Add Function1 #137
Add Function1 #137
Conversation
pakoito
requested review from
wiyarmir,
raulraja,
ffgiraldez and
JorgeCastilloPrz
Codecov Report
@@ Coverage Diff @@
## master #137 +/- ##
===========================================
+ Coverage 51.98% 52.09% +0.1%
- Complexity 160 163 +3
===========================================
Files 66 68 +2
Lines 1458 1478 +20
Branches 175 175
===========================================
+ Hits 758 770 +12
- Misses 629 637 +8
Partials 71 71
Continue to review full report at Codecov.
|
| operator fun <P> invoke(p: P): R | ||
| } | ||
|
|
||
| interface FunctionSurject<in P> { |
There was a problem hiding this comment.
where is needed this interface?
There was a problem hiding this comment.
Nowhere at the moment. It will, as it's dual of the Inject.
| override fun <A> pure(a: A): Function1<P, A> = | ||
| { _: P -> a }.k() | ||
|
|
||
| override fun <A, B> map(fa: HK<Function1.F, A>, f: (A) -> B): HK<Function1.F, B> = |
There was a problem hiding this comment.
return type could be Function1<F,B> ?
There was a problem hiding this comment.
Same as above, the parameter info isn't available to the compiler. It infers Function1<B, B>.
| Monad<Function1.F>, | ||
| MonadReader<Function1.F, P> { | ||
|
|
||
| override fun ask(): HK<Function1.F, P> = |
There was a problem hiding this comment.
what the reason to not return Function1<F,P>? not expose to callers FunctionInject?
There was a problem hiding this comment.
The compiler doesn't have enough information about the generic F in this context, so it fails.
| @@ -0,0 +1,9 @@ | |||
| package kategory | |||
|
|
|||
| interface FunctionInject<out R> { | |||
There was a problem hiding this comment.
we want those Bijections to be public?
There was a problem hiding this comment.
Yes, we have very few private things on the repo. It's all tools, really.
| fun <R> Function1F<R>.ev(): FunctionInject<R> = | ||
| this as FunctionInject<R> | ||
|
|
||
| // We don't we want an inherited class to avoid equivalence issues, so a simple HK wrapper will do |
| @RunWith(KTestJUnitRunner::class) | ||
| class Function1Test : UnitSpec() { | ||
| init { | ||
| testLaws(MonadLaws.laws(Function1.monad<Int>(), object : Eq<HK<Function1.F, Int>> { |
There was a problem hiding this comment.
Since this a new instance of MonadReader and so far we have Kleisli and this one I think it's time to add the missing MonadReader laws and apply them both here and in Kleisli, fine if addressed in a separate PR.
There was a problem hiding this comment.
Also just realized we are missing Category and MonoidK as typeclasses, also addressable in a separate PR. Function1 should also have an instance for MonoidK
| a.ev()(1) == b.ev()(1) | ||
| })) | ||
|
|
||
| "Function1Monad.binding should for comprehend over all values of multiple Function0" { |
There was a problem hiding this comment.
We may be able to add generic tests for Monad and Applicative to tests comprehensions and the applicative builder works for all monads and applicative instances instead of hardcoding these on each datatype test.
Another breakthrough in language limitations. This PR adds the first kind projection using skolemization!
The PR adds a HK wrapper for Functions with 1 parameter, along with an instance of monad() and monadReader()
For a background on the buzzwords, check these links:
https://github.com/non/kind-projector
https://gist.github.com/jdegoes/97459c0045f373f4eaf126998d8f65dc#skolemization