- Beginners - Haskell.org

program not running lazily
by Mason Lai 20 Jan '16

20 Jan '16

Hi, I'm teaching myself Haskell and attempting to use it for day-to-day tasks at work. I signed up for this list a few weeks ago, and this is my first post. Here's my problem: I'm given a list of 45 Strings, each nine Chars long. The only characters are 'A', 'C', 'G' and 'T'. (This is for a bioinformatics application.) I'll refer to any nine-length String composed of these Chars as a "9-mer". I need to generate a larger list of 9-mers such that no two 9-mers have a Levenshtein distance of less than 3. (The list I start with satisfies this requirement.) I can generate as many 9-mers as I possibly can, but this process is very slow. It's also not being computed lazily; calling head on the output list forces the entire list of 9-mers to be computed. *I'd like to understand why this list isn't being computed lazily, and how I can change my code to make it so. *My knowledge of monads is pretty poor as well, so a digression or a series of questions about why the line [9] >>= (`replicateM` "ACGT") works would also be helpful, as would general tips about writing clean code. This is an O(n^2) operation, so I'm not expecting it to be slow. However, I figured that I'd just be able to take the first N elements from the output list. Here's what I have: import Control.Monad main :: IO () main = interact processData processData :: String -> String processData = unlines . (`merge` ([9] >>= (`replicateM` "ACGT"))) . lines -- Merges two lists of things into a single list merge :: Eq a => [[a]] -> [[a]] -> [[a]] merge xs [] = xs merge xs ys = merge ((head ys) `addInto` xs) $ tail ys -- Adds a thing into a list if it is "different" enough addInto :: Eq a => [a] -> [[a]] -> [[a]] y `addInto` ys = if ((minimum $ map (dist y) ys) < 3) then ys else y:ys -- Calculate the Levenshtein distance -- Lloyd Allison algorithm. See Haskell wiki -- code omitted dist :: Eq a => [a] -> [a] -> Int My workaround to getting a smaller subset of the whole list is to replace [9] >>= (`replicateM` "ACGT") with take 10000 $ [9] >>= (`replicateM` "ACGT") Thanks! Mason Lai

1 0

Problem using stack
by Robert Weisser 20 Jan '16

20 Jan '16

I recently installed stack via Haskell for Mac OS X (https://ghcformacosx.github.io/) which gave me stack 0.1.8.0 x86_64 and ghc 7.10.2. Following the User Guide (http://docs.haskellstack.org/en/stable/GUIDE.html) I tried to create the first example (hello world, naturally). Everything went according to plan until I entered 'stack test', which failed. The output was supposed to be, 'Test suite not yet implemented.' I tried to find information online, but was unsuccessful. I play with Haskell for fun and know nothing about cabal, so I don't know if it is a cabal issue. The error message includes 'helloworld-test: executable not found'. helloworld-test is indeed missing. I searched for it using find: find . -name '*test' Here is the terminal session, followed by the contents of the generated helloworld.cabal file (with some user info removed): $ stack new helloworld new-template Downloading template "new-template" to create project "helloworld" in helloworld/ ... Writing default config file to: /Users/rpw/self/haskell/stack/user_guide/helloworld/stack.yaml Basing on cabal files: - /Users/rpw/self/haskell/stack/user_guide/helloworld/helloworld.cabal Checking against build plan lts-3.13 Selected resolver: lts-3.13 Wrote project config to: /Users/rpw/self/haskell/stack/user_guide/helloworld/stack.yaml $ cd helloworld/ $ stack setup stack will use the GHC on your PATH For more information on paths, see 'stack path' and 'stack exec env' To use this GHC and packages outside of a project, consider using: stack ghc, stack ghci, stack runghc, or stack exec $ stack build helloworld-0.1.0.0: configure Configuring helloworld-0.1.0.0... helloworld-0.1.0.0: build Preprocessing library helloworld-0.1.0.0... [1 of 1] Compiling Lib ( src/Lib.hs, .stack-work/dist/x86_64-osx/Cabal-1.22.4.0/build/Lib.o ) In-place registering helloworld-0.1.0.0... Preprocessing executable 'helloworld-exe' for helloworld-0.1.0.0... [1 of 1] Compiling Main ( app/Main.hs, .stack-work/dist/x86_64-osx/Cabal-1.22.4.0/build/helloworld-exe/helloworld-exe-tmp/Main.o ) Linking .stack-work/dist/x86_64-osx/Cabal-1.22.4.0/build/helloworld-exe/helloworld-exe ... helloworld-0.1.0.0: copy/register Installing library in /Users/rpw/self/haskell/stack/user_guide/helloworld/.stack-work/install/x86_64-osx/lts-3.13/7.10.2/lib/x86_64-osx-ghc-7.10.2/helloworld-0.1.0.0-6urpPe0MO7OHasGCFSyIAT Installing executable(s) in /Users/rpw/self/haskell/stack/user_guide/helloworld/.stack-work/install/x86_64-osx/lts-3.13/7.10.2/bin Registering helloworld-0.1.0.0... $ stack exec helloworld-exe someFunc $ stack test Test suite helloworld-test executable not found for helloworld Test suite failure for package helloworld-0.1.0.0 helloworld-test: executable not found Logs printed to console $ cat helloworld.cabal name: helloworld version: 0.1.0.0 synopsis: Initial project template from stack description: Please see README.md homepage: ... license: BSD3 license-file: LICENSE author: ... maintainer: ... copyright: None category: Example build-type: Simple -- extra-source-files: cabal-version: >=1.10 library hs-source-dirs: src exposed-modules: Lib build-depends: base >= 4.7 && < 5 default-language: Haskell2010 executable helloworld-exe hs-source-dirs: app main-is: Main.hs ghc-options: -threaded -rtsopts -with-rtsopts=-N build-depends: base , helloworld default-language: Haskell2010 test-suite helloworld-test type: exitcode-stdio-1.0 hs-source-dirs: test main-is: Spec.hs build-depends: base , helloworld ghc-options: -threaded -rtsopts -with-rtsopts=-N default-language: Haskell2010 source-repository head type: git location: ... $ Robert Weisser

