This commit is contained in:
s434695 2021-04-13 19:15:17 +02:00
parent 8f3e9c8fae
commit f14d53337e
4 changed files with 76 additions and 1 deletions

View File

@ -4,7 +4,7 @@ import ShadowLibrary.Core
import Text.XML.HXT.Core import Text.XML.HXT.Core
import Text.XML.HXT.XPath import Text.XML.HXT.XPath
--import Text.XML.HXT.Curl import Text.XML.HXT.Curl
import Data.List import Data.List
import Data.List.Utils (replace) import Data.List.Utils (replace)

43
app/pbsociety.hs Normal file
View File

@ -0,0 +1,43 @@
{-# 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='image-link']" -- pary adres-tytuł
>>> second (arr $ replace "\n" "")
>>> first (extractLinksWithText "//a[@class='image-link']")
-- ... 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 = "Pbsociety " ++ yearlyTitle
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://pbsociety.org.pl/repository/discover?filtertype_1=has_content_in_original_bundle&filter_relational_operator_1=equals&filter_1=true&filtertype_2=title&filter_relational_operator_2=contains&filter_2=&submit_apply_filter=Apply&query=&scope=%2F&rpp=260&sort_by=score&order=desc"
let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
lname="Polish Botanical Society Repository",
abbrev="Pbsociety",
lLevel=0,
webpage=start}
extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)

View File

@ -61,6 +61,19 @@ executable almanachmuszyny
, shadow-library , shadow-library
default-language: Haskell2010 default-language: Haskell2010
executable pbsociety
hs-source-dirs: app
main-is: pbsociety.hs
ghc-options: -threaded -rtsopts -with-rtsopts=-N
build-depends: base
, hxt
, hxt-xpath
, hxt-curl
, MissingH
, regex-posix
, shadow-library
default-language: Haskell2010
source-repository head source-repository head
type: git type: git

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