module Main where

import Util
import qualified Data.Map as Map
import qualified Data.Set as Set

main = do 
  s1 <- readFile "/home/jonatan/unified_sblex_update/dalin.txt"
  s2 <- getContents
  putStrLn $ process s2 (table (lines s1) Map.empty)

process [] m = []
process ('<':'L':'e':'x':'i':'c':'o':'n':'>':s) m = 
  "<Lexicon xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xsi:schemaLocation=\"https://svn.spraakdata.gu.se/sb-arkiv/pub/lexikon/dalin/dalin.xsd\">" ++ process s m
process ('<':'e':'i':'d':'>':s) m = 
    case span (/= '<') s of
      (lem,_:_:_:_:_:_:_:r) -> 
          case Map.lookup lem m of
            Nothing -> concat ["<eid>", lem, "</eid>\n"] ++ process r m
            Just s  -> concat ("<eid>" : lem : "</eid>\n" : 
                               [concat ["    <saldo>",s,"</saldo>\n"] 
                                         | s <- Set.toList s]) ++ process r m
process (c:cs) m = c:process cs m

table [] m = m
table (s:ss) m = table ss (ins s m)
 where ins s m = case untab s of
                   [_,_,_,_,_,lex,lem,_] -> Map.insertWith 
                                                Set.union 
                                                   lem (Set.fromList (words lex)) m
                   _ -> m