diff options
| author | aarne <aarne@cs.chalmers.se> | 2008-12-15 15:45:36 +0000 |
|---|---|---|
| committer | aarne <aarne@cs.chalmers.se> | 2008-12-15 15:45:36 +0000 |
| commit | e478228cb5738d67c7fa10e39e32013f8e99d955 (patch) | |
| tree | 306e7677fbb707f9730c8f20c6eee01f6677f05e | |
| parent | 10675e29cac9149c18fbf61903e52468426a0ea1 (diff) | |
better treatment of syncategorematic words in alignment
| -rw-r--r-- | src/PGF/VisualizeTree.hs | 16 |
1 files changed, 13 insertions, 3 deletions
diff --git a/src/PGF/VisualizeTree.hs b/src/PGF/VisualizeTree.hs index a15e380d6..ae91fd9de 100644 --- a/src/PGF/VisualizeTree.hs +++ b/src/PGF/VisualizeTree.hs @@ -65,7 +65,10 @@ lin2graph ss = prelude ++ nodes ++ links prelude = ["rankdir=LR ;", "node [shape = record] ;"] - -- find all words + -- the plain string, with syncategorematic words included + strings = filter (flip notElem "{[()]}" . head) . words + + -- find all lexicalized words lins :: String -> [(String,String)] lins [] = [] lins s = let (s1, s2) = if null s then ([],[]) else span (/='{') s in @@ -75,11 +78,18 @@ lin2graph ss = prelude ++ nodes ++ links -- separate a word to the link (1,2,3) and the word itself wlink :: String -> (String,String) wlink s = let (s1, s2) = span (/=']') s in - (tail s1, init (drop 1 s2)) + (tail s1, unwords (words (init (drop 1 s2)))) + + -- to merge in syncat words + slins i s = merge (strings s) (lins s) where + merge ws cs = case (ws,cs) of + (w:ws2,(m,c):cs2) | w==c -> (m,c) : merge ws2 cs2 + (w:ws2,_ ) -> ("w" ++ show i,w) : merge ws2 cs + _ -> [] -- make all marks unique to deal with discontinuities nlins :: [(Int,[((Int,String),String)])] - nlins = [(i, [((j,m),w) | (j,(m,w)) <- zip [0..] (lins s)]) | (i,s) <- zip [0..] ss] + nlins = [(i, [((j,m),w) | (j,(m,w)) <- zip [0..] (slins i s)]) | (i,s) <- zip [0..] ss] nodes = map mkStruct nlins |