From 092ef7b8f14be7f6b523a31404be7e69bd0d6f50 Mon Sep 17 00:00:00 2001
From: Marko Andjelic <markoandjelic31@gmail.com>
Date: Sat, 24 Jan 2026 00:56:23 +0000
Subject: [PATCH] Implement scalpel

---
 app/EpubParser.hs | 30 +++++++++---------------------
 svitak.cabal      |  3 ++-
 2 files changed, 11 insertions(+), 22 deletions(-)

diff --git a/app/EpubParser.hs b/app/EpubParser.hs
index f9bea9e..7d6e08c 100644
--- a/app/EpubParser.hs
+++ b/app/EpubParser.hs
@@ -3,7 +3,8 @@ module EpubParser where
 import Codec.Archive.Zip (toArchive, findEntryByPath, fromEntry, Archive)
 import qualified Data.ByteString.Lazy as BL
 import qualified Data.ByteString.Lazy.Char8 as C8
-import Text.HTML.TagSoup (parseTags, fromAttrib, (~==), Tag, innerText, sections)
+import Text.HTML.TagSoup (parseTags, Tag)
+import Text.HTML.Scalpel (scrape, text, tagSelector, attr)
 import System.FilePath (takeDirectory)
 import Control.Monad.Reader
 import qualified Data.Text as T
@@ -36,17 +37,10 @@ openEpub path = toArchive <$> BL.readFile path >>= \archive -> maybe (pure $ Lef
 getOpfPath :: Archive -> Maybe FilePath
 getOpfPath archive = do
   entry <- findEntryByPath "META-INF/container.xml" archive
-  let content = C8.unpack (fromEntry entry)
+  let tags = parseTags $ C8.unpack (fromEntry entry)
+  -- Logic: scrape (the_scraper) (the_tags)
+  scrape (attr "full-path" (tagSelector "rootfile")) tags
 
-  let tags = parseTags content
-
-  let rootfileTags = filter (~== "<rootfile>") tags
-
-  -- If we found at least one, get the "full-path" attribute
-  case rootfileTags of
-    (t:_) -> Just (fromAttrib "full-path" t)
-    []    -> Nothing
-  
 getOpfTags :: Archive -> FilePath -> [Tag String]
 getOpfTags archive opfpath =
   case findEntryByPath opfpath archive of
@@ -56,16 +50,10 @@ getOpfTags archive opfpath =
 getRootPrefix :: FilePath -> FilePath
 getRootPrefix path = 
     let dir = takeDirectory path
-    in if dir == "." 
-       then "" 
-       else dir ++ "/"
-
-tagify :: String -> String
-tagify item = "<" ++ item ++ ">"
+    in if dir == "."  then "" else dir ++ "/"
 
 getTagText :: String -> [Tag String] -> T.Text
-getTagText tagName tags =
-  case sections (~== tagify tagName) tags of
-    (x:_) -> T.pack (innerText x)
-    []    -> T.empty
+getTagText tagName tags = 
+  let scraper = text (tagSelector tagName)
+  in T.pack $ fromMaybe "" (scrape scraper tags)
 
diff --git a/svitak.cabal b/svitak.cabal
index e710b44..578ee42 100644
--- a/svitak.cabal
+++ b/svitak.cabal
@@ -66,7 +66,8 @@ executable svitak
                       mtl >= 2.3.2,
                       tagsoup,
                       text >= 2.1.2,
-                      zip-archive >= 0.4.3.2
+                      zip-archive >= 0.4.3.2,
+                      scalpel >= 0.6.2.2
 
     hs-source-dirs:   app
     default-language: Haskell2010