fixes in the script for the conversion of PennTreebank

This commit is contained in:
kr.angelov
2012-02-21 17:26:04 +00:00
parent 516c31835b
commit f7fa0395be
2 changed files with 5 additions and 4 deletions

View File

@@ -94,6 +94,7 @@ cidGenNP = mkCId "GenNP"
cidPredetNP = mkCId "PredetNP"
cidDetNP = mkCId "DetNP"
cidAdAP = mkCId "AdAP"
cidAdvAP = mkCId "AdvAP"
cidPositAdAAdj = mkCId "PositAdAAdj"
cideither7or_DConj = mkCId "either7or_DConj"
cidboth7and_DConj = mkCId "both7and_DConj"

View File

@@ -18,8 +18,8 @@ import Data.Tree
test = False
main = do
pgf <- readPGF "PennTreebank.pgf"
let Just language = readLanguage "PennTreebankCnc"
pgf <- readPGF "ParseEngAbs.pgf"
let Just language = readLanguage "ParseEng"
morpho = buildMorpho pgf language
s <- readFile "wsj.02-21"
ref <- newIORef (0,0,0)
@@ -84,8 +84,8 @@ penn =
v <- inside "JJ" (lemma "V2" "s VPPart")
pps <- many (cat "PP")
let adj = mkApp cidPastPartAP [mkApp v []]
ap0 = foldr (\ada ap -> mkApp cidAdAP [ada,ap]) (mkApp cidPositA [adj]) adas
ap = foldr (\pp ap -> mkApp cidAdAP [ap,pp]) ap0 pps
ap0 = foldr (\ada ap -> mkApp cidAdAP [ada,ap]) adj adas
ap = foldr (\pp ap -> mkApp cidAdvAP [ap,pp]) ap0 pps
return ap
`mplus`
do adas0 <- many pAdA