1
0
forked from GitHub/gf-rgl

Complete AdjectiveTur (#12)

* remove unnecessary flags from GrammarTur.gf

* Implement the linearization for UseA2

* Minor formatting fixes

* Manually specify the preposition "ile" for married_A2

* Implement the linearization for ComplA2

* Implement the linearization for ReflA2 ("kendi ile evli")

* Fix ReflA2 adjective case to Nom

* Add a TODO comment

* Add a linearization for AdAP

* Implement the linearization for UseA2

* Minor formatting fixes

* Manually specify the preposition "ile" for married_A2

* Implement the linearization for ComplA2

* Implement the linearization for ReflA2 ("kendi ile evli")

* Fix ReflA2 adjective case to Nom

* Add a TODO comment

* Add a linearization for AdAP

* Fix AdAP

* Add linearization for at_most_AdN

* Add linearization for at_least_AdN

* added dummies for the missing modules in Turkish

* Add the linearization for as_CAdv

* Add rudimentary linearization for UsePN

* Add the linearization for CAdvAP

* Add preliminary linearization for EmbedVP

* Add linearization stub for SentAP

* Fix code-formatting and indentation in HarmonyTur

* Patterns for front and back vowels

* Add an oper for attaching the -mA suffix

* Formatting fixes in ParadigmsTur

* Attempt to implement attachMe

* Almost-working linearization for SentAP

* Update TODO for SentAP

* Move the addition of suffix si to the linearization of EmbedVP

* Import PhraseTur in GrammarTur

* Fix the duplication of PhraseTur in GrammarTur

* Start working on PhraseTur

* Fix formatting of makeH4Table

* Formatting

* Begin implementing the gerund inflection `makeVerb`

* Attach Number and Case data to the Gerund constructor of param VForm

* Get rid of hacky appending of "si" in EmbedVP
This commit is contained in:
Ayberk Tosun
2017-09-30 06:29:03 -05:00
committed by Krasimir Angelov
parent 432a6295cc
commit 76d6749797
15 changed files with 726 additions and 542 deletions

View File

@@ -1,4 +1,5 @@
concrete AdjectiveTur of Adjective = CatTur ** open ResTur, Prelude in {
concrete AdjectiveTur of Adjective =
CatTur ** open ResTur, ParadigmsTur, Prelude in {
lin
@@ -18,4 +19,38 @@ concrete AdjectiveTur of Adjective = CatTur ** open ResTur, Prelude in {
s = \\n, c => adv.s ++ ap.s ! n ! c
} ;
AdAP ada ap = {
s = \\n, c => ada.s ++ ap.s ! n ! c
} ;
UseA2 a = {s = a.s} ;
ComplA2 a np = {
s = \\n, c => np.s ! a.c.c ++ a.c.s ++ a.s ! n ! c
} ;
-- TODO: Whether this is correct or not requires further examination.
ReflA2 a = {
s =
let
kendi : N = mkN "kendi"
in
\\n, c => kendi.s ! n ! c ++ a.c.s ++ a.s ! n ! Nom
} ;
-- Some examples of using CAdvAP:
-- Lang> gt -number=2 -depth=1 (CAdvAP ? ? ?) | l -lang=LangTur
-- > Paris kadar kötü
-- > o kadar kötü
CAdvAP cadv ap np = {
s = \\n, c => np.s ! Nom ++ cadv.s ++ ap.s ! n ! c
} ;
-- TODO: Instead of `++ BIND ++ "si"`, sc.s should be treated as a noun
-- and it should be inflected to `gen Sg {n = Sg; p = P3}`.
SentAP ap sc = {
s =
\\n, c => sc.s ++ (ap.s ! n ! c)
} ;
}

View File

@@ -1,8 +1,4 @@
concrete ConjunctionEng of Conjunction =
concrete ConjunctionTur of Conjunction =
CatTur ** open ResTur, Coordination, Prelude in {
lin
-- ConjNP : Conj -> ListNP -> NP
-- ConjNP conj ss =
}
}

