ShadowLibrary/Core.hs

@@ -33,7 +33,7 @@ import Data.Tree.NTree.TypeDefs
 import Data.Maybe
 import Control.Monad.Trans
 import Text.XML.HXT.XPath
-import Text.XML.HXT.Curl
+-- import Text.XML.HXT.Curl
 import Text.XML.HXT.HTTP
 
 import Text.Regex.TDFA
@@ -64,8 +64,8 @@ downloadDocument = readFromDocument [withParseHTML yes,
                                      withEncodingErrors no,
                                      withPreserveComment yes,
                                      withStrictInput yes,
---                                     withHTTP []
+                                     withHTTP []
-                                     withCurl [("curl--user-agent","AMU Digital Libraries Indexing Agent")]
+--                                     withCurl [("curl--user-agent","AMU Digital Libraries Indexing Agent")]
                                     ]
 
 downloadDocumentWithEncoding enc = readFromDocument [withParseHTML yes,
@@ -73,13 +73,13 @@ downloadDocumentWithEncoding enc = readFromDocument [withParseHTML yes,
                                                      withEncodingErrors no,
                                                      withPreserveComment yes,
                                                      withInputEncoding enc,
---					             withHTTP []]		
+					             withHTTP []]		
-                                                     withCurl []]
+--                                                     withCurl []]
 
 downloadXmlDocument = readFromDocument [withWarnings no,
                                         withEncodingErrors no,
---					withHTTP []]
+					withHTTP []]
-                                        withCurl [] ]
+--                                        withCurl [] ]
 
 
 data ShadowLibrary = ShadowLibrary { logoUrl :: Maybe String,

app/instytutTechnikiGorniczej.hs

@@ -1,49 +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 = extractLinksWithText "//li/a[contains(@href,'/lata')]"
-                   >>> first (extractLinksWithText "//li/a[contains(@href,'start=')]"
-                               >>> first (extractLinksWithText "//a[contains(@href, '/archiwum/images/19') or contains(@href, '/Trybuna')]"))
-
-
-toShadowItem :: (((String, String), String), String) -> ShadowItem
-toShadowItem (((url, articleTitle), _), yearlyTitle) =
-  (defaultShadowItem url title) {
-    originalDate = Just date,
-    itype = "periodical",
-    format = Just form,
-    finalUrl = url
-    }
-  where title = "Instytut Techniki Gorniczej " ++ yearlyTitle ++ " " ++ (replace "\r\n" "" (replace "\r\n          " "" articleTitle))
-        date = getDate $ replace "%20" "" url
-        form = unwords $ map stringToLower [getFormat url]
-
-getDate url =
-  case url =~~ "(19[0-9][0-9]|20[0-9][0-9]).*\\..{3}" :: Maybe [[String]] of
-    Just [[_, year]] -> year
-    otherwise -> error $ "unexpected url: " ++ url
-    
-getFormat url =
-  case url =~~ "([a-zA-Z]+$)" :: Maybe [[String]] of
-    Just [[_, format]] -> format
-    otherwise -> error $ "unexpected url: " ++ url
-
-main = do
-    let start = "http://komag.gliwice.pl/archiwum/historia-komag/"
-    let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
-                                       lname="Instytut Techniki Gorniczej",
-                                       abbrev="Komag",
-                                       lLevel=0,
-                                       webpage=start}
-    extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadowItem)

shadow-library.cabal

@@ -20,7 +20,6 @@ library
                      , HTTP
                      , hxt
                      , hxt-http
-                     , hxt-curl
                      , hxt-xpath
                      , MissingH
                      , monad-logger
@@ -60,17 +59,6 @@ executable almanachmuszyny
                       , shadow-library
    default-language:    Haskell2010
 
-executable instytutTechnikiGorniczej
-   hs-source-dirs:      app
-   main-is:             instytutTechnikiGorniczej.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

stack.yaml.lock

@@ -1,19 +0,0 @@
-# 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