command eb for example-based grammar conversion; see 'help eb' and the example in examples/animals/QuestionsI.gfe

examples/animals/Animals.gf

@@ -0,0 +1,12 @@
+-- The Question grammar specialized to animals.
+
+abstract Animals = Questions ** {
+
+  flags startcat=Phrase ;
+
+  fun
+    -- a lexicon of animals and actions among them
+    Dog, Cat, Mouse, Lion, Zebra : Entity ;
+    Chase, Eat, See : Action ;
+}
+

examples/animals/AnimalsEng.gf

@@ -0,0 +1,16 @@
+--# -path=.:present:prelude
+--resource/english:resource/abstract:resource/../prelude
+
+concrete AnimalsEng of Animals = QuestionsEng **
+  open LangEng, ParadigmsEng, IrregEng in {
+
+  lin
+    Dog = regN "dog" ;
+    Cat = regN "cat" ;
+    Mouse = mk2N "mouse" "mice" ;
+    Lion = regN "lion" ;
+    Zebra = regN "zebra" ;
+    Chase = dirV2 (regV "chase") ;
+    Eat = dirV2 eat_V ;
+    See = dirV2 see_V ;
+}

examples/animals/AnimalsFre.gf

@@ -0,0 +1,15 @@
+--# -path=.:present:prelude
+
+concrete AnimalsFre of Animals = QuestionsFre **
+  open LangFre, ParadigmsFre, IrregFre in {
+
+  lin
+    Dog   = regN "chien" ;
+    Cat   = regN "chat" ;
+    Mouse = regGenN "souris" feminine ;
+    Lion  = regN "lion" ;
+    Zebra = regGenN "zèbre" masculine ;
+    Chase = dirV2 (regV "chasser") ;
+    Eat   = dirV2 (regV "manger") ;
+    See   = voir_V2 ;
+}

examples/animals/AnimalsSwe.gf

@@ -0,0 +1,15 @@
+--# -path=.:present:prelude
+
+concrete AnimalsSwe of Animals = QuestionsSwe **
+  open LangSwe, ParadigmsSwe, IrregSwe in {
+
+  lin
+    Dog = regN "hund" ;
+    Cat = mk2N "katt" "katter" ;
+    Mouse = mkN "mus" "musen" "möss" "mössen" ;
+    Lion = mk2N "lejon" "lejon" ;
+    Zebra = regN "zebra" ;
+    Chase = dirV2 (regV "jaga") ;
+    Eat = dirV2 äta_V ;
+    See = dirV2 se_V ;
+}

examples/animals/Questions.gf

@@ -0,0 +1,10 @@
+-- Simple questions and answers, in present tense.
+
+abstract Questions = {
+  cat 
+    Phrase ; Entity ; Action ;
+  fun
+    Who    : Action -> Entity -> Phrase ;           -- who chases X
+    Whom   : Entity -> Action -> Phrase ;           -- whom does X chase
+    Answer : Entity -> Action -> Entity -> Phrase ; -- X chases Y
+}

examples/animals/QuestionsEng.gf

@@ -0,0 +1,2 @@
+concrete QuestionsEng of Questions = QuestionsI with 
+  (Lang = LangEng) ;

examples/animals/QuestionsFre.gf

@@ -0,0 +1,2 @@
+concrete QuestionsFre of Questions = QuestionsI with 
+  (Lang = LangFre) ;

