Merge branch 'master' of github.com:GrammaticalFramework/gf-rgl

.travis.yml

@@ -4,7 +4,7 @@ language: c
 os:
   - linux
   - osx
-  - windows
+#  - windows
 
 addons:
   apt:

src/abstract/Extend.gf

@@ -193,6 +193,9 @@ abstract Extend = Cat ** {
     Cons_nr_RNP : NP  -> RNPList -> RNPList ;   -- John, my family, myself
 ----    Cons_rn_RNP : RNP -> ListNP  -> RNPList ;   -- myself, John, Mary
 
+-- reflexive possessive on its own right, like in Swedish, Czech, Slovak
+
+    ReflPossPron : Quant ;  -- Swe sin,sitt,sina
 
 --- from Extensions
 

src/common/ExtendFunctor.gf

@@ -79,6 +79,7 @@ lin
   Base_rn_RNP = variants {} ;     -- RNP -> NP -> RNPList ; -- myself, John
   Cons_rr_RNP = variants {} ;     -- RNP -> RNPList -> RNPList ; -- my family, myself, John
   Cons_nr_RNP = variants {} ;     -- NP -> RNPList -> RNPList ; -- John, my family, myself
+  ReflPossPron = PossPron he_Pron ;  -- : Quant ;  -- Swe sin,sitt,sina
   ComplGenVV = variants {} ;     -- VV -> Ant -> Pol -> VP -> VP ; -- want not to have slept
   ComplSlashPartLast = ComplSlash ;
   --SlashV2V = variants {} ;     -- V2V -> Ant -> Pol -> VPS -> VPSlash ; -- force (her) not to have slept

src/czech/AllCze.gf

@@ -1,6 +1,7 @@
 --# -path=.:../abstract:../common:../api
 
 concrete AllCze of AllCzeAbs =
-  LangCze
+  LangCze,
+  ExtendCze
   ;
 

src/czech/AllCzeAbs.gf

@@ -1,6 +1,7 @@
 --# -path=.:../abstract:../common:prelude
 
 abstract AllCzeAbs =
-  Lang
+  Lang,
+  Extend
   ;
 

src/czech/CatCze.gf

@@ -24,6 +24,7 @@ concrete CatCze of Cat =
     VPSlash = {verb : VerbForms ; clit,compl : Agr => Str ; c : ComplementCase} ; ----
     V  = ResCze.VerbForms ;
     V2 = ResCze.VerbForms ** {c : ComplementCase} ;
+    VS,VQ = ResCze.VerbForms ;
 
     A  = ResCze.AdjForms ;
     AP = ResCze.Adjective ** {isPost : Bool} ; -- {s : Gender => Number => Case => Str}

src/czech/ExtendCze.gf

@@ -0,0 +1,38 @@
+concrete ExtendCze of Extend = CatCze ** 
+  ExtendFunctor - [
+    ReflPossPron
+    ---- constant not found (yet)
+    ,youPolFem_Pron
+    ,UttVPShort
+    ,UttAccIP
+    ,UttDatIP
+    ,SubjRelNP
+    ,StrandRelSlash
+    ,StrandQuestSlash
+    ,SlashBareV2S
+    ,PredIAdvVP
+    ,PredAPVP
+    ,ExistsNP
+    ,ExistS
+    ,ExistPluralCN
+    ,ExistNPQS
+    ,ExistMassCN
+    ,ExistIPQS
+    ,ExistCN
+    ,EmptyRelSlash
+    ,DetNPMasc
+    ,DetNPFem
+    ,ComplBareVS
+    ,CompIQuant
+    ,CompBareCN
+    ]
+  with (Grammar = GrammarCze)
+    **
+open
+  ResCze
+in {
+
+lin ReflPossPron = justDemPronFormsAdjective reflPossessivePron ;
+
+
+}

src/czech/ResCze.gf

@@ -632,8 +632,6 @@ adjFormsAdjective : AdjForms -> Adjective = \afs -> {
     a : Agr
     } ;
 
----- TODO: possessives
-
   personalPron : Agr -> PronForms = \a ->
     {a = a ; cnom = []} **
     case a of {
@@ -753,6 +751,7 @@ adjFormsAdjective : AdjForms -> Adjective = \afs -> {
 
