DocBook reader improvements.

1 files changed, 49 insertions, 18 deletions

diff --git a/src/Text/Pandoc/Readers/DocBook.hs b/src/Text/Pandoc/Readers/DocBook.hs
index 73a2e6abc..b8ff3f1ff 100644
--- a/src/Text/Pandoc/Readers/DocBook.hs
+++ b/src/Text/Pandoc/Readers/DocBook.hs
@@ -5,32 +5,63 @@ import Text.Pandoc.Builder
 import Text.XML.Light
 import Data.Monoid
 import Data.Char (isSpace)
+import Control.Monad.State
+import Control.Applicative ((<$>))
+
+type DB = State DBState
+
+data DBState = DBState{ dbSectionLevel :: Int }
+             deriving (Read, Show)
+
+defaultDBState :: DBState
+defaultDBState = DBState { dbSectionLevel = 0 }
 
 readDocBook :: ParserState -> String -> Pandoc
 readDocBook st inp = Pandoc (Meta [] [] []) $ toList blocks
-  where blocks = mconcat $ map (parseBlock st) $ parseXML inp
+  where blocks = mconcat $ evalState (mapM parseBlock $ parseXML inp)
+                             defaultDBState
 
-parseBlock :: ParserState -> Content -> Blocks
-parseBlock st (Text (CData _ s _)) = if all isSpace s
-                                        then mempty
-                                        else plain $ text s
-parseBlock st (Elem e) =
+parseBlock :: Content -> DB Blocks
+parseBlock (Text (CData _ s _)) = if all isSpace s
+                                        then return mempty
+                                        else return $ plain $ text s
+parseBlock (Elem e) =
   case qName (elName e) of
-        "para" -> para $ trimInlines $ mconcat
-                       $ map (parseInline st) $ elContent e
-        _      -> mconcat $ map (parseBlock st) $ elContent e
-parseBlock st (CRef _) = mempty
+        "para"  -> para <$> getInlines e
+        "sect1" -> sect 1
+        "sect2" -> sect 2
+        "sect3" -> sect 3
+        "sect4" -> sect 4
+        "sect5" -> sect 5
+        "sect6" -> sect 6
+        "title" -> return $ mempty
+        _       -> innerBlocks
+   where innerBlocks = mconcat <$> (mapM parseBlock $ elContent e)
+         getInlines e' = (trimInlines . mconcat) <$>
+                       (mapM parseInline $ elContent e')
+         isTitle e' = qName (elName e') == "title"
+         skipWhite (Text (CData _ s _):xs) | all isSpace s = skipWhite xs
+                                           | otherwise     = xs
+         skipWhite xs = xs
+         sect n = case skipWhite (elContent e) of
+                        ((Elem t):body)
+                          | isTitle t -> do
+                                  h <- header n <$> (getInlines t)
+                                  b <- mconcat <$> (mapM parseBlock body)
+                                  return $ h <> b
+                        _        -> (header n mempty <>) <$> innerBlocks
+parseBlock (CRef _) = return mempty
 
-parseInline :: ParserState -> Content -> Inlines
-parseInline st (Text (CData _ s _)) = text s
-parseInline st (Elem e) =
+parseInline :: Content -> DB Inlines
+parseInline (Text (CData _ s _)) = return $ text s
+parseInline (Elem e) =
   case qName (elName e) of
         "emphasis" -> case lookupAttrBy (\attr -> qName attr == "role")
                            (elAttribs e) of
-                             Just "strong" -> strong innerInlines
-                             _             -> emph innerInlines
+                             Just "strong" -> strong <$> innerInlines
+                             _             -> emph <$> innerInlines
         _          -> innerInlines
-   where innerInlines = trimInlines . mconcat . map (parseInline st)
-                        $ elContent e
-parseInline st (CRef _) = mempty
+   where innerInlines = (trimInlines . mconcat) <$>
+                          (mapM parseInline $ elContent e)
+parseInline (CRef _) = return mempty