examples/animals/QuestionsI.gf

@@ -0,0 +1,27 @@
+-- to compile: echo "eb -file=QuestionsI.gfe" | gf $GF_LIB_PATH/present/LangEng.gfo
+-- or use directly gf <mkAnimals.gfs
+
+incomplete concrete QuestionsI of Questions = open Lang in {
+  lincat
+    Phrase = Utt ;
+    Entity = N ;
+    Action = V2 ;
+
+  lin 
+    Who  love_V2 man_N           = (
+UttQS (UseQCl (TTAnt TPres ASimul) PPos (QuestVP whoSg_IP (ComplSlash (SlashV2a love_V2) (DetCN (DetQuant IndefArt NumPl) (UseN man_N)))))
+)
+ ;
+    Whom man_N love_V2           = (
+UttQS (UseQCl (TTAnt TPres ASimul) PPos (QuestSlash whoPl_IP (SlashVP (DetCN (DetQuant DefArt NumSg) (UseN man_N)) (SlashV2a love_V2))))
+  --- UttQS (UseQCl (TTAnt TPres ASimul) PPos (QuestSlash whoSg_IP (SlashVP (DetCN (DetQuant DefArt NumSg) (UseN man_N)) (SlashV2a love_V2))))
+)
+ ;
+    Answer woman_N love_V2 man_N = (
+UttNP (DetCN (DetQuant DefArt NumSg) (ApposCN (ApposCN (UseN woman_N) (DetCN (DetQuant IndefArt NumPl) (UseN love_N))) (DetCN (DetQuant IndefArt NumPl) (UseN man_N))))
+  --- UttNP (DetCN (DetQuant DefArt NumSg) (ApposCN (UseN woman_N) (DetCN (DetQuant IndefArt NumPl) (ApposCN (UseN love_N) (DetCN (DetQuant IndefArt NumPl) (UseN man_N))))))
+  --- UttS (UseCl (TTAnt TPres ASimul) PPos (PredVP (DetCN (DetQuant DefArt NumSg) (UseN woman_N)) (ComplSlash (SlashV2a love_V2) (DetCN (DetQuant IndefArt NumPl) (UseN man_N)))))
+)
+ ;
+
+}

examples/animals/QuestionsI.gfe

@@ -0,0 +1,15 @@
+-- to compile: echo "eb -file=QuestionsI.gfe" | gf $GF_LIB_PATH/present/LangEng.gfo
+-- or use directly gf <mkAnimals.gfs
+
+incomplete concrete QuestionsI of Questions = open Lang in {
+  lincat
+    Phrase = Utt ;
+    Entity = N ;
+    Action = V2 ;
+
+  lin 
+    Who  love_V2 man_N           = %ex Utt "who loves men" ;
+    Whom man_N love_V2           = %ex Utt "whom does the man love" ;
+    Answer woman_N love_V2 man_N = %ex Utt "the woman loves men" ;
+
+}

examples/animals/QuestionsSwe.gf

@@ -0,0 +1,2 @@
+concrete QuestionsSwe of Questions = QuestionsI with 
+  (Lang = LangSwe) ;

src/compiler/GF/Command/Commands.hs

@@ -23,6 +23,7 @@ import PGF.Printer
 import PGF.Probabilistic -- (getProbsFromFile,prProbabilities,defaultProbabilities)
 import PGF.Generate (genRandomProb) ----
 import GF.Compile.Export
+import GF.Compile.ExampleBased
 import GF.Infra.Option (noOptions, readOutputFormat, Encoding(..))
 import GF.Infra.UseIO
 import GF.Data.ErrM ----
@@ -231,6 +232,33 @@ allCommands cod env@(pgf, mos) = Map.fromList [
      longname = "empty",
      synopsis = "empty the environment"
      }),
+  ("eb", emptyCommandInfo {
+     longname = "example_based",
+     syntax = "eb (-probs=FILE | -lang=LANG)* -file=FILE.gfe",
+     synopsis = "converts .gfe files to .gf files by parsing examples to trees",
+     explanation = unlines [
+       "Reads FILE.gfe and writes FILE.gf. Each expression of form",
+       "'%ex CAT QUOTEDSTRING' in FILE.gfe is replaced by a syntax tree.",
+       "This tree is the first one returned by the parser; a biased ranking",
+       "can be used to regulate the order. If there are more than one parses",
+       "the rest are shown in comments, with probabilities if the order is biased.",  
+       "The probabilities flag and configuration file is similar to the commands",
+       "gr and rt. Notice that the command doesn't change the environment,",
+       "but the resulting .gf file must be imported separately."
+       ],
+     flags = [
+       ("file","the file to be converted (suffix .gfe must be given)"),
+       ("lang","the language in which to parse"),
+       ("probs","file with probabilities to rank the parses")
+       ],
+     exec = \opts _ -> do
+       let file = optFile opts
+       mprobs <- optProbs opts pgf
+       let conf = configureExBased pgf mprobs (optLang opts)
+       file' <- parseExamplesInGrammar conf file
+       return (fromString ("wrote " ++ file')),
+     needsTypeCheck = False
+     }),
   ("gr", emptyCommandInfo {
      longname = "generate_random",
      synopsis = "generate random trees in the current abstract syntax",
@@ -871,6 +899,8 @@ allCommands cod env@(pgf, mos) = Map.fromList [
 --       putStrLn $ prProbabilities ps 
        return $ Just ps
 
