I want to define a new abstract data type which is either a general Number or a Division construct. How would I do that in Haskell?
My first approach was:
data MyMath = MyNum Num
| Div MyMath MyMath
The problem is that the compiler complains about "Num" which is not a data type but a type class. So my second thought would be to solve the problem like this:
data MyMath = MyNum Int
| MyNum Float
| Div MyMath MyMath
But this would not work either as MyNum is used twice which is not allowed, additionaly this approach would not really be polymorphic. So what is the solution to this problem?
EDIT2: After (again) reading the answers I tried to use GADT data constructors. This is some artificial example code:
5 data MyMathExpr a where
6 MyNumExpr :: Num a => a -> MyMathExpr a
7 MyAddExpr :: MyMathExpr b -> MyMathExpr c -> MyMathExpr (b, c)
8 deriving instance Show(MyMathExpr a)
9 deriving instance Eq(MyMathExpr a)
10
11 data MyMathVal a where
12 MyMathVal :: Num a => a -> MyMathVal a
13 deriving instance Show(MyMathVal a)
14 deriving instance Eq(MyMathVal a)
15
16 foo :: MyMathExpr a -> MyMathVal a
17 foo (MyNumExpr num) = MyMathVal num
18 foo (MyAddExpr num1 num2) = MyMathVal (l + r)
19 where (MyMathVal l) = foo num1
20 (MyMathVal r) = foo num2
But something is wrong with line number 18:
test.hs:18:40:
Couldn't match type `b' with `(b, c)'
`b' is a rigid type variable bound by
a pattern with constructor
MyAddExpr :: forall b c.
MyMathExpr b -> MyMathExpr c -> MyMathExpr (b, c),
in an equation for `foo'
at test.hs:18:6
In the first argument of `(+)', namely `l'
In the first argument of `MyMathVal', namely `(l + r)'
In the expression: MyMathVal (l + r)
The same goes for `c'. I guess it's a stupid mistake which I just don't see. Do you?
This solves the problem you're addressing in the code, but doesn't cover for the boolean.
If you want to use a class constraint in a data declaration, you do it the way that you would with any other function:
data (Num a) => MyMath a = MyMath {x :: a}
You can use existential quantification for that:
> let data MyMath = forall n. Num n => MyNum n
> :t MyNum 3
MyNum 3 :: MyMath
> :t MyNum 3.5
MyNum 3.5 :: MyMath
There are many ways to do it. One way is with GADTs:
{-# LANGUAGE GADTs #-}
data MyMath where
MyNum :: Num a => a -> MyMath
MyBool :: Bool -> MyMath
Another way with GADTs:
{-# LANGUAGE GADTs #-}
data MyMath a where
MyNum :: Num a => a -> MyMath a
MyBool :: Num a => Bool -> MyMath a
As has been mentioned, your attempts don't involve any booleans, but I'll go with the text of your question instead.
You don't have to invent this type, check out Either in the Prelude. What you're looking for is thus Either a Bool, where you want a to be an instance of Num. If you want to actually enforce this, ready the edit below.
Edit: If you don't want to use Either, you can do data MyMath a = MyNum a | MyBool Bool. Now you can enforce a being an instance of Num if you want to, but you may want to consider this SO question and this answer to it first. There's really no need to enforce the instance for the data type; just do it for the functions using it intsead.
Related
Suppose we have a record type that is homogeneous.
type RecI = { a :: Int, b :: Int, c :: Int, d :: Int, e :: Int }
We want to get from it type with the same keys but different value type:
type RecS = { a :: String, b :: String, c :: String, d :: String, e :: String }
Is it possible to get RecS type without explicitly defining all the keys from RecI?
And the second part of the question, what is the best way to implement mapping function from one type to another:
mapItoS :: (Int -> String) -> RecI -> RecS
?
To get a free-ish conversion from Int to String at type level, just give your record a parameter, then instantiate it with Int to get RecI and with String to get RecS:
type Rec a = { a :: a, b :: a, c :: a, d :: a, e :: a }
type RecI = Rec Int
type RecS = Rec String
To implement mapItoS, you can first convert to a Foreign.Object using fromHomogeneous, then map the function over it, then convert back to the record.
