msl robot

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/almanachmuszyny.hs

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

app/msl.hs

@@ -0,0 +1,67 @@
+
+{-# 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
+
+
+extractH xpathCondition  = (downloadDocument &&& this)
+                            >>> first(getXPathTrees xpathCondition  
+                            >>> (listA (deep isText >>> getText)
+                            >>> arr (intercalate " ")
+                            ))
+
+extractInfo xpathCondition  = (downloadDocument &&& this)
+                            >>> first (getXPathTrees xpathCondition 
+                                >>> (getAttrValue "href"
+                                &&& (getXPathTrees "//a[@class='msl-box-link']//div[@class='msl-box-title']"
+                                    >>> (listA (deep isText >>> getText))
+                                    >>> arr (intercalate " ")
+                                    )
+                                &&& (getXPathTrees "//a[@class='msl-box-link']//div[@class='msl-box-info']"
+                                    >>> (listA (deep isText >>> getText))
+                                    >>> arr (intercalate " ")
+                                    )
+                                &&& (getXPathTrees "//a[@class='msl-box-link']//div[@class='time']"
+                                    >>> (listA (deep isText >>> getText))
+                                    >>> arr (intercalate " ")
+                                    )
+                                ))
+
+
+
+extractRecords = extractInfo "//a[@class='msl-box-link']"
+
+
+
+toShadow :: ((String, (String, (String, String))), String) -> ShadowItem
+toShadow ((url, (title, (author, year))), s) =
+    ShadowItem {
+        url = Just url,
+        title = title,
+        originalDate = Just year,
+        creator = Just author,
+        format = Just s,
+        itype = "periodical",
+        lang = Nothing,
+        finalUrl = url,
+        description = Nothing 
+        }
+
+
+main = do
+    --let start = "https://zasoby.msl.org.pl/mobjects/show"
+    let start = "https://zasoby.msl.org.pl/mobjects/show?date_range=&otype%5B4%5D=1&otype%5B5%5D=1&otype%5B6%5D=1&otype%5B7%5D=1" --przefiltrowane
+    let shadowLibrary = ShadowLibrary {logoUrl=Nothing,
+                                       lname="Zasoby MSL",
+                                       abbrev="MSL",
+                                       lLevel=0,
+                                       webpage=start}
+    extractItemsStartingFromUrl shadowLibrary start (extractRecords >>> arr toShadow)

shadow-library.cabal

@@ -59,6 +59,18 @@ executable almanachmuszyny
                       , shadow-library
    default-language:    Haskell2010
 
+executable msl
+    hs-source-dirs:      app
+    main-is:             msl.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

@@ -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