     } ;
 
+  reflPossessivePron : DemPronForms = mladyAdjForms "svy" ** {msnom = "svůj" ; pdat = "svým"} ;
 
   mkPron : Agr -> PronForms ** {poss : DemPronForms} = \a ->
     personalPron a ** {poss = possessivePron a} ;

src/czech/StructuralCze.gf

@@ -23,6 +23,7 @@ lin
     youSg_Pron = mkPron (Ag (Masc Anim) Sg P2) ;
     he_Pron = mkPron (Ag (Masc Anim) Sg P3) ;
     she_Pron = mkPron (Ag Fem Sg P3) ;
+    it_Pron = mkPron (Ag Neutr Sg P3) ;
     we_Pron = mkPron (Ag (Masc Anim) Pl P1) ;
     youPl_Pron = mkPron (Ag (Masc Anim) Pl P2) ;
     they_Pron = mkPron (Ag (Masc Anim) Pl P3) ;

src/hungarian/NounHun.gf

@@ -99,14 +99,13 @@ concrete NounHun of Noun = CatHun ** open
     quant ** num ** {
       s = \\c => case <isNum num,isIndefArt quant> of {
                    <True,True> => [] ; -- don't output "a 2 cars"
-                   _           => quant.s ! n ! c }
+                   _           => quant.s ! n ! c } -- az/azok a …
               ++ num.s ! Attrib ;      -- TODO: add inflection table in numbers
       sp = \\c => case <isNum num,isIndefArt quant> of {
                    <True,True> => [] ;
-                   _           => quant.sp ! n ! c }
+                   _           => quant.sp ! n ! c } -- az/azok …
               ++ num.s ! Indep ;
       n = n ;
-      dt = qt2dt quant.qt ;
       } ;
 
   -- : Quant -> Num -> Ord -> Det ;  -- these five best
@@ -170,20 +169,20 @@ concrete NounHun of Noun = CatHun ** open
   DefArt = mkQuant "a" "a" ** {
     s,
     sp = \\_,_ => pre {"a" ; "az" / v } ;
-    qt = DefQuant ;
+    dt = DefDet ;
     } ;
 
   -- : Quant
   IndefArt = mkQuant "egy" [] ** {
     s = \\n,_ => case n of {Sg => "egy" ; Pl => []} ;
     sp = \\n,_ => case n of {Sg => "egy" ; Pl => "sok"} ;
-    qt = IndefArticle ;
+    dt = IndefDet ;
     } ;
 
   -- : Pron -> Quant
   PossPron pron = pron ** {
     s,sp = \\_ => pron.s ;
-    qt = QuantPoss (agr2pstem pron.agr) ;
+    dt = DetPoss (agr2pstem pron.agr) ;
     caseagr = False ;
     } ;
 

src/hungarian/ParamHun.gf

@@ -17,17 +17,11 @@ oper
 -- Quant
 
 param
-  QuantType =
-    IndefArticle  -- Needed to prevent "a 2 cars"
-  | IndefQuant    -- Not IndefArt, not poss, not def
-  | DefQuant
-  | QuantPoss PossStem -- Which possessive stem it takes
-  ;
 
   DetType =
-    DefDet  -- distinction between Article and Other no longer needed
+    DefDet
-  | IndefDet -- still need def or indef
+  | IndefDet         -- Def or Indef relevant for verb agreement
-  | DetPoss PossStem -- Sill need to know which stem it takes if Poss
+  | DetPoss PossStem -- Which possessive stem it takes
   ;
 
   -- Singular stems. Plural is always same, no need to add here.
@@ -35,10 +29,10 @@ param
 
 oper
   -- standard trick to prevent "a one car"
-  isIndefArt : {qt : QuantType} -> Bool = \quant ->
+  isIndefArt : {dt : DetType} -> Bool = \quant ->
-    case quant.qt of {
+    case quant.dt of {
-      IndefArticle => True ;
+      IndefDet => True ;
-      _            => False
+      _        => False
     } ;
 
   dt2objdef : DetType -> ObjDef = \dt -> case dt of {
@@ -51,12 +45,6 @@ oper
     Indef => IndefDet
     } ;
 
