Replace ~> with NaturalTransformation

[markus1189]

After having read NaturalTransformation.scala several times for reference, I found it much easier to read by replacing NaturalTransformation[A,B] with the much better A ~> B.

This pr replaces the occurrences where possible in NaturalTransformation.scala, Free.scala and FreeApplicative.scala.

I am sure there are more places where you could do this, but I first wanted to check if this change is welcome at all ;)

[codecov-io]

Current coverage is 81.99%

Merging #991 into master will decrease coverage by 6.54%

1. 6 files (not in diff) in ...main/scala/cats/free were deleted. more
2. 3 files (not in diff) in ...main/scala/cats/free were created. more
3. 13 files (not in diff) in ...cala/cats/kernel/std were modified. more
   • Misses +111
   • Hits -111
4. 5 files (not in diff) in ...in/scala/cats/kernel were modified. more
   • Misses +37
   • Hits -37
5. 2 files (not in diff) in ...ala/cats/kernel/laws were modified. more
   • Misses +25
   • Hits -25
6. 3 files (not in diff) in ...in/scala/cats/syntax were modified. more
   • Misses +4
   • Hits -5
7. 7 files (not in diff) in .../main/scala/cats/std were modified. more
   • Misses -11
   • Hits +2
8. 2 files (not in diff) in ...n/scala/cats/functor were modified. more
9. 8 files (not in diff) in ...main/scala/cats/data were modified. more
   • Hits -5
10. 6 files (not in diff) in .../src/main/scala/cats were modified. more
    • Misses +3
    • Hits -6

@@             master       #991   diff @@
==========================================
  Files           214        215     +1   
  Lines          2719       2704    -15   
  Methods        2655       2639    -16   
  Messages          0          0          
  Branches         59         60     +1   
==========================================
- Hits           2407       2217   -190   
- Misses          312        487   +175   
  Partials          0          0          

Powered by Codecov. Last updated by 271c197...a706712

[ceedubs]

@markus1189 thanks for bringing this up!

Personally, I kind of prefer ~> to NaturalTransformation as well. However, I think we've kind of settled on using the non-symbolic form of operators and types even when a symbolic form is provided as an alias. This tends to make it less bewildering and easier to google or ask questions about for newcomers. But I can't seem to find anywhere that we have documented this convention in the contributor guide. In fact, it looks like our contributor guide needs a lot of love. I've just created #993 for this.

[markus1189]

That's a good point, also currently there already is a mixture of NaturalTransformation and ~>, both in NaturalTransformation and throughout the rest of the code base.

I agree with you that it might be confusing to see ~> without knowing what that is supposed to be. So is the tendency to convert more ~> to NaturalTransformation instead of the other way around? I think at least inside NaturalTransformation having a consistent style would be beneficial (currently e.g. andThen uses NaturalTransformation while or uses ~>. I'd be happy to change the pr tomorrow to translate it the other way around then ;)

[ceedubs]

@markus1189 I tend to agree with you that it's best to be consistent. I hate to ask you to fill in the hole that you have just taken the time to dig, but I'd be in favor of changing the ~> references to NaturalTransformation within the code. Maybe it's best if at least one other person chimes in with their opinion before you take the time to make such a change?

I think the docs for FreeMonad use ~> as well. If we are going to use it there, we should probably at least point out that it's an alias for NaturalTransformation, so F ~> G is equivalent to NaturalTransformation[F, G].

[markus1189]

Since no one else chimed in, I assume that we go for the replacement of symbolic ~> with NaturalTransformation. I think I got all of them and added a short note in the free monad tutorial ;)

[markus1189]

Here is the note in the tutorial

[has609]

@@ -238,7 +240,7 @@ recursive structure by:
This operation is called Free.foldMap:

-final def foldMap[M[_]](f: S ~> M)(M: Monad[M]): M[A] = ...
+final def foldMap[M[_]](f: NaturalTransformation[S,M])(M: Monad[M]): M[A] = ...

[markus1189]

Whoops, missed that one ;)

[ceedubs]

@markus1189 thanks for helping keep us consistent.

I noticed that ~> is still used in a few places in the docs. Is there a reason behind switching some occurrences over to NaturalTransformation but not others?

[markus1189]

@ceedubs do you mean the ones in the tutorial for free monads? I think that it doesn't hurt to have ~> there, because it is explicitly mentioned and I think most people outside of cats would prefer ~> anyway, such that it could be a good idea to make readers of the tutorial familiar with the symbolic operator.

Other than that I only found occurrences of ~> inside of tests which I now also replaced. Anything else I am not seeing?

[ceedubs]

@markus1189 I think I was thrown off by the fact that object InMemoryDatasourceInterpreter extends NaturalTransformation[DataOp,Id] shows up in the free tutorial. If we are settling on ~> for the tutorial, then maybe we should change that to object InMemoryDatasourceInterpreter extends (DataOp ~> Id)?

[ceedubs]

Sorry but this PR has merge conflicts.

[markus1189]

Should be good to go now ;)

[ceedubs]

👍 thanks, @markus1189!

[non]

I'm 👍 on this change too. Unfortunately it has conflicts!