Unfortunately there is no toHomogeneous function, because in general you can't be sure that the Foreign.Object actually contains all required keys. But no matter: in this particular case you can be sure that it does, so you can get away with unsafeCoerce:
mapItoS :: forall a b. (a -> b) -> Rec a -> Rec b
mapItoS f = fromHomogeneous >>> map f >>> unsafeCoerce
A small self plug which is strictly relevant to the question :-P I've just published a library which provides many instances which allow a PureScripter to work with homogeneous Record and Variant:
https://pursuit.purescript.org/packages/purescript-homogeneous
I think it should have much better inference than solutions like heterogeneous. Please check it out and let me know what do you think.
Well, just simplified as possible:
There is a function that takes functor and does whatever
sToInt :: ∀ a s. Functor s => s a -> Int
sToInt val = unsafeCoerce val
Usage of this function with functor S which param (v) is functor too.
-- declare date type S that is functor
data S (v :: Type -> Type) a = S (v a)
instance functorS :: Functor v => Functor (S v) where
map f (S a) = S (map f a)
sV :: ∀ v a. S v a
sV = unsafeCoerce 1
sss :: Int
sss = sToInt sV -- get the error here
No type class instance was found for
Data.Functor.Functor t2
The instance head contains unknown type variables. Consider adding a type annotation.
while applying a function sToInt
of type Functor t0 => t0 t1 -> Int
to argument sV
while checking that expression sToInt sV
has type Int
in value declaration sss
where t0 is an unknown type
t1 is an unknown type
t2 is an unknown type
So it doesn't like S Functor instance has v param Functor constraint, I wonder why getting this error and how to fix it for this case.
This doesn't have to do with v or with the specific shape of S. Try this instead:
sV :: forall f a. Functor f => f a
sV = unsafeCoerce 1
sss :: Int
sss = sToInt sV
You get a similar error.
Or here's an even more simplified version:
sV :: forall a. a
sV = unsafeCoerce 1
sss :: Int
sss = sToInt sV
Again, same error.
The problem is that sToInt must get a Functor instance as a parameter (that's what the Functor s => bit in its type signature says), and in order to pick which Functor instance to pass, the compiler needs to know the type of the value. Like, if it's Maybe a, it will pass the Functor Maybe instance, and if it's Array a, it will pass the Functor Array instance, and so on.
Usually the type can be inferred from the context. For example when you say map show [1,2,3], the compiler knows that map should come from Functor Array, because [1,2,3] :: Array Int.
But in your case there is nowhere to get that information: sV can return S v for any v, and sToInt can also take any functor type. There is nothing to tell the compiler what the type should be.
And the way to fix this is obvious: if there is no context information for the compiler to get the type from, you have to tell it what the type is yourself:
sss :: Int
sss = sToInt (sV :: S Maybe _)
This will be enough for the compiler to know that v ~ Maybe, and it will be able to construct a Functor (S Maybe) instance and pass it to sToInt.
Alternatively, if you want the consumer of sss to decide what v is, you can add an extra dummy parameter to capture the type, and require that the consumer pass in a Functor v instance:
sss :: forall v. Functor v => FProxy v -> Int
sss _ = sToInt (sV :: S v _)
ddd :: Int
ddd = sss (FProxy :: FProxy Maybe)
In Haskell you can do this with visible type applications instead of FProxy, but PureScript, sadly, doesn't support that yet.
Even more alternatively, if sToInt doesn't actually care for a Functor instance, you can remove that constraint from it, and everything will work as-is:
sToInt :: forall s a. s a -> Int
sToInt a = unsafeCoerce a
sV :: forall v a. S v a
sV = unsafeCoerce 1
sss :: Int
sss = sToInt sV
This works because PureScript allows for ambiguous (aka "unknown") types to exist as long as they're not used for selecting instances.
Is there any way to do something like
first = {x:0}
second = {x:1,y:1}
both = [first, second]
such that both is inferred as {x::Int | r} or something like that?
I've tried a few things:
[{x:3}] :: Array(forall r. {x::Int|r}) -- nope
test = Nil :: List(forall r. {x::Int|r})
{x:1} : test -- nope
type X r = {x::Int | r}
test = Nil :: List(X) -- nope
test = Nil :: List(X())
{x:1} : test
{x:1, y:1} : test -- nope
Everything I can think of seems to tell me that combining records like this into a collection is not supported. Kind of like, a function can be polymorphic but a list cannot. Is that the correct interpretation? It reminds me a bit of the F# "value restriction" problem, though I thought that was just because of CLR restrictions whereas JS should not have that issue. But maybe it's unrelated.