-  qt2dt : QuantType -> DetType = \qt -> case qt of {
-    QuantPoss x => DetPoss x ;
-    DefQuant => DefDet ;
-    _ => IndefDet
-    } ;
-
   agr2pstem : Person*Number -> PossStem = \pn ->
     case <pn.p1,pn.p2> of {
       <P1,Pl> => dSg_rPl1 ;

src/hungarian/ResHun.gf

@@ -203,13 +203,13 @@ oper
   Quant : Type = BaseQuant ** {
     s, -- form that comes before noun: "{this} car"
     sp : Number => Case => Str ; -- independent form, "I like {this}" (DetNP)
-    qt : QuantType ;
+    dt : DetType ;
     } ;
 
   mkQuant : (s,sp : Str) -> Quant = \s,sp -> {
     s = mkCaseNoun s ;
     sp = mkCaseNoun sp ;
-    qt = DefQuant ;
+    dt = DefDet ;
     caseagr = True ;
     poss = harm1 [] ;
     } ;

src/italian/ParadigmsIta.gf

@@ -160,7 +160,9 @@ oper
 -- case (masc and fem singular, masc plural, adverbial), given that
 -- comparison is formed by "più".
 
-    mkA : (solo,sola,soli,sole,solamente : Str) -> A ; -- irregular adjective
+    mkA : (solo,sola,soli,sole : Str) -> A ; -- possibly irregular adjective
+
+    mkA : (solo,sola,soli,sole,solamente : Str) -> A ; -- irregular adjective, with adverb
 
 -- With irregular comparison, there are as it were two adjectives:
 -- the positive ("buono") and the comparative ("migliore"). 
@@ -391,13 +393,16 @@ oper
 
   regV x = 
     let 
-      are = Predef.dp 3 x ;
+----      are = Predef.dp 3 x ;
       ci  = Predef.dp 2 (Predef.tk 3 x) ;
       i   = last ci ;
-      verb = case are of {
+      verb : {s : VForm => Str} = case x of {
-        "ire" =>  finire_100 x ;
+        _ + "ire" =>  finire_100 x ;
-        "ere" =>  temere_20 x ;
+        _ + "ere" =>  temere_20 x ;
-        "are" => case i of {
+        _ + "durre" =>  condurre_36 x ;
+        _ + "trarre" =>  trarre_91 x ;
+	_ + "rre" =>  porre_68 x ; --- all other verbs in -rre?
+        _ + "are" => case i of {
           "c" => cercare_7 x ;
           "g" => legare_8 x ;
           _ => case ci of {
@@ -512,6 +517,7 @@ oper
 
   mkA = overload {
     mkA : (bianco : Str) -> A = regA ;
+    mkA : (solo,sola,soli,sole : Str) -> A = \solo,sola,soli,sole -> mk5A solo sola soli sole (sola + "mente") ;
     mkA : (solo,sola,soli,sole, solamente : Str) -> A = mk5A ;
     mkA : A -> A -> A = mkADeg ;
     mkA : A -> CopulaType -> A = adjCopula ;

src/morphodict/Makefile

@@ -0,0 +1,5 @@
+ita:
+	cp -p MorphoDictItaAbs.header MorphoDictItaAbs.gf
+	cp -p MorphoDictIta.header MorphoDictIta.gf
+	python3 get_dict.py Ita tmp/IrregItaAbs.gf tmp/entries | grep -v "lin " >>MorphoDictItaAbs.gf
+	python3 get_dict.py Ita tmp/IrregItaAbs.gf tmp/entries | grep -v "fun " >>MorphoDictIta.gf

src/morphodict/MorphoDictIta.header

@@ -0,0 +1,15 @@
+concrete MorphoDictIta of MorphoDictItaAbs =
+  CatIta [N,A,V,Adv,Prep] **
+  open
+    ParadigmsIta, IrregIta
+  in
+ {
+
+-- converted from https://github.com/TALP-UPC/FreeLing/tree/master/data/it/dictionary/entries
+-- inheriting its license https://creativecommons.org/licenses/by-sa/3.0/
+
+lincat NSg,NPl = N ;
+oper mkNSg : Str -> Gender -> N = \s,g -> mkN s g ;
+oper mkNPl : Str -> Gender -> N = \s,g -> mkN s s g ;
+
+

