ngrams.hs

import Data.List (group, sort)
import Control.Arrow ((&&&))

nGrams :: Int -> [a] -> [[a]]
nGrams x xs
    | null xs = [[]]
    | x <= 0 = [[]]
    | x == 1 = [[y] | y <- xs]
    | x == 2 = map (\ (x, y) -> [x, y]) $ zip xs $ tail xs
    | otherwise = zipWith (:) xs (nGrams (x - 1) $ tail xs)

bigrams :: [a] -> [[a]]
bigrams = nGrams 2

countNGrams :: (Ord a) => Int -> [[a]] -> [([a], Int)]
countNGrams x xss = map (head &&& length) . group $ sort [xs | xs <- xss, length xs == x || x < 0 ]


countBigrams' :: (Ord a) => [a] -> [([a], Int)]
countBigrams' = countNGrams 2 . nGrams 2

-- -for E58 assignment
countBigrams :: String -> [(String, Int)]
countBigrams xs = map (fst &&& (\ (_, y) -> y - 1)) $ countNGrams 2 $ [ [x, y] | x <- "ATCG", y <- "ATCG" ] ++ nGrams 2 "ATCG"