Merge pull request #157 from inariksit/persian

Persian

src/persian/CatPes.gf

@@ -59,9 +59,9 @@ concrete CatPes of Cat = CommonX - [Adv] ** open ResPes, Prelude in {
     Pron = {s : Str ; ps : Str ; a : AgrPes};
     Det = ResPes.Determiner ;
     Predet = {s : Str} ;
-    Num  = {s : Str ; n : Number} ;
+    Num  = {s : Str ; n : Number ; isNum : Bool} ;
     Card = {s : Str; n : Number} ;
-    Ord = {s : Str; n : Number} ;
+    Ord = {s : Str; n : Number ; isNum : Bool} ;
     Quant = {s: Number => Str ; a:AgrPes ; fromPron : Bool};
     Art = {s : Str} ;
 
@@ -87,10 +87,10 @@ concrete CatPes of Cat = CommonX - [Adv] ** open ResPes, Prelude in {
     A = ResPes.Adjective ; --- {s : Gender => Number => Case => Str} ;
     A2 = ResPes.Adjective ** { c2 : Str} ;
 
-    N = {s : Ezafa => Number => Str ; animacy : Animacy ; definitness : Bool} ;
+    N = ResPes.Noun ;
 
-    N2 = {s : Ezafa => Number  => Str ; animacy : Animacy ; definitness : Bool} ** {c : Str};
-    N3 = {s : Ezafa => Number  => Str ; animacy : Animacy ; definitness : Bool} ** {c2 : Str ; c3 : Str } ;
+    N2 = ResPes.Noun ** {c : Str};
+    N3 = ResPes.Noun ** {c2 : Str ; c3 : Str} ;
     PN = {s : Str ; animacy : Animacy} ;
 
 }

src/persian/ExtendPes.gf

@@ -0,0 +1,17 @@
+--# -path=.:../common:../abstract:../prelude
+
+concrete ExtendPes of Extend =
+  CatPes ** ExtendFunctor - [
+    GenNP
+    ]
+  with (Grammar=GrammarPes)
+  ** open Prelude, ResPes in {
+
+lin
+  -- NP -> Quant ; -- this man's
+  GenNP np = np ** { 
+    s = \\num => np.s ! NPC bEzafa ; -- the possessed will get aEzafa, possesser is just unmarked; https://sites.la.utexas.edu/persian_online_resources/language-specific-grammar/ezfe/
+    fromPron = True -- not necessarily from Pron, but it should come after the noun, and if fromPron=True, then DetCN places determiner after cn.
+  } ;
+
+}

src/persian/LangPes.gf

@@ -1,8 +1,8 @@
---# -path=.:../abstract:../common:../hindustani
+--# -path=.:../abstract:../common
 
 concrete LangPes of Lang = 
-  GrammarPes,
-  LexiconPes
+  GrammarPes
+  ,LexiconPes
   ** {
 
 flags startcat = Phr ; unlexer=unwords ; lexer=words ;

src/persian/NounPes.gf

@@ -3,11 +3,11 @@ concrete NounPes of Noun = CatPes ** open ResPes, Prelude in {
   flags optimize=all_subs ;
 
   lin
-    DetCN det cn = {
-      s =  \\_ => case <det.isNum,det.fromPron> of {
-                 <False,True> => cn.s ! aEzafa ! det.n ++ det.s ; -- det.n  ;
-		 <False,False> => det.s  ++ cn.s ! bEzafa ! det.n ; -- det.n  ;
-		 <True,True> => cn.s ! aEzafa ! Sg ++ det.s ;
+    DetCN det cn = { s = \\_ =>
+      case <det.isNum,det.fromPron> of {
+        <False,True>  => cn.s ! aEzafa ! det.n ++ det.s ;
+        <False,False> => det.s ++ cn.s ! bEzafa ! det.n ;
+        <True,True>   => cn.s ! aEzafa ! Sg ++ det.s ; -- noun modified by a number is invariably singular
         <True,False>  => det.s ++ cn.s ! bEzafa ! Sg
      };
       a = agrPesP3 det.n ;
@@ -17,40 +17,32 @@ concrete NounPes of Noun = CatPes ** open ResPes, Prelude in {
     UsePN pn = {s = \\_ => pn.s ; a = agrPesP3 Sg ; animacy = pn.animacy } ;
     UsePron p = {s = \\_ => p.s ; a = p.a ; animacy = Animate} ;
 