src/morphodict/MorphoDictItaAbs.header

@@ -0,0 +1,9 @@
+abstract MorphoDictItaAbs =
+  Cat [N,A,V,Adv,Prep] **
+{
+
+-- converted from https://github.com/TALP-UPC/FreeLing/tree/master/data/it/dictionary/entries
+-- inheriting its license https://creativecommons.org/licenses/by-sa/3.0/
+
+cat NSg ; NPl ;
+

src/morphodict/get_dict.py

@@ -0,0 +1,176 @@
+import sys
+
+"""
+Converting a form-lemma-description lexicon to a GF lexicon.
+
+Example: Romance resources from http://nlp.lsi.upc.edu/freeling/index.php/node/12
+
+  $ python3 get_dict.py Ita tmp/IrregItaAbs.gf tmp/entries
+
+where entries is obtained by 
+
+  $ cat adjs adv noun verb other | sort -u >entries
+
+in FreeLing/data/it/dictionary/entries/
+In this file, each line has three words, e.g.
+
+  abbondanze abbondanza NCFP000
+
+from which the script produces two lines,
+
+  fun abbondanza_N : N ;
+  lin abbondanza_N = mkN "abbondanza" "abbondanze" feminine ;
+
+These can be directed to an abstract and concrete module by using grep, as shown in Makefile.
+
+Adaptation to a new language requires extending lang_args() with a new language code and the functions for that language.
+
+"""
+# an auxiliary
+
+def quoted(s):
+    return '"' + s + '"'
+
+######################
+## language-specific code
+######################
+
+## Ita = Italian
+
+# how each argument is obtained from the set of descriptions
+def ita_noun_args(irregs,entry):
+    arg1 = entry.get('NCFS000', entry.get('NCMS000',entry.get('NCFN000',entry.get('NCMN000','NONE'))))
+    arg2 = entry.get('NCFP000', entry.get('NCMP000','NONE'))
+    if 'NCFS000' in entry.keys():
+        arg3 = 'feminine'
+    else:
+        arg3 = 'masculine'
+    if arg2 == 'NONE':
+        return ['NSg', arg1, 'mkNSg', quoted(arg1), arg3]  # nouns that occur in singular only
+    elif arg1 == 'NONE':
+        return ['NPl', arg2, 'mkNPl', quoted(arg2), arg3]  # nouns that occur in pular only
+    else:
+        return ['N', arg1, 'mkN', quoted(arg1), quoted(arg2),arg3]
+
+    
+def ita_adj_args(irregs,entry):
+    arg1 = entry.get('AQ0MS00', entry.get('AQ0CS00',entry.get('AQ0CN00','NONE')))
+    arg2 = entry.get('AQ0FS00', entry.get('AQ0CS00',entry.get('AQ0CN00','NONE')))
+    arg3 = entry.get('AQ0MP00', entry.get('AQ0CP00',entry.get('AQ0CN00','NONE')))
+    arg4 = entry.get('AQ0FP00', entry.get('AQ0CS00',entry.get('AQ0CN00','NONE_' + str(entry))))
+    return ['A', arg1, 'mkA', quoted(arg1), quoted(arg2),quoted(arg3),quoted(arg4)]
+
+
+def ita_verb_args(irregs,entry):
+    arg1 = entry.get('VMN0000', 'NONE_' + str(entry))
+    if (arg1 + '_V') in irregs:
+        return ['V', arg1, '', 'IrregIta.' + arg1 + '_V']
+    else:
+        return ['V', arg1, 'mkV', quoted(arg1)]
+
+    
+def ita_adv_args(irregs,entry):
