{-# LANGUAGE Arrows, NoMonomorphismRestriction #-}
import ShadowLibrary.Core

import Text.XML.HXT.Core
import Text.XML.HXT.XPath
-- import Text.XML.HXT.Curl 
import Data.List
import Data.List.Utils (replace)

import Text.Regex.Posix
import Text.Printf


extractRecords = extractLinksWithText "//a[@class='ramka']"  -- pary adres-tytuł

                 >>> first (extractLinksWithText "//div/a[contains(@href, '.zip') and not (contains(@href,'text_offline.zip'))]") -- tutaj wybieramy paczki zip z plikami DjVu, jeśli byśmy chcieli bezpośrednio format DjVu wystarczy trochę zmienić zapytanie XPatha.

-- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem
toShadowItem :: ((String, String), String) -> ShadowItem
toShadowItem ((url, articleTitle), yearlyTitle) =
  (defaultShadowItem url title) {
    originalDate = Just date,
    itype = "periodical",
    format = Just "zip",
    finalUrl = url
    }
  where title = "Tajemnica Atari " ++ yearlyTitle ++ " " ++ (replace "\r\n" "" (replace "\r\n          " "" articleTitle))
        date = getDate url

getDate url =
  case url =~~ "/([1-9]_9[0-9]|1[0-2]_9[0-9]|[1-9]-[1-9]_9[0-9]|1[0-2]-1[0-2]_9[0-9])/" :: Maybe [[String]] of
    Just [[_, year]] -> year
    otherwise -> error $ "unexpected url: " ++ url


main = do
    let start = "http://krap.pl/mirrorz/atari/horror.mirage.com.pl/pixel/"
    let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
                                       lname="Tajemnica Atari",
                                       abbrev="TA",
                                       lLevel=0,
                                       webpage=start}
    extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)