-    PredetNP pred np = {
-      s = \\ez => pred.s ++ np.s ! ez ;
-      a = np.a;
-      animacy = np.animacy
+    PredetNP pred np = np ** {
+      s = \\ez => pred.s ++ np.s ! ez
       } ;
 
-    PPartNP np v2 = {
-      s =  \\ez => np.s ! ez ++ partNP (v2.s ! Root1)    ;
-      a = np.a ;
-      animacy = np.animacy
+    PPartNP np v2 = np ** {
+      s = \\ez => np.s ! ez ++ partNP (v2.s ! Root1)
       } ;
 
-    RelNP np rs = {
-	  s = \\ez => np.s ! ez  ++ rs.s ! np.a ;
-      a = np.a ;
-      animacy = np.animacy
+    RelNP np rs = np ** {
+      s = \\ez => np.s ! ez ++ rs.s ! np.a
       } ;
 
-    AdvNP np adv = {
-      s = \\ez => np.s ! NPC aEzafa ++ adv.s ;
-      a = np.a  ;
-      animacy = np.animacy
+    AdvNP np adv = np ** {
+      s = \\ez => np.s ! NPC aEzafa ++ adv.s
       } ;
 
     DetQuantOrd quant num ord = {
       s = quant.s ! num.n ++ num.s ++ ord.s ;
-	isNum = True;
+      isNum = orB num.isNum ord.isNum ;
       fromPron = quant.fromPron ;
       n = num.n
       } ;
 
     DetQuant quant num = {
       s = quant.s ! num.n ++ num.s;
-	isNum = True ; -- this does not work in case of 'these women' but  works in case of 'five women' 
+      isNum = num.isNum;
       fromPron = quant.fromPron ;
       n = num.n
       } ;
@@ -63,20 +55,20 @@ concrete NounPes of Noun = CatPes ** open ResPes, Prelude in {
 
     PossPron p = {s = \\_ => p.ps ; a = p.a ; fromPron = True} ;
 
-    NumSg = {s = [] ; n = Sg} ;
-    NumPl = {s = [] ; n = Pl} ;
+    NumSg = {s = [] ; n = Sg ; isNum = False} ;
+    NumPl = {s = [] ; n = Pl ; isNum = False} ;
 -- from here
-    NumCard n = n ** {hasCard = True} ;
+    NumCard n = n ** {isNum = True} ;
 
-    NumDigits n = {s = n.s ! NCard ; n = n.n} ;
-    OrdDigits n = {s = n.s ! NOrd; n = n.n} ;
+    NumDigits n = n ** {s = n.s ! NCard; isNum = True} ;
+    OrdDigits n = n ** {s = n.s ! NOrd ; isNum = True} ;
 
-    NumNumeral numeral = {s = numeral.s ! NCard; n = numeral.n} ;
-    OrdNumeral numeral = {s = numeral.s ! NOrd ; n = numeral.n} ;
+    NumNumeral n = n ** {s = n.s ! NCard; isNum = True} ;
+    OrdNumeral n = n ** {s = n.s ! NOrd ; isNum = True} ;
 -- to here
-    AdNum adn num = {s = adn.s ++ num.s ; n = num.n} ;
+    AdNum adn num = num ** {s = adn.s ++ num.s} ;
 
-    OrdSuperl a = {s = a.s ! bEzafa ++ taryn; n = Sg} ; -- check the form of adjective
+    OrdSuperl a = {s = a.s ! bEzafa ++ taryn; n = Sg ; isNum=False} ; -- check the form of adjective
 