View File

@@ -4,12 +4,19 @@ concrete GrammarTur of Grammar =
NounTur,
VerbTur,
AdjectiveTur,
AdverbTur,
NumeralTur,
StructuralTur,
SentenceTur,
AdverbTur
QuestionTur,
RelativeTur,
ConjunctionTur,
TextX,
StructuralTur,
PhraseTur,
IdiomTur,
TenseX
** {
flags startcat = Phr ; unlexer = text ; lexer = text ;
flags startcat = Phr ;
} ;

View File

@@ -5,38 +5,55 @@
resource HarmonyTur = ParamX ** open Prelude, Predef in {
flags
coding=utf8 ;
--Harmony
param
-- Consonant are divided into 2 groups: Voiceless vs Voiced or Hard vs Soft.
-- This parameter type is used for consonant harmony, namely hardening and softening rules.
-- Harmony
param
-- Consonant are divided into 2 groups: Voiceless vs Voiced or Hard vs Soft.
-- This parameter type is used for consonant harmony, namely hardening and
-- softening rules.
Softness = Soft | Hard ;
-- Parameter type for consonant harmony:
-- Suffixes should have three forms at the worst case for consonant harmony, these forms are
-- used when stem ends with:
-- 1) soft consonant
-- 2) hard consonant
-- 3) vowel
-- Parameter type for consonant harmony:
-- Suffixes should have three
-- forms at the worst case for consonant harmony, these forms are
-- used when stem ends with:
-- 1) soft consonant
-- 2) hard consonant
-- 3) vowel
HarConP = SCon Softness | SVow ;
-- Parameter type for vowel harmony:
-- Suffixes should have 4 forms, because of two dimensional vowel harmony
HarVowP = I_Har | U_Har | Ih_Har | Uh_Har ;
-- Parameter type for vowel harmony:
-- Suffixes should have 4 forms, because of two dimensional vowel harmony
HarVowP = I_Har | U_Har | Ih_Har | Uh_Har ;
oper
-- Some pattern macros used by some opers (especially those related to harmonies) in ResTur.gf and ParadigmsTur.gf
-- Some pattern macros used by some opers (especially those related to
-- harmonies) in ResTur.gf and ParadigmsTur.gf
frontVowel : pattern Str = #("e"|"i"|"ö"|"ü") ;
backVowel : pattern Str = #("a"|"ı"|"o"|"u") ;
--Capital forms of vowels are also added, otherwise harmony of proper nouns like "Of" can not be determined
vowel : pattern Str = #("a"|"e"|"ı"|"i"|"u"|"ü"|"o"|"ö"|"î"|"â"|"û"|"A"|"E"|"I"|"İ"|"U"|"Ü"|"O"|"Ö"|"Î"|"Â"|"Û") ;
consonant : pattern Str = #("b"|"v"|"d"|"z"|"j"|"c"|"g"|"ğ"|"l"|"r"|"m"|"n"|"y"|"p"|"f"|"t"|"s"|"ş"|"ç"|"k"|"h") ;
--Extended consonant are used when proccessing words that contain non-letter characters like "stand-by"
extConson : pattern Str = #("b"|"v"|"d"|"z"|"j"|"c"|"g"|"ğ"|"l"|"r"|"m"|"n"|"y"|"p"|"f"|"t"|"s"|"ş"|"ç"|"k"|"h"|"'"|"-") ;
--The following are the hard (voiced) consonant in Turkish Alphabet (Order is determined by "Fıstıı Şahap" :) )
-- Capital forms of vowels are also added, otherwise harmony of proper
-- nouns like "Of" can not be determined
vowel : pattern Str =
#("a"|"e"|"ı"|"i"|"u"|"ü"|"o"|"ö"|
"|"â"|"û"|"A"|"E"|"I"|"İ"|"U"|"Ü"|
"O"|"Ö"|"Î"|"Â"|"Û") ;
consonant : pattern Str =
#("b"|"v"|"d"|"z"|"j"|"c"|"g"|"ğ"| "l"|"r"|"m"|"n"|
"y"|"p"|"f"|"t"|"s"|"ş"|"ç"|"k"|"h") ;
-- Extended consonant are used when proccessing words that contain
-- non-letter characters like "stand-by"
extConson : pattern Str =
#("b"|"v"|"d"|"z"|"j"|"c"|"g"|"ğ"|"l"|"r"|"m"|"n"|"y"|
"p"|"f"|"t"|"s"|"ş"|"ç"|"k"|"h"|"'"|"-") ;
