Lexing and parsing an e-mail address

An elegant way to clean data is by defining a lexer to split up a string into tokens. In this recipe, we will parse an e-mail address using the attoparsec library. This will naturally allow us to ignore the surrounding whitespace.

Getting ready

Import the attoparsec parser combinator library:

$ cabal install attoparsec

How to do it…

Create a new file, which we will call Main.hs, and perform the following steps:

1. Use the GHC OverloadedStrings language extension to more legibly use the Text data type throughout the code. Also, import the other relevant libraries:

   {-# LANGUAGE OverloadedStrings #-}
   import Data.Attoparsec.Text
   import Data.Char (isSpace, isAlphaNum)

2. Declare a data type for an e-mail address:

   data E-mail = E-mail 
     { user :: String
     , host :: String
     } deriving Show

3. Define how to parse an e-mail address. This function can be as simple or as complicated as required:

   e-mail :: Parser E-mail
   e-mail = do
     skipSpace
     user <- many' $ satisfy isAlphaNum
     at <- char '@'
     hostName <- many' $ satisfy isAlphaNum
     period <- char '.'
     domain <- many' (satisfy isAlphaNum)
     return $ E-mail user (hostName ++ "." ++ domain)

4. Parse an e-mail address to test the code:

   main :: IO ()
   main = print $ parseOnly e-mail "nishant@shukla.io"

5. Run the code to print out the parsed e-mail address:

   $ runhaskell Main.hs
   
   Right (E-mail {user = "nishant", host = "shukla.io"})
   

How it works…

We create an e-mail parser by matching the string against multiple tests. An e-mail address must contain some alphanumerical username, followed by the 'at' sign (@), then an alphanumerical hostname, a period, and lastly the top-level domain.

The various functions used from the attoparsec library can be found in the Data.Attoparsec.Text documentation, which is available at https://hackage.haskell.org/package/attoparsec/docs/Data-Attoparsec-Text.html.