From f38b6b716b20dafcbf43b804194b5b8f519e4caa Mon Sep 17 00:00:00 2001
From: kuba <s478874@st.amu.edu.pl>
Date: Mon, 18 Apr 2022 11:15:43 +0200
Subject: [PATCH] msl robot

---
 app/msl.hs             |  67 +++++++++++++++++++++++++++++++++++++++++
 packages.microsoft.gpg | Bin 0 -> 641 bytes
 stack.yaml.lock        |  19 ++++++++++++
 3 files changed, 86 insertions(+)
 create mode 100644 app/msl.hs
 create mode 100644 packages.microsoft.gpg
 create mode 100644 stack.yaml.lock

diff --git a/app/msl.hs b/app/msl.hs
new file mode 100644
index 0000000..deb3f6f
--- /dev/null
+++ b/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)
diff --git a/packages.microsoft.gpg b/packages.microsoft.gpg
new file mode 100644
index 0000000000000000000000000000000000000000..0cffae08d061d6ae8959b4442dd2dbce26c340e8
GIT binary patch
literal 641
zcmV-{0)G9O0SyFJF<6WN2mrt;hGAw>aTw8-&ZU0T;kCvj9-@p2hRX&3tKIp#mk44K
zWSkQ%c4{HQA}s#tsn;1=(LHS^JIEF`F{wkR(&D>9o|7yP=~Y+NKK3;C`4axI=rYyq
zrB<*k)Bs<eZ?MAk^X?R}9mZpXVuRz(v^#Jq(pQ2|#0sHPg!(!*??tS}mv*2z=(9a}
zk(M(pQA=$d24_YaWzI>kBDEsDx5Yd^pAvz|*XxLN^Rs1tKcJz?E$yQpt@e`p4a|{c
zMCET7wg8i5k#M`dMmXStEMdYRelqXC@i+Karv74@Qsaue67_2{<S%2g%}D3pNbZuB
z%)O#322Iz*-`T?Ml~Di@0RRECH%)0{a&L2QW^^DZQe|vqVRL05b7^O8X>MmJAUtPq
zXLDs^b#iHRc|dJxV{&hEZ)S8ZV{dIfi2*eP69EDM9|ZzdF<6WO8v_Ol2?z%R0t6KT
z2m=Ea0s#UZ0Rk6*0162Z>pqmNz7i?V8k-0IeXlDmy&evAAJ<<LlQGlwzZdRrDd3(~
zsHVRp%+r9rTtiK{CfH|pkvwsS6uO=;(Xp}i`wO;l5#_rl`_j!A2gYvMyhNh0MiCaR
ztb#&*98nQAx}S0`Ic?ZxtxGcPC9wA<@pMz9xe78UJ?#c)hJ52tGEiDL#hn7kocFG4
zyqg{8HX3D-NYzXekX6+vn|E%V&4PVz5ry!IF_UZISI?n~9_NX`T?p>zH{duFkO6N3
z&j)GB_6-&>FO*4L(_oK+UEBCNPtqruWb}jtrH<<UbeZ!uBSMk6{VBEkC;E!ihSBpb
bdebCk7R5d_(1lHI@MA$+)%A({@{_PGwc;jf

literal 0
HcmV?d00001

diff --git a/stack.yaml.lock b/stack.yaml.lock
new file mode 100644
index 0000000..cc80ef3
--- /dev/null
+++ b/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