final version

This commit is contained in:
Jakub Kaczmarek 2022-04-06 20:13:51 +02:00
parent 079e69cbba
commit b315a478a3
2 changed files with 52 additions and 0 deletions

41
app/parafiasrem.hs Normal file
View File

@ -0,0 +1,41 @@
{-# 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[contains(@href, 'index.php?download')]" -- pary adres-tytuł
>>> second (arr $ replace "\r\n " " ") -- czyścimy drugi element pary, czyli tytuł z niepotrzebnych białych znaków
toShadowItem :: (String, String) -> ShadowItem
toShadowItem (url, articleTitle) =
(defaultShadowItem url title) {
originalDate = Just date,
itype = "periodical",
format = Just "pdf",
finalUrl = url
}
where title = "Parafia Srem " ++ 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 = "https://sremfara.pl/miesiecznik_30.html"
let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
lname="Parafia Srem",
abbrev="ParSre",
lLevel=0,
webpage=start}
extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)

View File

@ -71,6 +71,17 @@ executable jujitsu
, shadow-library
default-language: Haskell2010
executable parafiasrem
hs-source-dirs: app
main-is: parafiasrem.hs
ghc-options: -threaded -rtsopts -with-rtsopts=-N
build-depends: base
, hxt
, hxt-xpath
, MissingH
, regex-posix
, shadow-library
default-language: Haskell2010
source-repository head
type: git