Implement scalpel

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
@@ -36,16 +37,9 @@ 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 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
+  let tags = parseTags $ C8.unpack (fromEntry entry)
+  -- Logic: scrape (the_scraper) (the_tags)
+  scrape (attr "full-path" (tagSelector "rootfile")) tags
 
 getOpfTags :: Archive -> FilePath -> [Tag String]
 getOpfTags archive opfpath =
@@ -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
+  let scraper = text (tagSelector tagName)
+  in T.pack $ fromMaybe "" (scrape scraper tags)
 

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