Compare commits
3 Commits
Author | SHA1 | Date | |
---|---|---|---|
|
b315a478a3 | ||
|
079e69cbba | ||
|
2eb795f797 |
48
app/jujitsu.hs
Normal file
48
app/jujitsu.hs
Normal file
@ -0,0 +1,48 @@
|
||||
|
||||
{-# 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[font]" -- pary adres-tytuł
|
||||
|
||||
-- ... a tutaj te trójki przerabiamy do docelowej struktury ShadowItem
|
||||
toShadowItem :: (String, String) -> ShadowItem
|
||||
toShadowItem (url, title) =
|
||||
(defaultShadowItem url title) {
|
||||
originalDate = date,
|
||||
itype = "periodical",
|
||||
format = Just "pdf",
|
||||
finalUrl = 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
|
||||
Just [[_, year]] -> year
|
||||
otherwise -> error $ "unexpected url: " ++ url
|
||||
|
||||
|
||||
main = do
|
||||
let start = "http://kobudo.pl/artykuly_jujutsu.html"
|
||||
let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
|
||||
lname="Jujitsu",
|
||||
abbrev="Juji",
|
||||
lLevel=0,
|
||||
webpage="http://kobudo.pl"}
|
||||
|
||||
extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)
|
41
app/parafiasrem.hs
Normal file
41
app/parafiasrem.hs
Normal 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)
|
@ -59,6 +59,29 @@ executable almanachmuszyny
|
||||
, shadow-library
|
||||
default-language: Haskell2010
|
||||
|
||||
executable jujitsu
|
||||
hs-source-dirs: app
|
||||
main-is: jujitsu.hs
|
||||
ghc-options: -threaded -rtsopts -with-rtsopts=-N
|
||||
build-depends: base
|
||||
, hxt
|
||||
, hxt-xpath
|
||||
, MissingH
|
||||
, regex-posix
|
||||
, 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
|
||||
|
19
stack.yaml.lock
Normal file
19
stack.yaml.lock
Normal 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
|
Loading…
Reference in New Issue
Block a user