Return as ShadowLibrary

This commit is contained in:
Jakub Kaczmarek 2022-03-16 16:31:16 +01:00
parent 2eb795f797
commit 079e69cbba
2 changed files with 31 additions and 11 deletions

View File

@ -12,23 +12,24 @@ import Text.Regex.Posix
import Text.Printf
extractRecords = extractLinksWithText "//a[img]" -- pary adres-tytuł
-- >>> second (arr $ replace "\r\n " " ") -- czyścimy drugi element pary, czyli tytuł z niepotrzebnych białych znaków
-- >>> first (arr ((++"tr") . init)) -- modyfikujemy pierwszy element pary, czyli adres URL
-- >>> 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)
extractRecords = extractLinksWithText "//a[font]" -- pary adres-tytuł
-- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem
toShadowItem :: ((String, String), String) -> ShadowItem
toShadowItem ((url, articleTitle), yearlyTitle) =
toShadowItem :: (String, String) -> ShadowItem
toShadowItem (url, title) =
(defaultShadowItem url title) {
originalDate = Just date,
originalDate = date,
itype = "periodical",
format = Just "pdf",
finalUrl = url
}
where title = "Jujitsu " ++ yearlyTitle ++ " " ++ (replace "\r\n" "" (replace "\r\n " "" articleTitle))
date = getDate url
where date = extractYear title
extractMonth :: String -> Maybe String
extractMonth n =
case n =~~ ("/(styczeń|luty|marzec|kwiecien|maj|czerwiec|lipiec|sierpień|wrzesień|październik|listopad|grudzień)/" :: String) of
Just month -> Just month
otherwise -> Nothing
getDate url =
case url =~~ "/(19[0-9][0-9]|20[0-9][0-9])/" :: Maybe [[String]] of
@ -44,4 +45,4 @@ main = do
lLevel=0,
webpage="http://kobudo.pl"}
extractItemsStartingFromUrl shadowLibrary start (extractRecords)
extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)

19
stack.yaml.lock Normal file
View File

@ -0,0 +1,19 @@
# This file was autogenerated by Stack.
# You should not edit this file by hand.
# For more information, please see the documentation at:
# https://docs.haskellstack.org/en/stable/lock_files
packages:
- completed:
hackage: hxt-xpath-9.1.2.2@sha256:9cd590ae93a04573db8f90fa4094625ebd97dded45da7667c577ce6b38a42900,1999
pantry-tree:
size: 2225
sha256: aee2f75974e868ff429b8ff349a29667536c60397098f5dfedc968d1951511bb
original:
hackage: hxt-xpath-9.1.2.2
snapshots:
- completed:
size: 507596
url: https://raw.githubusercontent.com/commercialhaskell/stackage-snapshots/master/lts/11/9.yaml
sha256: 42f472dbf06482da1b3319241f3e3b3593a45bd7d4f537d2789f21386b9b2ad3
original: lts-11.9