typelevel/cats

Welcome! Got a question? Do you have -Ypartial-unification turned on? Other FAQs: http://typelevel.org/cats/faq.html

Jasper Moeys

@Jasper-M

Is there some well known name for this method foo?

def foo[A, B, C](f: A => B)(g: (A, B) => C): A => C =
  a => g(a, f(a))

Fabio Labella

@SystemFw

@Jasper-M that seems equivalent of Applicative.ap for F = A =>, with uncurrying and tuple == parameter lists

but it's not worth writing it differently than what you have I don't think

you can notice by changing (A, B) => C into A => B => C

then factorising the whole signature into (A => B) => (A => (B => C)) => (A => C)

then abstracting A => into F[_]

F[B] => F[B => C] => F[C]

and then alfa-rename and switch order into

F[A => B] => F[A] => F[B], which is the key signature of the Apply typeclass

Fabio Labella

@SystemFw

for F[E] = E => *

(renamed some type variables for clarity)

Jasper Moeys

@Jasper-M

Thanks! I felt like there had to be some cats concept for this

Fabio Labella

@SystemFw

one of the key soft skills is an eye for spotting F[_] patterns

in particular the starting point could have have swiping f and g and noticing basically (A => B) => (A => C)

that should trigger your functorial warning, since functor, applicative and monad can all be seen to transform things into F[B] => F[C]

although it's easier to rename things to F[A] => F[B] for consistency with the way things are usually written down

map: (A => B) => (F[A] => F[B])
ap: F[A => B] => (F[A] => F[B])
flatMap: (A => F[B]) => (F[A] => F[B])

Jasper Moeys

@Jasper-M

one of the key soft skills is an eye for spotting F[_] patterns

I'll tell that to my manager when he starts about communication and presenting

Fabio Labella

@SystemFw

which also helps with understanding that the real difference between power in the three typeclasses reflects the difference in power between A => B (uneffectful), F[A =>B](effectful but no context sensitivity) and A => F[B] (the effect can depend on the output of the previous computation, since A => F

I'll tell that to my manager when he starts about communication and presenting

ahahhaha :+1:

Adam Rosien

@arosien

traverse: (A => G[B]) => (F[A] => G[F[B]])

(kind of like flatMap wooo)

https://typelevel.org/cats/datatypes/either.html#solution-2-adts-all-the-way-down talks about error ADTs, what's SOP for common errors that you want to share across modules?

i think usually you have a separate ADT that is shared across modules. if you can have direct dependencies, then the upper ADT can refer directly to the module error types. otherwise you can cast or translate the module errors into something like Throwable which the upper ADT wraps

Nathaniel Fischer

@kag0

@arosien thanks, that makes sense. I was thinking of doing the opposite - having service specific errors extend common errors. but that just seemed messy.

Adam Rosien

@arosien

agreed!

you probably don't need it, and it's more complicated, but you could have a typeclass that translates "lower" level errors into a "higher" level error, but that means the higher level error loses information (which may be ok)

Nathaniel Fischer

@kag0

that could be interesting
I was already thinking of having a chain of causes similar to exceptions, so details could be retained that way

Soren

@srnb_gitlab

This just means that for a foo: Foo => Bar I can get a mFoo: M[Foo] => M[Bar], right?

(H is a monad)

Soren

@srnb_gitlab

Also, does "[a] monad on the category of something" mean that something is a Monad? (I don't know a lot of the notation and terminology for all of this)

Ryan Peters

@sloshy

@srnb_gitlab yes. Consider the age-old "a Monad is just a monoid in the category of endofunctors" for another example

Soren

@srnb_gitlab

@sloshy Cool. I'm implementing the H monad from https://arxiv.org/pdf/1804.09408.pdf now

And I think I get most of it

Unfortunately its easier to implement map and flatten than flatMap

Soren

@srnb_gitlab

@sloshy Am I misreading or does the H monad only define the pure operation over directed hypergraphs?

Soren

@srnb_gitlab

Nevermind, I forgot that I can represent any value as a unary sourced hypergraph

What do I do when I don't think its possible to implement tailRecM?

Soren

@srnb_gitlab

https://github.com/sorenbug/graphmonad/blob/master/core/src/main/scala/tf/bug/graphmonad/h/SourcedHyperGraph.scala#L44

PandaSeal20

@PandaSeal20

Hi all, I am an intermediate scala developer I suppose you might say. I am introducing cats to my work codebases. One problem I seem to have run into is that there is one data type that I find really useful and use for almost everything - EitherT. I want to branch out and use other datatypes, I just find I can code so much faster using this and it does exactly what I need in a lot of cases. I've also used resource, validated and a couple of other ones which worked well. Any advice on how to branch out in Cats, what are some other useful parts of the cats library I should be using? At work I have to do a lot of validation, so having a nice long for comprehension that will shortcircuit on the left works nicely.

Louis Bettens

@lourkeur

Is this a known infringement of Order laws by FloatOrder and DoubleOrder?

cats.kernel.laws.discipline.OrderTests[Float].laws.partialCompare(Float.NaN, Float.NaN) 
  // => IsEq(false, true)
cats.kernel.laws.discipline.OrderTests[Float].laws.partialCompare(-0f,0f) 
  // => IsEq(false, true)

This comes down to java.lang.Float being inconsistent with builtin float by design

Yilin Wei

@yilinwei

@srnb_gitlab tailRecM practically speaking is a stack safe flatMap (the Either represents whether to terminate or keep recursing). You may not be able to implement a stack safe one easily, but you should at least be able to match the signature. Take a look at the existing data types tailRecM to get some ideas of how to write them.

Louis Bettens

@lourkeur

@lourkeur #855

Ryan Peters

@sloshy

@PandaSeal20 Are you mainly using cats data types from the sounds of it? Or do you also write code using cats typeclasses? For me that's the most liberating part: writing some ultra-generic code that has very little concrete dependencies so I can reuse the functions I write in surprising new ways.

You mentioned resource - from cats effect? I've gotten a lot of utility from Ref + Deferred as well (not cats Defer, but cats effect Deferred)

Probably the biggest utility I get though is from streaming abstractions built on cats, so libraries like FS2 and Monix. I gave a talk recently for NEScala this year about why you should try to use them nearly everywhere 😅 but it's not uploaded to the public internet yet.

Ryan Peters

@sloshy

Speaking of validation, recently I've been writing a "validation combinators" kind of library for work, which uses Either, Parallel, and Kleisli to let you combine functions for validating. Kleisli is very useful for this kind of thing.

Soren

@srnb_gitlab

@yilinwei I'm worried about when to finally return. Once head is Right? Once head and tail are Right? Once head, tail, values, and edges are Right?

Fabio Labella

@SystemFw

I wouldn't worry about stack safety at this stage, you're just playing with a paper

you can extend StackSafeMonad to get an implementation of tailRecM for free given flatMap

Where communities thrive

People

Repo info

Activity