almost

app/czasopisma_francja.hs

@@ -0,0 +1,49 @@
+
+{-# 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,'.pdf') and contains(text(),'czasopisma_francja')]"
+
+toShadowItem :: (String, String) -> ShadowItem
+toShadowItem (url, articleTitle) =
+  (defaultShadowItem url title) {
+    originalDate = Just date,
+    itype = "periodical",
+    format = Just "pdf",
+    finalUrl = url
+    }
+  where title = "czasopisma_francja " ++ getNr url
+        date = getDate url
+
+replaceSeparatorWithSpace = map (\c -> if c=='_' then ' '; else if c=='-' then ' '; else c)
+replaceSeparatorWithDot = map (\c -> if c=='_' then '.'; else c)
+
+getNr url =
+  case url =~~ "nr[-_][0-9][0-9]" :: Maybe String of
+    Just nr -> replaceSeparatorWithSpace nr
+    otherwise -> error $ "unexpected url: " ++ url
+
+getDate url =
+  case url =~~ "20[0-9][0-9]_[0-1][0-9]" :: Maybe String of
+    Just date -> replaceSeparatorWithDot date
+    otherwise -> error $ "unexpected url: " ++ url
+
+
+main = do
+    let start = "http://www.almanachmuszyny.pl/"
+    let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
+                                       lname="czasopisma_francja",
+                                       abbrev="CzasFr",
+                                       lLevel=0,
+                                       webpage=start}
+    extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)

shadow-library.cabal

@@ -58,6 +58,18 @@ executable almanachmuszyny
                       , regex-posix
                       , shadow-library
    default-language:    Haskell2010
+   
+   executable czasopisma_francja
+   hs-source-dirs:      app
+   main-is:             czasopisma_francja.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

stack.yaml.lock

@@ -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