app/teatrLalek.hs

@@ -1,39 +0,0 @@
-
-{-# 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 = extractLinks "//a[contains(@href, '.pdf')]"
-
-
-toShadowItem :: String -> ShadowItem
-toShadowItem url =
-  (defaultShadowItem url title) {
-    originalDate = Just date,
-    itype = "periodical",
-    format = Just "pdf",
-    finalUrl = url
-    }
-  where
-    date = last $ getAllTextMatches $ url =~ "(19[0-9][0-9]|20[0-9][0-9])" :: String
-    titleToProcess = last $ getAllTextMatches $ url =~ "/[a-zA-Z ]+[-_]" :: String
-    title = titleToProcess =~ "[a-zA-Z ]+" :: String
-
-
-main = do
-    let start = "http://polunima.pl/teatr-lalek/"
-    let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
-                                       lname="Teatr Lalek",
-                                       abbrev="Teatr",
-                                       lLevel=0,
-                                       webpage=start}
-    extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)

shadow-library.cabal

@@ -34,19 +34,18 @@ library
                      , tz
   default-language:    Haskell2010
 
-executable teatrLalek
+-- executable maly-modelarz-exe
-    hs-source-dirs:      app
+--    hs-source-dirs:      app
-    main-is:             teatrLalek.hs
+--    main-is:             malymodelarz.hs
-    ghc-options:         -threaded -rtsopts -with-rtsopts=-N
+--    ghc-options:         -threaded -rtsopts -with-rtsopts=-N
-    build-depends:       base
+--    build-depends:       base
-                       , hxt
+--                       , hxt
-                       , hxt-curl
+--                       , hxt-curl
-                       , hxt-xpath
+--                       , hxt-xpath
-                       , MissingH
+--                       , MissingH
-                       , regex-posix
+--                       , regex-posix
-                       , shadow-library
+--                       , shadow-library
-                       , regex-tdfa
+--    default-language:    Haskell2010
-    default-language:    Haskell2010
 
 executable almanachmuszyny
    hs-source-dirs:      app