+   optFile opts = valStrOpts "file" "_gftmp" opts
+
    optType opts = 
      let str = valStrOpts "cat" (showCId $ lookStartCat pgf) opts
      in case readType str of
@@ -988,3 +1018,4 @@ prMorphoAnalysis (w,lps) =
 
 morphoMissing :: Morpho -> [String] -> [String]
 morphoMissing mo ws = [w | w <- ws, null (lookupMorpho mo w)]
+

src/compiler/GF/Compile/ExampleBased.hs

@@ -0,0 +1,59 @@
+module GF.Compile.ExampleBased (parseExamplesInGrammar,configureExBased) where
+
+import PGF
+import PGF.Probabilistic
+
+parseExamplesInGrammar :: ExConfiguration -> FilePath -> IO FilePath
+parseExamplesInGrammar conf file = do
+  src <- readFile file                             -- .gfe
+  let file' = take (length file - 3) file ++ "gf"  -- .gf
+  convertFile conf src file'
+  return file'
+
+convertFile :: ExConfiguration -> String -> FilePath -> IO ()
+convertFile conf src file = do
+  writeFile file "" -- "-- created by example-based grammar writing in GF\n"
+  conv src
+ where
+  conv s = do
+    (cex,end) <- findExample s
+    if null end then return () else do
+      convEx cex
+      conv end
+  findExample s = case s of
+    '%':'e':'x':cs -> return $ getExample cs
+    c:cs -> appf [c] >> findExample cs
+    _ -> return (undefined,s)
+  getExample s = 
+    let 
+      (cat,exend) = break (=='"') s
+      (ex,   end) = break (=='"') (tail exend)
+    in ((unwords (words cat),ex), tail end)  -- quotes ignored
+  pgf = resource_pgf conf
+  lang = language conf 
+  convEx (cat,ex) = do
+    appn "("
+    let typ = maybe (error "no valid cat") id $ readType cat
+    let ts = rank $ parse pgf lang typ ex
+    case ts of
+      []   -> appv ("WARNING: cannot parse example " ++ ex)
+      t:tt -> appn t >> mapM_ (appn . ("  --- " ++)) tt 
+    appn ")"
+  rank ts = case probs conf of
+    Just probs -> [showExpr [] t ++ "  -- " ++ show p | (t,p) <- rankTreesByProbs probs ts]
+    _ -> map (showExpr []) ts
+  appf = appendFile file
+  appn s = appf s >> appf "\n"
+  appv s = appn s >> putStrLn s
+
+data ExConfiguration = ExConf {
+  resource_file :: FilePath,
+  resource_pgf  :: PGF,
+  probs    :: Maybe Probabilities,
+  verbose  :: Bool,
+  language :: Language
+  }
+
+configureExBased :: PGF -> Maybe Probabilities -> Language -> ExConfiguration
+configureExBased pgf mprobs lang = ExConf [] pgf mprobs False lang
+