Compare commits

...

4 Commits

2 changed files with 47 additions and 3 deletions

44
app/swidnik.hs Normal file
View File

@ -0,0 +1,44 @@
{-# 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
swidnik
extractRecords = extractLinksWithText "//a/div[@src='http://www.bibliotekaswidnik.pl/images/banners/pdf_klik.jpg"']" -- pary adres-tytuł
>>> second (arr $ replace "\r\n " " ") -- czyścimy drugi element pary, czyli tytuł z niepotrzebnych białych znaków
>>> first (extractLinksWithText "//li/a[contains(@href,'.pdf')]") -- 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) -> ShadowItem
toShadowItem ((url, articleTitle), yearlyTitle) =
(defaultShadowItem url title) {
originalDate = Just date,
itype = "periodical",
format = Just "pdf",
finalUrl = url
}
where title = "Glos " ++ yearlyTitle ++ " " ++ (replace "\r\n" "" (replace "\r\n " "" articleTitle))
date = getDate url
getDate url =
case url =~~ "/(19[0-9][0-9]|20[0-9][0-9])/" :: Maybe [[String]] of
Just [[_, year]] -> year
otherwise -> error $ "unexpected url: " ++ url
main = do
let start = "s.bibliotekaswidnik.pl/index.php/116-glos-swidnika
let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
lname="Glos",
abbrev="Glos",
lLevel=0,
webpage=start}
extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)

View File

@ -47,9 +47,9 @@ library
-- , shadow-library -- , shadow-library
-- default-language: Haskell2010 -- default-language: Haskell2010
executable almanachmuszyny executable swidnik
hs-source-dirs: app hs-source-dirs: app
main-is: almanachmuszyny.hs main-is: swidnik.hs
ghc-options: -threaded -rtsopts -with-rtsopts=-N ghc-options: -threaded -rtsopts -with-rtsopts=-N
build-depends: base build-depends: base
, hxt , hxt
@ -62,4 +62,4 @@ executable almanachmuszyny
source-repository head source-repository head
type: git type: git
location: https://github.com/name/project location: https://git.wmi.amu.edu.pl/s444354/twilight-library