better treatment of syncategorematic words in alignment

This commit is contained in:
aarne
2008-12-15 15:45:36 +00:00
parent 2bccc3e405
commit 9982b15b5b

View File

@@ -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