2 2

proving fmap law for recursive data types
by 鲍凯文 16 Jan '16

16 Jan '16

Hi, Suppose you have something like: data Tree a = Leaf a | Branch (Tree a) (Tree a) instance Functor Tree where fmap f (Leaf a) = Leaf $ f a fmap f (Branch l r) = Branch (fmap f l) (fmap f r) To check that fmap id = id, I can see that the case for Leaf is ok: fmap id (Leaf a) = Leaf $ id a = Leaf a How would you prove it for the second case? Is some sort of inductive proof necessary? I found this: http://ssomayyajula.github.io/posts/2015-11-07-proofs-of-functor-laws-with-…, which goes over an inductive (on the length of the list) proof for the list type, but I'm not sure how to do it for a Tree. Thanks, toz

2 1

Ambiguous module name `Prelude': it was found in multiple packages (trying to install HXQ)
by Stanislaw Findeisen 16 Jan '16

16 Jan '16

Hi! I am trying to install HXQ: https://hackage.haskell.org/package/HXQ . But somehow it doesn't work: $ cabal install hxq Resolving dependencies... Configuring HXQ-0.19.0... Building HXQ-0.19.0... Failed to install HXQ-0.19.0 Build log ( /home/ct/.cabal/logs/HXQ-0.19.0.log ): Configuring HXQ-0.19.0... Building HXQ-0.19.0... Preprocessing library HXQ-0.19.0... src/Text/XML/HXQ/XQuery.hs:1:1: Ambiguous module name `Prelude': it was found in multiple packages: base haskell98-2.0.0.2 cabal: Error: some packages failed to install: HXQ-0.19.0 failed during the building phase. The exception was: ExitFailure 1 Why? Thanks! -- http://people.eisenbits.com/~stf/ http://www.eisenbits.com/ OpenPGP: 9EC2 5620 2355 B1DC 4A8F 8C79 0EC7 C214 E5AE 3B4E

5 7

Implementing instance of '^' operator
by pmcilroy＠gmail.com 16 Jan '16

16 Jan '16

As a ‘hello world’ example for type definitions, I like to define a numeric type that can handle the mod p multiplicative group, where p is prime. This requires: • Implementing interface functions • Defining non-trivial implementations, where constructor must be private, etc. • Invoking an abstract superclass concrete instance method from within the subclass method definition The latter appears not to be possible in Haskell. Is this true? Here’s the basic code, but I punted on x^n. It looks like I’d have to paste in the entire original definition of ‘^’. data Modp a = Modp a a deriving (Eq, Show) mkModp p n | isPrime p = Modp p (n `mod` p) | otherwise = error $ show p ++ " is not a prime" instance Integral a => Num (Modp a) where (Modp q n) + (Modp p m) | p==q = Modp p $ (n+m) `mod` p | otherwise = error $ "unequal moduli" (Modp p n) * (Modp q m) | p==q = Modp p $ (n*m) `mod` p | otherwise = error $ "unequal moduli" negate (Modp p n) = Modp p (p-n) -- can't reuse base because ^ is impl. directly in prelude {- (Modp p x) ^ n | n <= p = (Modp p x) `baseExp` n | n1 == 0 = (Modp p x) | n > p = x ^ n1 where baseExp = ^ in Num n1 = n `mod` p -} instance Integral a => Fractional (Modp a) where recip (Modp p n) = (Modp p n)^(p-2) isPrime p = True -- stub

