Question

I have an input where “Divinités” (9 characters) would be represented as “Divinit\303\251s” (actual text data of 16 characters long), how can I convert it to Haskell’s properly encoded Text (or ByteString, or String)?

Answer 1

First you need to unescape string converting each escape sequence to one Char. Then use utf8-string package to decode the result to actual utf8 string.

import Data.Char
import Codec.Binary.UTF8.String (decodeString)

input :: String
input = "Divinit\\303\\251s"

main = maybe (return ()) putStrLn $ convertString input

convertString :: [Char] -> Maybe [Char]
convertString = fmap decodeString . unescape

unescape :: [Char] -> Maybe [Char]
unescape [] = Just []
unescape ('\\' : tail) = do
  headResult <- fmap toEnum . octalDigitsToInt . take 3 $ tail
  tailResult <- unescape . drop 3 $ tail
  return $ headResult : tailResult
unescape (head : tail) = fmap (head :) . unescape $ tail

octalDigitsToInt :: [Char] -> Maybe Int
octalDigitsToInt = 
  fmap sum . sequence .
    map (\(i, c) -> fmap (8^i*) $ octalDigitToInt c) .
      zip [0..] . reverse

octalDigitToInt :: Char -> Maybe Int
octalDigitToInt c | isOctDigit c = Just $ digitToInt c
octalDigitToInt _ = Nothing