diff options
| author | aarne <unknown> | 2003-12-04 12:08:29 +0000 |
|---|---|---|
| committer | aarne <unknown> | 2003-12-04 12:08:29 +0000 |
| commit | 15f94710f0403c760ed4ec1a8328c89400c4d94c (patch) | |
| tree | 93e1c81f180685a8a2afc4837aea511f48e312ca /src/GF/UseGrammar | |
| parent | 6a9dc9e5f5ddea8130b88a88d1e07f489d0906f9 (diff) | |
Added French for new API. Started alpha conv. Fixed bugs.
Diffstat (limited to 'src/GF/UseGrammar')
| -rw-r--r-- | src/GF/UseGrammar/Custom.hs | 4 | ||||
| -rw-r--r-- | src/GF/UseGrammar/Linear.hs | 17 | ||||
| -rw-r--r-- | src/GF/UseGrammar/Parsing.hs | 7 | ||||
| -rw-r--r-- | src/GF/UseGrammar/Tokenize.hs | 5 |
4 files changed, 23 insertions, 10 deletions
diff --git a/src/GF/UseGrammar/Custom.hs b/src/GF/UseGrammar/Custom.hs index 64cb29680..c117c0335 100644 --- a/src/GF/UseGrammar/Custom.hs +++ b/src/GF/UseGrammar/Custom.hs @@ -241,8 +241,8 @@ customTokenizer = ,(strCI "code", const $ lexHaskell) ,(strCI "text", const $ lexText) ,(strCI "unglue", \gr -> map tS . decomposeWords (stateMorpho gr)) ----- ,(strCI "codelit", lexHaskellLiteral . stateIsWord) ----- ,(strCI "textlit", lexTextLiteral . stateIsWord) + ,(strCI "codelit", lexHaskellLiteral . stateIsWord) + ,(strCI "textlit", lexTextLiteral . stateIsWord) ,(strCI "codeC", const $ lexC2M) ,(strCI "codeCHigh", const $ lexC2M' True) -- add your own tokenizers here diff --git a/src/GF/UseGrammar/Linear.hs b/src/GF/UseGrammar/Linear.hs index e60f8da79..0bd053803 100644 --- a/src/GF/UseGrammar/Linear.hs +++ b/src/GF/UseGrammar/Linear.hs @@ -42,13 +42,17 @@ linearizeToRecord gr mk m = lin [] where xs' <- mapM (\ (i,x) -> lin (i:ts) x) $ zip [0..] xs r <- case at of - A.AtC f -> look f >>= comp xs' + A.AtC f -> lookf c t f >>= comp xs' A.AtL s -> return $ recS $ tK $ prt at A.AtI i -> return $ recS $ tK $ prt at - A.AtV x -> lookCat c >>= comp [tK (prt at)] - A.AtM m -> lookCat c >>= comp [tK (prt at)] + A.AtV x -> lookCat c >>= comp [tK (prt_ at)] + A.AtM m -> lookCat c >>= comp [tK (prt_ at)] - return $ fmk $ mkBinds binds r + r' <- case r of -- to see stg in case the result is variants {} + FV [] -> lookCat c >>= comp [tK (prt_ t)] + _ -> return r + + return $ fmk $ mkBinds binds r' look = lookupLin gr . redirectIdent m . rtQIdent comp = ccompute gr @@ -60,6 +64,11 @@ linearizeToRecord gr mk m = lin [] where lookCat = return . errVal defLindef . look ---- should always be given in the module + -- to show missing linearization as term + lookf c t f = case look f of + Ok h -> return h + _ -> lookCat c >>= comp [tK (prt_ t)] + -- thus the special case: diff --git a/src/GF/UseGrammar/Parsing.hs b/src/GF/UseGrammar/Parsing.hs index b5b587c91..48b6ffac6 100644 --- a/src/GF/UseGrammar/Parsing.hs +++ b/src/GF/UseGrammar/Parsing.hs @@ -64,9 +64,10 @@ tokens2trms opts sg cn parser as = do _ | null ts0 -> checkWarn "No success in cf parsing" >> return [] _ | raw -> do ts1 <- return (map cf2trm0 ts0) ----- should not need annot - mapM (checkErr . (annotate gr) . trExp) ts1 ---- complicated + mapM (checkErr . (annotate gr) . trExp) ts1 ---- complicated; often fails _ -> do - (ts1,_) <- checkErr $ mapErr postParse ts0 + (ts1,ss) <- checkErr $ mapErr postParse ts0 + if null ts1 then raise ss else return () ts2 <- mapM (checkErr . (annotate gr) . trExp) ts1 ---- if forgive then return ts2 else do let tsss = [(t, allLinsOfTree gr cn t) | t <- ts2] @@ -75,7 +76,7 @@ tokens2trms opts sg cn parser as = do if null ps then raise $ "Failure in morphology." ++ if verb - then "\nPossible corrections: " +++++ + then "\nPossible corrections: " +++++ unlines (nub (map sstr (concatMap snd tsss))) else "" else return ps diff --git a/src/GF/UseGrammar/Tokenize.hs b/src/GF/UseGrammar/Tokenize.hs index ac28276f5..b264075ba 100644 --- a/src/GF/UseGrammar/Tokenize.hs +++ b/src/GF/UseGrammar/Tokenize.hs @@ -129,6 +129,9 @@ unknown2string isKnown = map mkOne where mkOne t@(TC s) = if isKnown s then t else mkTL s mkOne t = t -lexTextLiteral isKnown = unknown2string isKnown . lexText +lexTextLiteral isKnown = unknown2string (eitherUpper isKnown) . lexText lexHaskellLiteral isKnown = unknown2string isKnown . lexHaskell +eitherUpper isKnown w@(c:cs) = isKnown (toLower c : cs) || isKnown (toUpper c : cs) +eitherUpper isKnown w = isKnown w + |
