haskell-formatter-1.0.0: src/library/Language/Haskell/Formatter/Toolkit/ListTool.hs
{-|
Description : List utilities
-}
module Language.Haskell.Formatter.Toolkit.ListTool
(maybeLast, dropWhileAtMost, mergeLongerSuccessions, takeEvery,
concatenateRuns, concatenateShiftedRuns)
where
import qualified Data.List as List
import qualified Data.Maybe as Maybe
import qualified Data.Monoid as Monoid
{-| The last element, or 'Nothing' if there is none.
prop> maybeLast [] == Nothing
prop> maybeLast (l ++ [e]) == Just e -}
maybeLast :: [a] -> Maybe a
maybeLast = Maybe.listToMaybe . reverse
{-| @dropWhileAtMost p l@ is like @dropWhile p@, but drops at most @l@ elements.
>>> dropWhileAtMost (== ' ') 2 " a bc "
" a bc " -}
dropWhileAtMost :: (a -> Bool) -> Int -> [a] -> [a]
dropWhileAtMost predicate limit list
= Monoid.mappend (dropWhile predicate deformable) rigid
where (deformable, rigid) = splitAt limit list
{-| @mergeLongerSuccessions p c l@ keeps only the first @c@ elements of
successive elements of @l@ satisfying the predicate @p@.
>>> mergeLongerSuccessions Data.Char.isSpace 2 " ab c d\LF e "
" ab c d\n e " -}
mergeLongerSuccessions :: (a -> Bool) -> Int -> [a] -> [a]
mergeLongerSuccessions predicate count = snd . List.foldl' merge (0, [])
where merge (successionLength, list) element
= if predicate element then
if successionLength < count then (succ successionLength, extended)
else (count, list)
else (0, extended)
where extended = Monoid.mappend list [element]
{-| @takeEvery p l@ takes every @p@th element of @l@ from the first one.
>>> takeEvery 2 "apple"
"ape"
prop> takeEvery 1 l == l -}
takeEvery :: Int -> [a] -> [a]
takeEvery _ [] = []
takeEvery period list@(first : _) = first : takeEvery period (drop period list)
{-| @concatenateRuns p l@ repeatedly concatenates @p@ lists of @l@.
>>> concatenateRuns 2 ["a", "b", "c", "d", "e"]
["ab","cd","e"] -}
concatenateRuns :: Int -> [[a]] -> [[a]]
concatenateRuns _ [] = []
concatenateRuns period lists = concat run : concatenateRuns period rest
where (run, rest) = splitAt period lists
{-| @concatenateShiftedRuns p s l@ first takes @s@ lists of @l@, followed by
repeatedly concatenating @p@ lists.
>>> concatenateShiftedRuns 2 1 ["a", "b", "c", "d", "e"]
["a","bc","de"]
prop> p <= 0 || concatenateShiftedRuns p 0 l == concatenateRuns p l -}
concatenateShiftedRuns :: Int -> Int -> [[a]] -> [[a]]
concatenateShiftedRuns period shift lists
= case shift of
0 -> concatenateUnshifted lists
_ -> concat shifted : concatenateUnshifted unshifted
where (shifted, unshifted) = splitAt shift lists
where concatenateUnshifted = concatenateRuns period