Parsing html with Haskell

Editors.hs

---------------------------------------------------------------
-- |
-- Module      : Editors
-- Copyright   : (c) Hitesh Jasani, 2007
-- License     : BSD3
--
-- Maintainer  : Hitesh Jasani <[email protected]>
-- Stability   : provisional
-- Portability : 
--
-- Created     : 2007-11-06
-- Version     : 0.1
--
-- Parse a wikipedia page and pull out all the editors.
--
-- This code depends upon Neil Mitchell's TagSoup library
-- from darcs (circa Nov 1, 2007)
---------------------------------------------------------------

module Main where

import Control.Arrow ( (&&&) )
import Text.HTML.TagSoup
import qualified Text.HTML.TagSoup.Match as M
import Text.HTML.Download ( openURL )
import System.Environment ( getArgs )

main = do
  args <- getArgs
  case args of
    ["save"]  -> cachePage wikiHaskell cachefile
    otherwise -> do src <- readCache cachefile
                    mapM_ showEdits . map (getName &&& getComment) $ tags src
  where
    tags = findSpans . parseTags
    findSpans = sections (~== TagOpen "span" [("class","history-user")])
    showEdits (n,c) = putStrLn $ n ++ ": " ++ c

getName = innerText . take 1 . drop 2

getComment ts =
  let cSt = runtil ts
  in if (isTagCloseName "li" $ head cSt)
       then "(no comment)"
       else reverse . drop 2 . reverse . concat . map to_s .
            takeWhile (~/= TagOpen "span" [("class", "mw-history-undo")]) $ cSt
  where
    runtil ts = dropWhile (\x -> (x ~/= (TagOpen "span" [("class","comment")]))
                                 && not (M.tagClose (=="li") x)) ts
    to_s (TagOpen "span" _) = ""
    to_s (TagText s)        = s
    to_s (TagOpen "a"    _) = ""
    to_s (TagClose s)       = ""

readCache fn = readFile fn

cachefile = "page.cache"

cachePage url fn = do
  src <- openURL url
  writeFile fn src

wikiHaskell = "http://en.wikipedia.org/w/index.php?title=Haskell_%28programming_language%29&action=history"