3 2

haskell platform
by DJ 13 Jan '16

13 Jan '16

Getting back to Haskell after giving up three years ago. I am running Linux Mint 17.1. Why is the Haskell Platform for linux (or my distro anyway) stuck at 2013? I see that ghc is several versions newer than what comes with Haskell Platform. That seems like a rather severe lag. Why is that? Does it matter? Best, - Jake -

6 9

Typeclass question
by Alan Buxton 13 Jan '16

13 Jan '16

What am I doing wrong in this admittedly contrived example? The code below will compile. It works as expected, unless I try to do "toMyParam Nothing". See below: λ: let arr = [P1 3.0, P2 'x'] λ: toMyParam False P2 'F' λ: toMyParam (Just 'x') P2 'x' λ: toMyParam Nothing <interactive>:38:1: No instance for (ToMyParam a0) arising from a use of `toMyParam' The type variable `a0' is ambiguous Code below: data MyParam = P1 Double | P2 Char deriving Show class ToMyParam a where toMyParam :: a -> MyParam instance ToMyParam Bool where toMyParam False = P2 'F' toMyParam True = P2 'T' instance ToMyParam Char where toMyParam = P2 instance ToMyParam Double where toMyParam = P1 instance ToMyParam a => ToMyParam (Maybe a) where toMyParam Nothing = P1 0.0 toMyParam (Just x) = toMyParam x

2 1

Profiling tiny executions in ghc
by pmcilroy＠gmail.com 12 Jan '16

12 Jan '16

Is there a way to count numbers of calls, allocations or constructions to get an estimate of program performance in ghc? I am curious about the performance of the various implementations of blowup in previous discussions. Timing and profile sampling is not convenient.

2 2

questions on evaluation.
by Peter McIlroy 12 Jan '16

12 Jan '16

Fabian, For your problem of bang -> baanngggg You can derive Rein's hairy idiom from a more straightforward technique for getting a recursive definition: First get a basic implementation using ++ and recursion: * add a helper function with any "state variables" as arguments. * Ignore concatenation while thinking about a problem. * Use pattern matching for recursion step. This gives the basic (naïve) implementation: blowup :: [a] -> [a] blowup x = concat (blowup' 1 x) -- returns (e.g.) ["b", "aa", "nnn", "gggg" ] blowup' a b :: Integer a => a -> [b] -> [[b]] blowup' n (x:xs) = (replicate n x : blowup' n+1 xs) blowup' _ _ = [] -- base case You can distribute concatenation into blowup', to get recursion over concatenation, but this really is going the wrong direction: blowup' n (x:xs) = (replicate n x) ++ (blowup' n+1 xs) blowup' _ _ = [] Instead, derive Rein's hairy-looking version by (as always) ignoring concatenation and replacing the state variable in blowup' with zip or zipWith. observe: > zip [1..] "bang" [(1,'b'),(2,'a'),(3,'n'),(4,'g')] So blowup' n x becomes blowup' x: blowup x = concat (blowup' x) blowup' x = map (\(a,b) replicate a b) zip [1..] x ==> definition of zipWith blowup' x = zipWith replicate [1..] x ==> blowup x = concat (zipWith replicate [1..] x) You can always push concat down into a function by unwrapping it with join, getting a more efficient version that creates the result directly. blowup x = concat (zipWith replicate [1..] x) ==> blowup x = (join . zipWith replicate [1..] ) x ==> blowup = join . zipWith replicate [1..]

3 3

question on evaluation
by Fabien R 11 Jan '16

11 Jan '16

Hello, I want to define a function blowup that takes "bang" as input and returns "baannngggg". I come up with these functions; myReverse :: [a] -> [a] myReverse [] = [] myReverse (x:xs) = myReverse xs ++ [x] buildLenList :: String -> [Int] buildLenList "" = [] buildLenList (_:xs) = [1 + length xs ] ++ buildLenList xs myRepeat :: Char -> Int -> String myRepeat x 0 = [] myRepeat x n = [x] ++ myRepeat x (n - 1) blowup :: String -> String blowup [] = [] blowup (x:xs) = myRepeat x (head ( (buildLenList (x:xs)))) ++ blowup xs With this code, blowup "bang" returns "bbbbaaanng". So I thought to insert myReverse between head and buildLenList but in that case, the result is only "bang". It seems that the evaluation of buildLenList is not working as I thought. I tried to debug that using ghci debugger but failed (still learning). Can someone explain how the evaluation is done here ? -- Fabien

3 4