https://twitter.com/1HaskellADay/status/474083245307351041

haskell_a_day_2014_06_05.hs

module NGrams where
import Data.Char
import Data.List

-- | Given a String, compute all the n letters n-grams of its words (excluding all non alphanums characters).
--   Word separations matter.
--
--   Example:
--
--   >>> lettersNGrams 3 "Hello, world!"
--   ["Hel","ell","llo","wor","orl","rld"]
--
--   >>> lettersNGrams 3 "Hello,world!"
--   ["Hel","ell","llo","wor","orl","rld"]
--
--   >>> lettersNGrams 10 "Hello, world!"
--   ["Hello","world"]
lettersNGrams :: Int -> String -> [String]
lettersNGrams n s = concatMap (nGrams n) (words (map (\c -> if isAlpha c then c else ' ') s))


-- Helper n-gram functions from yesterday
nGrams :: Int -> [a] -> [[a]]
nGrams n = unfoldr go
  where go [] = Nothing
        go xs = let
          (ng, remain) = splitAt n xs
          next = take 1 remain >> drop 1 xs -- if (null $ take 1 remain) then [] else drop 1 xs
          in Just (ng, next)