Haskell IntroductionData.Aeson - JSON in HaskellHaskell Applicative FunctorHaskell Arbitrary-rank polymorphism with RankNTypesHaskell ArithmeticHaskell ArrowsHaskell AttoparsecHaskell BifunctorHaskell CabalHaskell Category TheoryHaskell Common functors as the base of cofree comonadsHaskell Common GHC Language ExtensionsHaskell Common monads as free monadsHaskell ConcurrencyHaskell Containers - Data.MapHaskell Creating Custom Data TypesHaskell Data.TextHaskell DatabasesHaskell Date and TimeHaskell Fixity declarationsHaskell FoldableHaskell Foreign Function InterfaceHaskell Free MonadsHaskell Function call syntaxHaskell Function compositionHaskell FunctorHaskell Generalized Algebraic Data TypesHaskell GHCJSHaskell Google Protocol BuffersHaskell Graphics with GlossHaskell Gtk3Haskell Higher-order functionsHaskell Infix operatorsHaskell IOHaskell LensHaskell List ComprehensionsHaskell ListsHaskell LoggingHaskell ModulesHaskell Monad TransformersHaskell MonadsHaskell MonoidHaskell OptimizationHaskell Overloaded LiteralsHaskell ParallelismHaskell Parsing HTML with taggy-lens and lensHaskell Partial ApplicationHaskell Phantom typesHaskell PipesHaskell ProfunctorHaskell ProxiesHaskell QuickCheckHaskell Reactive-bananaHaskell Reader / ReaderTHaskell Record SyntaxHaskell Recursion SchemesHaskell Rewrite rules (GHC)Haskell RoleHaskell Sorting AlgorithmsHaskell StackHaskell State MonadHaskell Streaming IOHaskell StrictnessHaskell Syntax in FunctionsHaskell Testing with TastyHaskell TraversableHaskell Tuples (Pairs, Triples, ...)Haskell Type algebraHaskell Type ApplicationHaskell Type ClassesHaskell Type FamiliesHaskell Typed holesHaskell Using GHCiHaskell VectorsHaskell Web DevelopmentHaskell XMLHaskell zipWithMTemplate Haskell & QuasiQuotes

Haskell Lens

From WikiOD

Lens is a library for Haskell that provides lenses, isomorphisms, folds, traversals, getters and setters, which exposes a uniform interface for querying and manipulating arbitrary structures, not unlike Java's accessor and mutator concepts.

Remarks[edit | edit source]

What is a Lens?[edit | edit source]

Lenses (and other optics) allow us to separate describing how we want to access some data from what we want to do with it. It is important to distinguish between the abstract notion of a lens and the concrete implementation. Understanding abstractly makes programming with lens much easier in the long run. There are many isomorphic representations of lenses so for this discussion we will avoid any concrete implementation discussion and instead give a high-level overview of the concepts.

Focusing[edit | edit source]

An important concept in understanding abstractly is the notion of focusing. Important optics focus on a specific part of a larger data structure without forgetting about the larger context. For example, the lens _1 focuses on the first element of a tuple but doesn't forget about what was in the second field.

Once we have focus, we can then talk about which operations we are allowed to perform with a lens. Given a Lens s a which when given a datatype of type s focuses on a specific a, we can either

  1. Extract the a by forgetting about the additional context or
  2. Replace the a by providing a new value

These correspond to the well-known get and set operations which are usually used to characterise a lens.

Other Optics[edit | edit source]

We can talk about other optics in a similar fashion.

Optic Focuses on...
Lens One part of a product
Prism One part of a sum
Traversal Zero or more parts of a data structure
Isomorphism ...

Each optic focuses in a different way, as such, depending on which type of optic we have we can perform different operations.

Composition[edit | edit source]

What's more, we can compose any of the two optics we have so-far discussed in order to specify complex data accesses. The four types of optics we have discussed form a lattice, the result of composing two optics together is their upper bound.


For example, if we compose together a lens and a prism, we get a traversal. The reason for this is that by their (vertical) composition, we first focus on one part of a product and then on one part of a sum. The result being an optic which focuses on precisely zero or one parts of our data which is a special case of a traversal. (This is also sometimes called an affine traversal).

In Haskell[edit | edit source]

The reason for the popularity in Haskell is that there is a very succinct representation of optics. All optics are just functions of a certain form which can be composed together using function composition. This leads to a very light-weight embedding which makes it easy to integrate optics into your programs. Further to this, due to the particulars of the encoding, function composition also automatically computes the upper bound of two optics we compose. This means that we can reuse the same combinators for different optics without explicit casting.