     DefArt = {s = \\_ => [] ; a = defaultAgrPes ; fromPron = False} ;
     IndefArt = {s = table { Sg => IndefArticle ; Pl => []} ; a =defaultAgrPes ; fromPron = False} ;
@@ -119,22 +111,18 @@ concrete NounPes of Noun = CatPes ** open ResPes, Prelude in {
       definitness = True;
       } ;
 
-    AdjCN ap cn = {
-      s = \\ez,n =>  cn.s ! aEzafa ! n ++ ap.s ! ez; -- check the form of adjective and also cn.s!ez!n changed from cn.s!aEzafa!n to have correct enclicitic form other wise it creats wrong enclictic form of old man
-      animacy = cn.animacy ;
-      definitness = cn.definitness
+    AdjCN ap cn = cn ** {
+      s = \\ez,n =>  cn.s ! aEzafa ! n ++ ap.s ! ez -- check the form of adjective and also cn.s!ez!n changed from cn.s!aEzafa!n to have correct enclicitic form other wise it creats wrong enclictic form of old man
       } ;
 
-    RelCN cn rs = {
+    RelCN cn rs = cn ** {
       s = \\ez,n => cn.s ! enClic ! n ++ rs.s ! agrPesP3 n ;
-      animacy = cn.animacy ;
-      definitness = cn.definitness
       } ;
 
-    AdvCN cn ad = {s = \\ez,n => cn.s ! aEzafa ! n ++ ad.s ; animacy = cn.animacy ; definitness = cn.definitness} ;
+    AdvCN cn ad = cn ** {s = \\ez,n => cn.s ! aEzafa ! n ++ ad.s} ;
 
-    SentCN cn sc = {s = \\ez,n => cn.s ! ez ! n ++ sc.s ; animacy = cn.animacy ; definitness = cn.definitness} ;
+    SentCN cn sc = cn ** {s = \\ez,n => cn.s ! ez ! n ++ sc.s} ;
 
-    ApposCN cn np = {s = \\ez,n => cn.s ! ez ! n  ++ np.s ! NPC aEzafa ; animacy = cn.animacy ; definitness = True} ; -- ezafa form of city to be used
+    ApposCN cn np = cn ** {s = \\ez,n => cn.s ! ez ! n ++ np.s ! NPC aEzafa ; definitness = True} ; -- ezafa form of city to be used
 
 }

src/persian/ParadigmsPes.gf

@@ -1,6 +1,6 @@
---# -path=.:../abstract:../../prelude:../common
+--# -path=.:../abstract:../prelude:../common
 --
-----1 Pnbu Lexical Paradigms
+----1 Persian Lexical Paradigms
 
 resource ParadigmsPes = open
   Predef,
@@ -23,13 +23,25 @@ oper
   singular = Sg ; plural = Pl;
 
   animate = Animate ; inanimate = Inanimate ; --i
-  mkN01 : Str -> Animacy -> Noun ;
-  mkN01 str ani = mkN str (str ++ "ها") ani;
-  mkN02 : Str -> Animacy -> Noun ;
-  mkN02 str ani = case (last str) of {
-    "ه" => mkN str ((init str) + "گان") ani ;
-    ("ا"|"و") => mkN str (str + "یان") ani ;
-    _ => mkN str (str+"ان") ani
+  mkN01 : (sg : Str) -> Animacy -> Noun ; -- Takes singular form and animacy, forms plural with ها
+  mkN01 sg ani =
+    let pl = zwnj sg "ها" ; -- Using zero-width non-joiner, defined in ResPes
+    in MorphoPes.mkN sg pl ani ;
+
+  mkN02 : (sg : Str) -> Animacy -> Noun ; -- Takes singular form and animacy, pattern matches singular and forms plural with either گان, یان or ان
+  mkN02 str ani = case last str of {
+    "ه"       => MorphoPes.mkN str (init str + "گان") ani ;
+    ("ا"|"و") => MorphoPes.mkN str (str + "یان")      ani ;
+    _         => MorphoPes.mkN str (str + "ان")       ani
+  };
+
+  mkN = overload {
+    mkN : (sg : Str) -> N -- Takes singular form, returns an inanimate noun with ها as the plural form
+      = \sg -> mkN01 sg inanimate ;
+    mkN : (sg,pl : Str) -> N -- Takes singular and plural form, returns an inanimate noun
+      = \sg,pl -> MorphoPes.mkN sg pl inanimate ;
+    mkN : (sg,pl : Str) -> Animacy -> N -- Worst-case constructor: takes singular and plural forms and animacy
+      = \sg,pl,ani -> MorphoPes.mkN sg pl ani
   } ;
 {-
 