Is there any way to declare the list/array to support this?
What you're looking for is "existential types", and PureScript just doesn't support those at the syntax level the way Haskell does. But you can roll your own :-)
One way to go is "data abstraction" - i.e. encode the data in terms of operations you'll want to perform on it. For example, let's say you'll want to get the value of x out of them at some point. In that case, make an array of these:
type RecordRep = Unit -> Int
toRecordRep :: forall r. { x :: Int | r } -> RecordRep
toRecordRep {x} _ = x
-- Construct the array using `toRecordRep`
test :: Array RecordRep
test = [ toRecordRep {x:1}, toRecordRep {x:1, y:1} ]
-- Later use the operation
allTheXs :: Array Int
allTheXs = test <#> \r -> r unit
If you have multiple such operations, you can always make a record of them:
type RecordRep =
{ getX :: Unit -> Int
, show :: Unit -> String
, toJavaScript :: Unit -> Foreign.Object
}
toRecordRep r =
{ getX: const r.x
, show: const $ show r.x
, toJavaScript: const $ unsafeCoerce r
}
(note the Unit arguments in every function - they're there for the laziness, assuming each operation could be expensive)
But if you really need the type machinery, you can do what I call "poor man's existential type". If you look closely, existential types are nothing more than "deferred" type checks - deferred to the point where you'll need to see the type. And what's a mechanism to defer something in an ML language? That's right - a function! :-)
newtype RecordRep = RecordRep (forall a. (forall r. {x::Int|r} -> a) -> a)
toRecordRep :: forall r. {x::Int|r} -> RecordRep
toRecordRep r = RecordRep \f -> f r
test :: Array RecordRep
test = [toRecordRep {x:1}, toRecordRep {x:1, y:1}]
allTheXs = test <#> \(RecordRep r) -> r _.x
The way this works is that RecordRep wraps a function, which takes another function, which is polymorphic in r - that is, if you're looking at a RecordRep, you must be prepared to give it a function that can work with any r. toRecordRep wraps the record in such a way that its precise type is not visible on the outside, but it will be used to instantiate the generic function, which you will eventually provide. In my example such function is _.x.
Note, however, that herein lies the problem: the row r is literally not known when you get to work with an element of the array, so you can't do anything with it. Like, at all. All you can do is get the x field, because its existence is hardcoded in the signatures, but besides the x - you just don't know. And that's by design: if you want to put anything into the array, you must be prepared to get anything out of it.
Now, if you do want to do something with the values after all, you'll have to explain that by constraining r, for example:
newtype RecordRep = RecordRep (forall a. (forall r. Show {x::Int|r} => {x::Int|r} -> a) -> a)
toRecordRep :: forall r. Show {x::Int|r} => {x::Int|r} -> RecordRep
toRecordRep r = RecordRep \f -> f r
test :: Array RecordRep
test = [toRecordRep {x:1}, toRecordRep {x:1, y:1}]
showAll = test <#> \(RecordRep r) -> r show
Passing the show function like this works, because we have constrained the row r in such a way that Show {x::Int|r} must exist, and therefore, applying show to {x::Int|r} must work. Repeat for your own type classes as needed.
And here's the interesting part: since type classes are implemented as dictionaries of functions, the two options described above are actually equivalent - in both cases you end up passing around a dictionary of functions, only in the first case it's explicit, but in the second case the compiler does it for you.
Incidentally, this is how Haskell language support for this works as well.
Folloing #FyodorSoikin answer based on "existential types" and what we can find in purescript-exists we can provide yet another solution.
Finally we will be able to build an Array of records which will be "isomorphic" to:
exists tail. Array { x :: Int | tail }
Let's start with type constructor which can be used to existentially quantify over a row type (type of kind #Type). We are not able to use Exists from purescript-exists here because PureScript has no kind polymorphism and original Exists is parameterized over Type.