Lenses for records[edit | edit source]

Simple record[edit | edit source]

{*# LANGUAGE TemplateHaskell #-}
import Control.Lens

data Point = Point {
    _x :: Float,
    _y :: Float
makeLenses ''Point

Lenses x and y are created.

let p = Point 5.0 6.0 
p ^. x     -- returns 5.0
set x 10 p -- returns Point { _x = 10.0, _y = 6.0 }
p & x +~ 1 -- returns Point { _x = 6.0, _y = 6.0 }

Managing records with repeating fields names[edit | edit source]

data Person = Person { _personName :: String }
makeFields ''Person

Creates a type class HasName, lens name for Person, and makes Person an instance of HasName. Subsequent records will be added to the class as well:

data Entity = Entity { _entityName :: String }
makeFields ''Entity

The Template Haskell extension is required for makeFields to work. Technically, it's entirely possible to create the lenses made this way via other means, e.g. by hand.

Manipulating tuples with Lens[edit | edit source]


("a", 1) ^. _1 -- returns "a"
("a", 1) ^. _2 -- returns 1


("a", 1) & _1 .~ "b" -- returns ("b", 1)


("a", 1) & _2 %~ (+1) -- returns ("a", 2)

both Traversal

(1, 2) & both *~ 2 -- returns (2, 4)

Lens and Prism[edit | edit source]

A Lens' s a means that you can always find an a within any s. A Prism' s a means that you can sometimes find that s actually just is a but sometimes it's something else.

To be more clear, we have _1 :: Lens' (a, b) a because any tuple always has a first element. We have _Just :: Prism' (Maybe a) a because sometimes Maybe a is actually an a value wrapped in Just but sometimes it's Nothing.

With this intuition, some standard combinators can be interpreted parallel to one another

  • view :: Lens' s a -> (s -> a) "gets" the a out of the s
  • set :: Lens' s a -> (a -> s -> s) "sets" the a slot in s
  • review :: Prism' s a -> (a -> s) "realizes" that an a could be an s
  • preview :: Prism' s a -> (s -> Maybe a) "attempts" to turn an s into an a.

Another way to think about it is that a value of type Lens' s a demonstrates that s has the same structure as (r, a) for some unknown r. On the other hand, Prism' s a demonstrates that s has the same structure as Either r a for some r. We can write those four functions above with this knowledge:

** `Lens' s a` is no longer supplied, instead we just *know* that `s ~ (r, a)`

view :: (r, a) -> a
view (r, a) = a

set :: a -> (r, a) -> (r, a)
set a (r, _) = (r, a)
** `Prism' s a` is no longer supplied, instead we just *know* that `s ~ Either r a`

review :: a -> Either r a
review a = Right a

preview :: Either r a -> Maybe a
preview (Left _) = Nothing
preview (Right a) = Just a

Stateful Lenses[edit | edit source]

Lens operators have useful variants that operate in stateful contexts. They are obtained by replacing ~ with = in the operator name.

(+~) :: Num a => ASetter s t a a -> a -> s -> t
(+=) :: (MonadState s m, Num a) => ASetter' s a -> a -> m ()

Note: The stateful variants aren't expected to change the type, so they have the Lens' or Simple Lens' signatures.

Getting rid of & chains[edit | edit source]

If lens-ful operations need to be chained, it often looks like this:

change :: A -> A
change a = a & lensA %~ operationA
             & lensB %~ operationB
             & lensC %~ operationC

This works thanks to the associativity of &. The stateful version is clearer, though.

change a = flip execState a $ do
    lensA %= operationA
    lensB %= operationB
    lensC %= operationC

If lensX is actually id, the whole operation can of course be executed directly by just lifting it with modify.

Imperative code with structured state[edit | edit source]

Assuming this example state:

data Point = Point { _x :: Float, _y :: Float }
data Entity = Entity { _position :: Point, _direction :: Float }
data World = World { _entities :: [Entity] }

makeLenses ''Point
makeLenses ''Entity
makeLenses ''World

We can write code that resembles classic imperative languages, while still allowing us to use benefits of Haskell:

updateWorld :: MonadState World m => m ()
updateWorld = do
    -- move the first entity
    entities . ix 0 . position . x += 1

    -- do some operation on all of them
    entities . traversed . position %= \p -> p `pointAdd` ...

    -- or only on a subset
    entities . traversed . filtered (\e -> e ^. position.x > 100) %= ...

Lenses compose[edit | edit source]

If you have a f :: Lens' a b and a g :: Lens' b c then f . g is a Lens' a c gotten by following f first and then g. Notably:

  • Lenses compose as functions (really they just are functions)
  • If you think of the view functionality of Lens, it seems like data flows "left to right"—this might feel backwards to your normal intuition for function composition. On the other hand, it ought to feel natural if you think of .-notation like how it happens in OO languages.

More than just composing Lens with Lens, (.) can be used to compose nearly any "Lens-like" type together. It's not always easy to see what the result is since the type becomes tougher to follow, but you can use the lens chart to figure it out. The composition x . y has the type of the least-upper-bound of the types of both x and y in that chart.

