separated the modules for compound and ordinary words in building DictFin. Both have about 40k lemmas

lib/src/finnish/kotus/Kotus.hs

@@ -2,7 +2,7 @@ import qualified Data.Map as M
 import qualified Data.Set as S
 import System
 
--- build DictFin from KOTUS word list. See Makefile.
+-- build DictFin from KOTUS word list. See Makefile for how to run
 -- AR 28/12/2010
 
 main = do
@@ -20,13 +20,16 @@ main = do
                           Just (end,beg) <- [lookCompound (whole,e) dictMap]]
 ---  mapM_ (\ (x,(y,_)) -> putStrLn (unwords [x,"-", word y])) compoundList -- to see compounds
   let (adjSet,advSet) = let (adjs,advs) = unzip adjList in (S.fromList adjs,S.fromList advs)
-  let dictList1 = map (mkAdjAdv adjSet advSet) dictList
-  let dictList2 = case xx of 
-        "-compounds":_ -> concatMap (mkCompound (M.fromList compoundList)) dictList1
-        "-all":_  -> dictList1 ++ concatMap (mkCompound (M.fromList compoundList)) dictList1
-        _ -> dictList1
-  let dict2 = map snd dictList2
-  mapM_ mkRules dict2
+  let compoundMap  = M.fromList compoundList
+  let dictList1    = map (mkAdjAdv adjSet advSet) dictList
+  let dictListComp = concatMap (mkCompound compoundMap) dictList1
+  let dictList2    = filter (flip M.notMember compoundMap . fst) dictList1
+  let dictList3 = case xx of 
+        "-compounds":_ -> dictListComp
+        "-all":_  -> dictList2 ++ dictListComp
+        _ -> dictList2
+  let dict3 = map snd dictList3
+  mapM_ mkRules dict3
 
 
 
@@ -95,7 +98,6 @@ adjCandidates adj = case reverse adj of
   _ -> [adj]
 
 
-
 -------------------------------------------------
 -- produce rules
 
@@ -104,7 +106,7 @@ mkRules e = do
   putRule $ mkLin (fun e) (par e) (word e)
  where
    putRule
-    | isDummy e = putStrLn . ("-- " ++)
+    | isDummy e = const (return ()) -- putStrLn . ("-- " ++)
     | isDerived e = putStrLn . ("--+ " ++)
     | isPlurTant e = putStrLn . ("--? " ++)
     | otherwise = putStrLn 

lib/src/finnish/kotus/Makefile

@@ -1,17 +1,19 @@
-gf: abs cnc
+all: dict
 
-src:
+dict:
 	runghc Kotus.hs src/kotus-sanalista_v1.xml >src.tmp
-
-compounds:
-	runghc Kotus.hs src/kotus-sanalista_v1.xml -compounds >src.tmp
-
-abs: src
 	cp prelDictFinAbs DictFinAbs.gf
 	grep ":" src.tmp >>DictFinAbs.gf
 	echo "}" >>DictFinAbs.gf
-
-cnc: src
 	cp prelDictFin DictFin.gf
 	grep "=" src.tmp >>DictFin.gf
 	echo "}" >>DictFin.gf
+
+compounds:
+	runghc Kotus.hs src/kotus-sanalista_v1.xml -compounds >src-comp.tmp
+	cp prelDictCompFinAbs DictCompFinAbs.gf
+	grep ":" src-comp.tmp >>DictCompFinAbs.gf
+	echo "}" >>DictCompFinAbs.gf
+	cp prelDictCompFin DictCompFin.gf
+	grep "=" src-comp.tmp >>DictCompFin.gf
+	echo "}" >>DictCompFin.gf

lib/src/finnish/kotus/prelDictCompFin

@@ -0,0 +1,6 @@
+--# -path=.:alltenses
+
+concrete DictCompFin of DictCompFinAbs = DictFin ** open MorphoFin, Kotus, Prelude in {
+
+flags coding = utf8 ;
+

lib/src/finnish/kotus/prelDictCompFinAbs

@@ -0,0 +1,2 @@
+abstract DictCompFinAbs = DictFinAbs ** {
+