newtype Exists f = Exists (forall a. f (a :: #Type))
We can follow and reimplement (<Ctrl-c><Ctrl-v> ;-)) definitions from Data.Exists and build a set of tools to work with such Exists values:
module Main where
import Prelude
import Unsafe.Coerce (unsafeCoerce)
import Data.Newtype (class Newtype, unwrap)
newtype Exists f = Exists (forall a. f (a :: #Type))
mkExists :: forall f a. f a -> Exists f
mkExists r = Exists (unsafeCoerce r :: forall a. f a)
runExists :: forall b f. (forall a. f a -> b) -> Exists f -> b
runExists g (Exists f) = g f
Using them we get the ability to build an Array of Records with "any" tail but we have to wrap any such a record type in a newtype before:
newtype R t = R { x :: Int | t }
derive instance newtypeRec :: Newtype (R t) _
Now we can build an Array using mkExists:
arr :: Array (Exists R)
arr = [ mkExists (R { x: 8, y : "test"}), mkExists (R { x: 9, z: 10}) ]
and process values using runExists:
x :: Array [ Int ]
x = map (runExists (unwrap >>> _.x)) arr
Suppose you are using the FlexibleInstances extension and have the class
class C a where
f :: a b -> Maybe b
how would you implement it for a list of lists of a datatype. In particular, how would the type be written. The only thing I could find is how to do it for a single list, but not a list of lists or lists of any other datatypes.
This works:
instance C [] where
...
But this doesn't
data D = ...
instance C [[D]] where
...
How can I express something like this?
You need a newtype
class C a where
f :: a b -> b -- the class before the OP edited
newtype LL a = LL [[a]]
instance C LL where
f (LL xss) = ...
However, it is impossible to write a completely meaningful instance, since if the lists-of-lists is empty, it is impossible to extract an element. the best we could do is
instance C LL where
f (LL xss) = case concat xss of
(x:_) -> x
_ -> error "f: no elements"
I'm not sure if that is a good idea.
As an alternative, you could use type families or functional dependencies. Here's a solution with type families.
{-# LANGUAGE TypeFamilies, FlexibleInstances #-}
class C a where
type T a
f :: a -> Maybe (T a)
instance C [[b]] where
type T [[b]] = b
f xss = case concat xss of
[] -> Nothing
(x:_) -> Just x
I would like to have a function that checks a property of a class instance:
class ListWithAtLeastOneElement a where
list :: [a]
check :: Bool
check = (length list) >= 1
but I get this error when compiling in GHC:
"The class method 'check' mentions none of the type variables of the class ListWithAtLeastOneElement a When checking the class method: check :: Bool In the class declaration for 'ListWithAtLeastOneElement'"
Is there a better way of doing what I want, or a way to get this to compile in GHC?
You seem to think a class in Haskell is like a class in an OO-language. That is not true. You should use a data type or newtype or type synonym.
newtype NonemptyList a = List1 [a]
fromList :: [a] -> Maybe (NonemptyList a)
fromList [] = Nothing
fromList xs#(_:_) = Just $ List1 xs
check :: NonemptyList a -> Bool
check (List1 xs) = length xs >= 1
As jmg said, this isn't valid Haskell so it's good GHC isn't compiling it! If you're used to Java evidently you should think of Haskell's type classes like Java interface. If that doesn't help then perhaps you should read LYAH's chapter on classes.
For you problem, it appears you're wanting a list-like data type that can never be null. You don't need to test for such a property, you can statically ensure it by using a data
type that can never be empty:
-- Notice this data type can never have zero 'a' values!
data NonEmptyList a = NEL a (NonEmptyList a) | Singleton a
-- We can define basic operators for this, just like list has
-- You can't pattern match with them, but there are work-arounds for that if you want to ask
(.:) = NEL -- concatenate non-empty lists
nelHead :: NonEmptyList a -> a
nelHead (NEL a _) = a
nelHead (Singleton a) = a
nelTail :: NonEmptyList a -> Maybe (NonEmptyList a)
nelTail (NEL _ b) = Just b
nelTail _ = Nothing
nelTake :: Int -> NonEmptyList a -> NonEmptyList a
nelTake 1 (NEL a _) = Singleton a
nelTake 1 (Singleton a) = Singleton a
nelTake n (NEL a rest) = a .: nelTake (n-1) rest
nelDrop :: Int -> NonEmptyList a -> NonEmptyList a
nelDrop _ (Singleton _) = error "Perhaps I should have used the 'maybe' type"
nelDrop 1 (NEL a r) = r
nelDrop n (NEL a r) = nelDrop (n-1) r
And so on and so forth. It's worth noting nelTake and nelDrop are partial but nelHead is total, funny as this is the opposite of regular lists.
You would need to make check a function
check :: [a] -> Bool
That said you are better off with a data type for non-empty lists rather than a class, classes in Haskell do not serve the same purpose as classes in object oriented languages.