md5-hash-emails.hs

{-# LANGUAGE OverloadedStrings #-}

module Main where
  import Data.Digest.Pure.MD5
  import Data.List.Split
  import Data.List
  import System.Environment
  import qualified Data.ByteString.Lazy as LBS
  import qualified Data.ByteString.Char8 as C

  main = do
    args <- getArgs 
    let filename = head args
    tsvData <- C.readFile filename 
    let
      lns = C.lines $ tsvData 
      mdLines = map splitByTabAndMd5 lns
    C.writeFile "temp.txt" $ C.unlines mdLines
    return ()

  splitByTabAndMd5 s = C.concat $ map (flip C.snoc ' ' . md5IfEmail) $ C.words s

  md5IfEmail :: C.ByteString -> C.ByteString
  md5IfEmail s = if C.isInfixOf "@" s then emailToMD5 s else s

  emailToMD5 :: C.ByteString -> C.ByteString
  emailToMD5 s = C.concat ["\"", C.pack $ show $ md5 $ LBS.fromStrict $ stripQuotes s, "\""]

  stripQuotes :: C.ByteString -> C.ByteString
  stripQuotes s = C.take l $ C.drop 1 s
    where
      l = C.length s - 2

File: MPA_users_all.txt: Little-endian UTF-16 Unicode text, with CRLF line terminators