+    arg1 = entry.get('RG', 'NONE_' + str(entry))
+    return ['Adv', arg1, 'mkAdv', quoted(arg1)]
+
+
+def ita_prep_args(irregs,entry):
+    arg1 = entry.get('SPS00', 'NONE_' + str(entry)) ## includes di,in,... but not their contractions
+    return ['Prep', arg1, 'mkPrep', quoted(arg1)]
+
+def ignore_args(irregs,entry):
+    arg1 = 'NONE_' + str(entry)
+    return ['IGNORED', list(entry.values())[0], '', quoted(arg1)]
+
+    
+def ita_args(key):
+    if key[0] == 'A':
+        return ita_adj_args
+    elif key[0] == 'N':
+        return ita_noun_args
+    elif key[0] == 'V':
+        return ita_verb_args
+    elif key[0] == 'R':
+        return ita_adv_args
+    elif key[0] == 'S':
+        return ita_prep_args
+    else:
+        return ignore_args
+
+def lang_args(lang,key):
+    if lang == 'Ita':
+        return ita_args(key)
+    else:
+        print("unknown language", lang)
+
+##########################################################
+## from this point, the code is generic for all languages
+##########################################################
+
+        
+def get_irregs(ifile):
+    file = open(ifile)
+    irregs = [] 
+    for line in file:
+        ws = line.split()
+        if ws and ws[0] == 'fun':
+            irregs.append(ws[1])
+    file.close()
+    return irregs
+
+
+def get_dict(filename):
+    file = open(filename)
+    dict = {}
+    for line in file:
+        words = line.split()
+        if len(words) == 3:
+            cat = words[2][0] # N,A,V,R,S ...
+            key = cat + '_' + words[1]
+            dict[key] = dict.get(key,{})
+            dict[key][words[2]] = words[0]
+    return dict
+
+def clean_fun(lemma,cat):
+    if lemma.isalpha():
+        return lemma + '_' + cat
+    else:
+        fun = ''
+        for c in lemma:
+            if c=="'":
+                fun += "\\'"
+            else:
+                fun += c
+        return "'" + fun + '_' + cat + "'"
+
+def print_entry(irregs,mapping,entry):
+    args = mapping(irregs,entry)
+    fun = clean_fun(args[1],args[0])
+    comment = min(1,str(args).count('NONE')) * '-- '
+    rule = [comment + "fun", fun, ":", args[0], ";"]
+    print(' '.join(rule))
+    rule = [comment + "lin", fun, "=", args[2]]
+    for arg in args[3:]:
+        rule.append(arg)
+    rule.append(';')
+    print(' '.join(rule))
+
+    
+def main():
+    if len(sys.argv) != 4:
+        print("usage: python3 get_dict <lang> <irregsfile> <dictfile>")
+        return 1
+    lang = sys.argv[1]
+    irregsfile = sys.argv[2]
+    dictfile = sys.argv[3]
+    irregs = get_irregs(irregsfile)
+    dict = get_dict(dictfile)
+    for key,entry in dict.items():
+        print_entry(irregs,lang_args(lang,key),entry)
+    print("}")
+    return 0
+
+        
+main()
+
+

