[Haskell-cafe] [random-fu] sampleRVar with StdGen in RWS?

[Sergiu Ivanov]

Thus quoth  Vanessa McHale  on Fri Aug 03 2018 at 15:06 (+0200):
> It seems like the instance comes from
> http://hackage.haskell.org/package/random-source-0.3.0.6/docs/Data-Random-Internal-Source.html#v:getRandomPrim
> then, no? I only see an instance for `StateT` unfortunately.

Yes, that's what I see as well.

> I suppose you could write an orphan instance for RWST if you really
> wanted to.

That's what I'll most probably do for my small project, but I was
wondering whether it made sense to make a pull request to random-fu.  To
me, having a MonadRandom for both StateT and RWST seems very natural.

-
Sergiu


> On 08/03/2018 07:54 AM, Sergiu Ivanov wrote:
>> Hello Vanessa,
>>
>> MonadRandom from Control.Monad.Random does indeed have an RWST instance,
>> but I'd like to stick with random-fu because of the long list of
>> provided distributions, and the two seem incompatible :
>>
>>   https://stackoverflow.com/a/13946470
>>
>> I did try to force a very monoidal write-state:
>>
>>   testRWS :: StdGen -> Int
>>   testRWS = fst . evalRWS work ()
>>     where
>>       work = do
>>         tell "hello"
>>         sampleRVar $ uniform 1 10
>>
>> I get
>>
>>   <interactive>:2:25: error:
>>   • No instance for (MonadRandom (RWST () String StdGen Identity)) arising from a use of ‘work’
>>   • In the first argument of ‘evalRWS’, namely ‘work’
>>   In the second argument of ‘(.)’, namely ‘evalRWS work ()’
>>   In the expression: fst . evalRWS work ()
>>
>> which seems to be saying the same thing as before :-(
>>
>> -
>> Sergiu
>>
>>
>> Thus quoth  Vanessa McHale  on Fri Aug 03 2018 at 14:43 (+0200):
>>> It seems like you need your write-state to be a monoid in order to infer
>>> a `MonadRandom` instance. Based on your code, I'm guessing the compiler
>>> can't infer that.
>>>
>>> Have a look at
>>> https://hackage.haskell.org/package/MonadRandom-0.5.1.1/docs/Control-Monad-Random-Class.html#t:MonadRandom
>>>
>>>
>>> On 08/03/2018 07:31 AM, Sergiu Ivanov wrote:
>>>> Dear Café,
>>>>
>>>> I would like to use random-fu to do some pseudo-random simulations for a
>>>> given StdGen (so that I can run the same simulation multiple times, if
>>>> needed).
>>>>
>>>> The following works:
>>>>
>>>>   testState :: StdGen -> Int
>>>>   testState = evalState (sampleRVar $ uniform 1 10)
>>>>
>>>> The following doesn't:
>>>>
>>>>   testRWS :: StdGen -> Int
>>>>   testRWS = fst . evalRWS (sampleRVar $ uniform 1 10) ()
>>>>
>>>> I get
>>>>
>>>>   <interactive>:2:26: error:
>>>>   • No instance for (MonadRandom (RWST () b0 StdGen Identity)) arising from a use of ‘sampleRVar’
>>>>   • In the first argument of ‘evalRWS’, namely ‘(sampleRVar $ uniform 1 10)’
>>>>   In the second argument of ‘(.)’, namely ‘evalRWS (sampleRVar $ uniform 1 10) ()’
>>>>   In the expression: fst . evalRWS (sampleRVar $ uniform 1 10) ()
>>>>
>>>> Indeed, I do see a MonadRandom instance for StateT, but none for RWST [0].
>>>>
>>>> Is there a reason to not have a MonadRandom instance for RWST?
>>>>
>>>> Am I looking in the wrong place?
>>>>
>>>> -
>>>> Sergiu
>>>>
>>>> [0] https://hackage.haskell.org/package/random-source-0.3.0.6/docs/Data-Random-Source.html
>>>>
>>>>
>>>> _______________________________________________
>>>> Haskell-Cafe mailing list
>>>> To (un)subscribe, modify options or view archives go to:
>>>> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
>>>> Only members subscribed via the mailman list are allowed to post.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 487 bytes
Desc: not available
URL: <http://mail.haskell.org/pipermail/haskell-cafe/attachments/20180803/fd318ebf/attachment.sig>