From 210b80656e4142ae9348bcb7284b89b8e7409ec9 Mon Sep 17 00:00:00 2001
From: AdamOsiowy123 <adaosi@st.amu.edu.pl>
Date: Mon, 11 Apr 2022 10:49:27 +0200
Subject: [PATCH] add custom method to extract links

---
 app/miastobierun.hs | 25 ++++++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)

diff --git a/app/miastobierun.hs b/app/miastobierun.hs
index 5e9c064..cea4e26 100644
--- a/app/miastobierun.hs
+++ b/app/miastobierun.hs
@@ -9,11 +9,26 @@ import Data.List.Utils (replace)
 import Text.Regex.Posix
 import Text.Printf
 
+--extractCustomLinks xpathCondition1 = (downloadDocument &&& this)
+--                                       >>> first (getXPathTrees xpathCondition1
+--                                                  >>> (getAttrValue "href"
+--                                                       &&& getAttrValue "title"
+--                                                       ))
+--                                       >>> arr rotateSecTh
+--                                       >>> first expandURIFixed
 
-extractRecords = extractLinksWithText "//td[@cell-header='Nazwa pliku']/a"  -- pary adres-tytuł
-                 >>> second (arr $ replace "\r\n" "") -- czyścimy drugi element pary, czyli tytuł z niepotrzebnych białych znaków
+extractCustomLinks xpathCondition1 xpathCondition2 xpathCondition3= (downloadDocument &&& this)
+                                       >>> first (getXPathTrees xpathCondition1
+                                                  >>> first (getXPathTrees xpathCondition2 >>> (getAttrValue "href" &&& getAttrValue "title"))
+                                                  >>> second (getXPathTrees xpathCondition3 >>> getAttrValue "cell-header")
+                                                 )
+                                       >>> arr rotateSecThX
+                                       >>> first expandURIFixed
+                                       >>> second expandURIFixed
+
+extractRecords = extractCustomLinks "//tr[@class=''] | //tr[@class='_a']" "//td[@cell-header='Nazwa pliku']/a" "//td[@class='size']"
+--extractRecords = extractCustomLinks "//td[@cell-header='Nazwa pliku']/a"
 
--- ... a tutaj te dwójki przerabiamy do docelowej struktury ShadowItem
 toShadowItem :: (String, String) -> ShadowItem
 toShadowItem (url, articleTitle) =
   (defaultShadowItem url title) {
@@ -23,9 +38,9 @@ toShadowItem (url, articleTitle) =
     finalUrl = url,
     description = Just desc
     }
-  where title = "Miasto Bierun: " ++ articleTitle
+  where title = "Miasto Bierun: " ++ " url: " ++ url ++ " title: " ++ articleTitle
         date = getDate articleTitle
-        desc = getArticleNr articleTitle
+        desc = getArticleNr articleTitle ++ " size: "
 
 getDate title =
   case title =~~ "(19[0-9][0-9]|20[0-9][0-9])" :: Maybe [[String]] of