src/slovak/AllSlo.gf

@@ -1,6 +1,7 @@
 --# -path=.:../abstract:../common:../api
 
 concrete AllSlo of AllSloAbs =
-  LangSlo
+  LangSlo,
+  ExtendSlo
   ;
 

src/slovak/AllSloAbs.gf

@@ -1,6 +1,7 @@
 --# -path=.:../abstract:../common:prelude
 
 abstract AllSloAbs =
-  Lang
+  Lang,
+  Extend
   ;
 

src/slovak/CatSlo.gf

@@ -24,6 +24,7 @@ concrete CatSlo of Cat =
     VPSlash = {verb : VerbForms ; clit,compl : Agr => Str ; c : ComplementCase} ; ----
     V  = ResSlo.VerbForms ;
     V2 = ResSlo.VerbForms ** {c : ComplementCase} ;
+    VS,VQ  = ResSlo.VerbForms ;
 
     A  = ResSlo.AdjForms ;
     AP = ResSlo.Adjective ** {isPost : Bool} ; -- {s : Gender => Number => Case => Str}

src/slovak/ExtendSlo.gf

@@ -0,0 +1,38 @@
+concrete ExtendSlo of Extend = CatSlo ** 
+  ExtendFunctor - [
+    ReflPossPron
+    ---- constant not found (yet)
+    ,youPolFem_Pron
+    ,UttVPShort
+    ,UttAccIP
+    ,UttDatIP
+    ,SubjRelNP
+    ,StrandRelSlash
+    ,StrandQuestSlash
+    ,SlashBareV2S
+    ,PredIAdvVP
+    ,PredAPVP
+    ,ExistsNP
+    ,ExistS
+    ,ExistPluralCN
+    ,ExistNPQS
+    ,ExistMassCN
+    ,ExistIPQS
+    ,ExistCN
+    ,EmptyRelSlash
+    ,DetNPMasc
+    ,DetNPFem
+    ,ComplBareVS
+    ,CompIQuant
+    ,CompBareCN
+    ]
+  with (Grammar = GrammarSlo)
+    **
+open
+  ResSlo
+in {
+
+lin ReflPossPron = justDemPronFormsAdjective reflPossessivePron ;
+
+
+}

src/slovak/ResSlo.gf

@@ -869,6 +869,15 @@ adjFormsAdjective : AdjForms -> Adjective = \afs -> {
 
     } ;
 
+  reflPossessivePron : DemPronForms = otcovA "svoj" ** {
+        msnom = "svoj" ; msgen = "svojho" ; msdat = "svojmu" ;
+	msins = "svojím" ;
+	ampnom = "svoji" ;
+	nsnom, fpnom = "svoje" ;
+	pgen = "svojich" ;
+	pdat = "svojim" ;
+	pins = "svojimi" ;
+	} ;
 
   mkPron : Agr -> PronForms ** {poss : DemPronForms} = \a ->
     personalPron a ** {poss = possessivePron a} ;

src/slovak/StructuralSlo.gf

@@ -24,6 +24,7 @@ lin
     youSg_Pron = mkPron (Ag (Masc Anim) Sg P2) ;
     he_Pron = mkPron (Ag (Masc Anim) Sg P3) ;
     she_Pron = mkPron (Ag Fem Sg P3) ;
+    it_Pron = mkPron (Ag Neutr Sg P3) ;
     we_Pron = mkPron (Ag (Masc Anim) Pl P1) ;
     youPl_Pron = mkPron (Ag (Masc Anim) Pl P2) ;
     they_Pron = mkPron (Ag (Masc Anim) Pl P3) ;

src/swedish/ExtendSwe.gf

@@ -12,14 +12,16 @@ concrete ExtendSwe of Extend = CatSwe **
     ICompAP,ProDrop,EmbedSSlash,
     AdAdV, PositAdVAdj, GerundCN, GerundNP, GerundAdv, PresPartAP, PastPartAP, PastPartAgentAP,
     RNP, RNPList, ReflRNP, ReflPron, ReflPoss, PredetRNP, ConjRNP,
-    Base_rr_RNP, Base_nr_RNP, Base_rn_RNP, Cons_rr_RNP, Cons_nr_RNP,
+    Base_rr_RNP, Base_nr_RNP, Base_rn_RNP, Cons_rr_RNP, Cons_nr_RNP, ReflPossPron,
     CompoundN, CompoundAP, AdvIsNP,
     UttAccNP
   ]
   with (Grammar = GrammarSwe)
     **
  open CommonScand, ResSwe, ParamX, VerbSwe, Prelude, DiffSwe, StructuralSwe, MorphoSwe,
-      NounSwe, Coordination, AdjectiveSwe, SentenceSwe, AdverbSwe, RelativeSwe, (P = ParadigmsSwe) in {
+      NounSwe, Coordination, AdjectiveSwe, SentenceSwe, AdverbSwe, RelativeSwe, (P = ParadigmsSwe),
+      (M = MakeStructuralSwe)
+in {
 
   flags coding=utf8 ;
 
@@ -207,6 +209,8 @@ concrete ExtendSwe of Extend = CatSwe **
     Cons_rr_RNP x xs = consrTable Agr comma x xs ;
     Cons_nr_RNP x xs = consrTable Agr comma {s = \\a => x.s ! NPAcc} xs ;
 
+    ReflPossPron = M.mkQuant "sin" "sitt" "sina" ;
+    
   lin
     ApposNP np1 np2 = {s = \\nform => np1.s ! nform ++ comma ++ np2.s ! nform; a = np1.a; isPron = False} ;