@@ -45,20 +57,19 @@ oper
 -- Compound Nouns
 
   mkCmpdNoun1 : Str -> N -> N
-     = \s,noun -> {s =\\ez,n => s ++ noun.s ! ez ! n  ;  animacy = noun.animacy ; definitness = noun.definitness ; lock_N = <>};
+    = \s,noun -> noun ** {s =\\ez,n => s ++ noun.s ! ez ! n};
   mkCmpdNoun2 : N -> Str -> N
-     = \noun,s -> {s =\\ez,n =>  noun.s ! ez ! n  ++ s ; animacy = noun.animacy ; definitness =noun.definitness ; lock_N = <>};
+    = \noun,s -> noun ** {s =\\ez,n => noun.s ! ez ! n  ++ s};
 
 
 -- Proper names
-   mkPN : Str -> Animacy -> PN =
-
-     \str,ani -> {s = str ; animacy = ani ; lock_PN = <>} ;
+   mkPN : Str -> Animacy -> PN
+    = \str,ani -> lin PN {s = str ; animacy = ani} ;
 
 
 -- Personal Pronouns
-  personalPN : Str -> Number -> PPerson -> Pron =
-     \str,nn,p ->  {s = str ; a = AgPes nn p ; ps = str ; lock_Pron = <>};
+  personalPN : Str -> Number -> PPerson -> Pron
+    = \str,nn,p -> lin Pron {s = str ; a = AgPes nn p ; ps = str};
 {-
 -- Demonstration Pronouns
   demoPN : Str -> Str -> Str -> Quant =
@@ -66,10 +77,10 @@ oper
 -- Determiner
 -}
   mkDet = overload {
-  mkDet : Str -> Number -> Det =
-    \s1,n -> makeDet s1 n False ** { lock_Det = <>};
-  mkDet : Str -> Number -> Bool -> Det =
-    \s1,n,b -> makeDet s1 n b ** { lock_Det = <>};
+    mkDet : Str -> Number -> Det
+      = \s1,n -> lin Det (makeDet s1 n False);
+    mkDet : Str -> Number -> Bool -> Det
+      = \s1,n,b -> lin Det (makeDet s1 n b)
   };
  {-
 -- Intergative pronouns
@@ -83,11 +94,11 @@ oper
 
   mkA = overload {
     mkA : Str -> A
-      = \str ->   mkAdj str str ** { lock_A = <>} ;
+      = \str -> lin A (mkAdj str str);
     mkA : Str-> Str -> A
-      = \str,adv ->   mkAdj str adv ** { lock_A = <>} ;  
+      = \str,adv -> lin A (mkAdj str adv);
     mkA : Str -> Str -> A2
-	  = \a,c -> mkAdj a a  ** { c2 = c ; lock_A2 = <>} ;
+    = \a,c -> lin A2 (mkAdj a a ** {c2 = c})
   } ;
 
 --2 Verbs
@@ -113,26 +124,28 @@ oper
     } ;
 
   mkV3 : V -> Str -> Str -> V3;
-    mkV3 v p q = v ** { c2 = p ; c3 = q ; lock_V3 = <>} ;
+    mkV3 v p q = lin V3 (v ** {c2 = p ; c3 = q}) ;
   mkV2V : V -> Str -> Str -> Bool -> V2V ;
-    mkV2V v s1 s2 b = v ** {isAux = b ; c1 = s1 ; c2 = s2 ; lock_V2V = <>} ;
+    mkV2V v s1 s2 b = lin V2V (v ** {isAux = b ; c1 = s1 ; c2 = s2}) ;
 
