twilight-library/app/elektronikapraktyczna.hs

{-# 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

import Data.Char (toLower)

class Nothingish a where
    nada :: a
instance Nothingish [a] where
    nada = []

eliminate :: (Nothingish a) => Maybe a -> a
eliminate (Just a) = a
eliminate Nothing  = nada

toLowerString :: [Char] -> [Char]
toLowerString str = [ toLower x | x <- str]

mToString :: Maybe String -> String
mToString n 
  | n == Nothing = ""
  | otherwise = eliminate n

extractMonth :: String -> String
extractMonth n =
  case n =~~ ("[A-za-z]+" :: String) of
    Just month -> "-" ++ eliminate (baseMonthNameToNumber (toLowerString month))
    otherwise -> ""

mExtractYear :: String -> String
mExtractYear n =
  case n =~~ ("(1[6789]|20)[0-9][0-9]" :: String) of
    Just year -> year
    otherwise -> ""

changeDate :: String -> Maybe String
changeDate a = Just (eliminate (extractYear a) ++ extractMonth a)

extractLinksWithArticleTitle xpathCondition = (downloadDocument &&& this)
                                      >>> first (getXPathTrees xpathCondition
                                                >>> (
                                                  (getXPathTrees "//div[@class='text']" >>> (listA (deep isText >>> getText)
                                                            >>> arr (intercalate " "))) 
                                                    &&& 
                                                    (getXPathTrees "//div[@class='files__item']/a[contains(@href,'.pdf')]" >>> (getAttrValue "href"))
                                                ))

--extractRecords = extractLinksWithText "(//a[@class='magazine-list__year-item'])[last()]"  -- pary adres-tytuł
extractRecords = extractLinksWithText "//a[@class='magazine-list__year-item']"  -- pary adres-tytuł
                 >>> second (arr $ replace "\r\n                        " "") -- czyścimy drugi element pary, czyli tytuł z niepotrzebnych białych znaków
                 >>> second (arr $ replace "    " "") -- czyścimy drugi element pary, czyli tytuł z niepotrzebnych białych znaków
--                 >>> first (arr ((++"tr") . init))  -- modyfikujemy pierwszy element pary, czyli adres URL
                 >>> first (extractLinksWithText "//div[@class='magazine-list__item']/a") -- pobieramy stronę z adresu URL i wyciągamy linki z tej strony pasujące do wyrażenia XPathowego
                 >>> first (second (arr $ replace "\r\n                                " ""))
                 >>> first (first (
                                    -- extractLinksWithArticleTitle "//div[@class='files__item']/a[contains(@href,'.pdf')]" 
                                    (extractLinksWithArticleTitle "//div[@class='magazine-single__content-title article text']")
                                    >>> 
                                                first (
                                                  first (arr $ replace "                                                                                    " "")
                                                  >>> first (arr $ replace "\r\n        " "")
                                                )
                                                -- >>> first (arr $ replace "\r\n" "")
                                    -- >>> first (arr $ replace "//" "/")
                                  )
                                  -- >>> second (arr $ changeDate) -- Zmiana nazwy miesiąca na wartość liczbową
                                  >>> second (arr $ replace "                                " "") >>> second (arr $ replace "     " "")
                            ) -- pobieramy stronę z adresu URL i wyciągamy linki z tej strony pasujące do wyrażenia XPathowego
                 -- ostatecznie wyjdą trójki ((adres URL, tytuł artykułu), tytuł rocznika)

-- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem
toShadowItem :: ((((String, String), String), String), String) -> ShadowItem
toShadowItem ((((chapterTitle, url), finalUrl), articleTitle), yearlyTitle) =
  (defaultShadowItem url title) {
    originalDate = changeDate articleTitle,
    itype = "periodical",
    format = Just "pdf",
    finalUrl = finalUrl
    }
  where title = "Elektronika Praktyczna  " ++ articleTitle ++ " - " ++ chapterTitle 
        date = yearlyTitle

getDate yearlyTitle =
  case yearlyTitle =~~ "/(19[0-9][0-9]|20[0-9][0-9])/" :: Maybe [[String]] of
    Just [[_, year]] -> year
    otherwise -> error $ "unexpected yearlyTitle: " ++ yearlyTitle


main = do
    let start = "https://ep.com.pl/archiwum/"
    let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
                                       lname="Elektronika praktyczna",
                                       abbrev="EP",
                                       lLevel=0,
                                       webpage=start}
    extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)