HNil = getValuesFull(labels)
def getLabelWithValues[L <: HList, P, WithValues](labels: L)(
implicit folder: RightFolder.Aux[L, (HNil.type, Map[String, Any]), combineLabelWithValue.type, P],
ic: IsComposite.Aux[P, WithValues, _]
): WithValues = {
val state = Map("a" -> 5, "b" -> "five")
val resultTuple = labels.foldRight((HNil, state))(combineLabelWithValue)
ic.head(resultTuple)
}
def getValues[WithValues <: HList, Values <: HList](withValues: WithValues)(
implicit mapper: Mapper.Aux[GetLabelValue.type, WithValues, Values]
): Values = {
withValues.map(GetLabelValue)
}
def getValuesFull[L <: HList, P, WithValues <: HList, Values <: HList](labels: L)(
implicit folder: RightFolder.Aux[L, (HNil.type, Map[String, Any]), combineLabelWithValue.type, P],
ic: IsComposite.Aux[P, WithValues, _],
mapper: Mapper.Aux[GetLabelValue.type, WithValues, Values]
): Values = {
val state = Map("a" -> 5, "b" -> "five")
val resultTuple = labels.foldRight((HNil, state))(combineLabelWithValue)
val withValues: WithValues = ic.head(resultTuple)
withValues.map(GetLabelValue)
}
The issue here is that you're ending up trying to map over an HList where the HNil is statically typed as HNil.type. This doesn't work in general—e.g. in a simplified case like this:
import shapeless._, ops.hlist.Mapper
val mapped1 = Mapper[poly.identity.type, HNil]
val mapped2 = Mapper[poly.identity.type, HNil.type]
mapped1 will compile, but mapped2 won't.
The trick is to change the HNil.type in your RightFolder types to HNil and then to call foldRight with HNil: HNil. This will make everything work just fine.
There are a few other suggestions I'd make (destructure the tuple in place of P instead of using IsComposite, skip the Aux on mapper and return mapper.Out instead of having a Value type parameter, etc.), but they're probably out of the scope of this question.