 -- compund verbs
   compoundV = overload {
-   compoundV : Str -> V -> V = \s,v -> {s = \\vf => s ++ v.s ! vf ; lock_V = <>} ;     
-   compoundV : Str -> V2 -> V = \s,v -> {s = \\vf => s ++ v.s ! vf ; lock_V = <>} ;
+    compoundV : Str -> V -> V
+      = \s,v -> v ** {s = \\vf => s ++ v.s ! vf} ;
+    compoundV : Str -> V2 -> V
+      = \s,v -> lin V {s = \\vf => s ++ v.s ! vf} ;
   };
 
   invarV : Str -> V -- for verbs like  " بایستن " ("must"), which don't inflect
     = \s -> lin V {s = \\_ => s} ;
 
 ----2 Adverbs
-  mkAdv : Str -> Adv = \str -> {s =  str ; lock_Adv = <>};
+  mkAdv : Str -> Adv = \str -> lin Adv {s = str} ;
 
 ----2 Prepositions
 
   mkPrep : Str -> Prep ;
-    mkPrep str = {s = str ; lock_Prep = <>};
+    mkPrep str = lin Prep {s = str};
 {-
 --3 Determiners and quantifiers
 
@@ -186,23 +199,5 @@ oper
 --
 --  V0 : Type ;
 --  AS, A2S, AV, A2V : Type ;
---
-----.
-----2 Definitions of paradigms
-----
----- The definitions should not bother the user of the API. So they are
----- hidden from the document.
---
---  Gender = MorphoHin.Gender ; 
---  Number = MorphoHin.Number ;
---  Case = MorphoHin.Case ;
---  human = Masc ; 
---  nonhuman = Neutr ;
---  masculine = Masc ;
---  feminine = Fem ;
---  singular = Sg ;
---  plural = Pl ;
---  nominative = Nom ;
---  genitive = Gen ;
--}  
+--}
 }

src/persian/ResPes.gf

@@ -62,6 +62,10 @@ resource ResPes =  ParamX **  open Prelude,Predef in {
     Neg => CNeg b
     } ;
     
+    -- Zero-width non-joiner, used for certain morphemes
+    -- See https://en.wikipedia.org/wiki/Persian_alphabet#Word_boundaries
+    ZWNJ : Str = "‌" ;
+    zwnj : Str -> Str -> Str = \s1,s2 -> s1 + ZWNJ + s2 ;
  -----------------------
  --- Verb Phrase
  -----------------------

src/persian/SymbolPes.gf

@@ -28,7 +28,7 @@ concrete SymbolPes of Symbol = CatPes ** open Prelude, ResPes in {
 
   SymbS sy = sy ;
   SymbNum sy = { s = sy.s ; n = Pl } ;
-  SymbOrd sy = { s = sy.s ++ "wN" ; n = Pl} ;
+  SymbOrd sy = { s = sy.s ++ "wN" ; n = Pl; isNum=False} ;
 
 lincat
 

src/persian/VerbPes.gf

@@ -41,9 +41,10 @@ concrete VerbPes of Verb = CatPes ** open ResPes in {
     ReflVP v = insertObjPre (\\a =>  reflPron ! a) v ;
     PassV2 v = predV v ; -- need to be fixed
     CompAP ap ={s = \\_ => ap.s ! bEzafa} ; -- check form of adjective
-    CompNP np = {s = \\a => np.s ! NPC bEzafa} ;
     CompAdv adv = {s = \\_ => adv.s } ;
     CompCN cn = {s = \\a => cn.s ! bEzafa !  giveNumber a } ;
-
+    -- IL 2019-01-31 changed bEzafa to enClic according to
+    -- https://sites.la.utexas.edu/persian_online_resources/nouns/noun-in-a-predicative-position/
+    CompNP np = {s = \\a => np.s ! NPC enClic} ;
 
 }