Writing a lens without Template Haskell[edit | edit source]

To demystify Template Haskell, suppose you have

data Example a = Example { _foo :: Int, _bar :: a }


makeLenses 'Example

produces (more or less)

foo :: Lens' (Example a) Int
bar :: Lens (Example a) (Example b) a b

There's nothing particularly magical going on, though. You can write these yourself:

foo :: Lens' (Example a) Int
**  :: Functor f => (Int -> f Int) -> (Example a -> f (Example a))    ;; expand the alias
foo wrap (Example foo bar) = fmap (\newFoo -> Example newFoo bar) (wrap foo)

bar :: Lens (Example a) (Example b) a b
**  :: Functor f => (a -> f b) -> (Example a -> f (Example b))    ;; expand the alias
bar wrap (Example foo bar) = fmap (\newBar -> Example foo newBar) (wrap bar)

Essentially, you want to "visit" your lens' "focus" with the wrap function and then rebuild the "entire" type.

Traversals[edit | edit source]

A Traversal' s a shows that s has 0-to-many as inside of it.

toListOf :: Traversal' s a -> (s -> [a])

Any type t which is Traversable automatically has that traverse :: Traversal (t a) a.

We can use a Traversal to set or map over all of these a values

> set traverse 1 [1..10]

> over traverse (+1) [1..10]

A f :: Lens' s a says there's exactly one a inside of s. A g :: Prism' a b says there are either 0 or 1 bs in a. Composing f . g gives us a Traversal' s b because following f and then g shows how there there are 0-to-1 bs in s.

Classy Lenses[edit | edit source]

In addition to the standard makeLenses function for generating Lenses, Control.Lens.TH also offers the makeClassy function. makeClassy has the same type and works in essentially the same way as makeLenses, with one key difference. In addition to generating the standard lenses and traversals, if the type has no arguments, it will also create a class describing all the datatypes which possess the type as a field. For example

data Foo = Foo { _fooX, _fooY :: Int }
  makeClassy ''Foo

will create

class HasFoo t where
   foo :: Simple Lens t Foo

instance HasFoo Foo where foo = id

fooX, fooY :: HasFoo t => Simple Lens t Int

Fields with makeFields[edit | edit source]

(This example copied from this StackOverflow answer)

Let's say you have a number of different data types that all ought to have a lens with the same name, in this case capacity. The makeFields slice will create a class that accomplish this without namespace conflicts.

{*# LANGUAGE FunctionalDependencies
           , MultiParamTypeClasses
           , TemplateHaskell

module Foo

import Control.Lens

data Foo
  = Foo { fooCapacity :: Int }
  deriving (Eq, Show)
$(makeFields ''Foo)

data Bar
  = Bar { barCapacity :: Double }
  deriving (Eq, Show)
$(makeFields ''Bar)

Then in ghci:

λ» let f = Foo 3
|     b = Bar 7
b :: Bar
f :: Foo
λ» fooCapacity f
it :: Int
λ» barCapacity b
it :: Double
λ» f ^. capacity
it :: Int
λ» b ^. capacity
it :: Double

λ» :info HasCapacity 
class HasCapacity s a | s -> a where
  capacity :: Lens' s a
    -- Defined at Foo.hs:14:3
instance HasCapacity Foo Int -- Defined at Foo.hs:14:3
instance HasCapacity Bar Double -- Defined at Foo.hs:19:3

So what it's actually done is declared a class HasCapacity s a, where capacity is a Lens' from s to a (a is fixed once s is known). It figured out the name "capacity" by stripping off the (lowercased) name of the data type from the field; I find it pleasant not to have to use an underscore on either the field name or the lens name, since sometimes record syntax is actually what you want. You can use makeFieldsWith and the various lensRules to have some different options for calculating the lens names.

In case it helps, using ghci -ddump-splices Foo.hs:

[1 of 1] Compiling Foo              ( Foo.hs, interpreted )
Foo.hs:14:3*18: Splicing declarations
    makeFields ''Foo
    class HasCapacity s a | s -> a where
      capacity :: Lens' s a
    instance HasCapacity Foo Int where
      {-# INLINE capacity #-}
      capacity = iso (\ (Foo x_a7fG) -> x_a7fG) Foo
Foo.hs:19:3*18: Splicing declarations
    makeFields ''Bar
    instance HasCapacity Bar Double where
      {-# INLINE capacity #-}
      capacity = iso (\ (Bar x_a7ne) -> x_a7ne) Bar
Ok, modules loaded: Foo.

So the first splice made the class HasCapcity and added an instance for Foo; the second used the existing class and made an instance for Bar.

This also works if you import the HasCapcity class from another module; makeFields can add more instances to the existing class and spread your types out across multiple modules. But if you use it again in another module where you haven't imported the class, it'll make a new class (with the same name), and you'll have two separate overloaded capacity lenses that are not compatible.