-- The following are the hard (voiced) consonant in Turkish Alphabet
-- (Order is determined by "Fıstıı Şahap" :) )
hardCons : pattern Str = #("f"|"s"|"t"|"k"|"ç"|"ş"|"h"|"p") ;
-- Type definition and constructor of Harmony.
-- Type definition and constructor of Harmony.
Harmony = {
vow : HarVowP ;
con : HarConP
@@ -58,48 +75,55 @@ resource HarmonyTur = ParamX ** open Prelude, Predef in {
_ + ("u"|"o"|"û"|"U"|"O"|"Û") + #extConson* => U_Har ;
_ + ("ü"|"ö"|"Ü"|"Ö") + #extConson* => Uh_Har ;
_ => Ih_Har --this is for yiyor ("y" is base in that case)
} ;
} ;
-- Param base : a word, of which softness is to be determined
-- Returns whether Soft or Hard form of suffix will be used when adding a suffix to base
getSoftness : Str -> Softness = \base -> case dp 1 base of {
#hardCons => Hard ;
_ => Soft
} ;
-- Param base : a word, of which softness is to be determined
-- Returns whether Soft or Hard form of suffix will be used when adding a
-- suffix to base
getSoftness : Str -> Softness =
\base ->
case dp 1 base of {
#hardCons => Hard ;
_ => Soft
} ;
-- Param larC : the consonant form of suffix of which softness is to be determined
-- Returns whether Soft or Hard form of base will be used when adding a suffix to parameter
getBeginType : Str -> Softness = \larC -> case take 1 larC of {
#vowel => Soft ;
_ => Hard
} ;
-- Param larC : the consonant form of suffix of which softness is to be determined
-- Returns whether Soft or Hard form of base will be used when adding a
-- suffix to parameter
getBeginType : Str -> Softness =
\larC ->
case take 1 larC of {
#vowel => Soft ;
_ => Hard
} ;
-- Param base : a word
-- Returns which SuffixForm will be used when adding a suffix to base
-- Param base : a word
-- Returns which SuffixForm will be used when adding a suffix to base
getHarConP : Str -> HarConP =
\base -> case dp 1 base of {
#vowel => SVow ;
_ => SCon (getSoftness base)
} ;
-- returns softened form of a base
softenBase : Str -> Str = \base ->
let root = tk 1 base in
case base of {
_+ "p" => root + "b" ;
_+ "ç" => root + "c" ;
_+ "t" => root + "d" ;
_+ #consonant + "k" => root + "g" ;
_+ #vowel + ("k"|"g") => root + "ğ" ;
_ => base
} ;
-- harmony of progresive form of a verb is always the same, therefore it is constructed here to avoid unnecessary computation
progHar : Harmony
= {
vow = U_Har ;
con = SCon Soft ;
} ;
#vowel => SVow ;
_ => SCon (getSoftness base)
} ;
-- Returns softened form of a base
softenBase : Str -> Str =
\base ->
let
root = tk 1 base
in
case base of {
_+ "p" => root + "b" ;
_+ "ç" => root + "c" ;
_+ "t" => root + "d" ;
_+ #consonant + "k" => root + "g" ;
_+ #vowel + ("k"|"g") => root + "ğ" ;
_ => base
} ;
-- harmony of progresive form of a verb is always the same, therefore it is
-- constructed here to avoid unnecessary computation
progHar : Harmony = {
vow = U_Har ;
con = SCon Soft ;
} ;
}

5
src/turkish/IdiomTur.gf Normal file
View File

@@ -0,0 +1,5 @@
concrete IdiomTur of Idiom = CatTur ** open Prelude, ResTur in {
}

View File

@@ -245,7 +245,7 @@ concrete LexiconTur of Lexicon = CatTur **
wide_A = mkA "geniş" ;
yellow_A = mkA "sarı" ;
young_A = mkA "genç" ;
married_A2 = mkA2 (mkA "evli") with_Prep ;
married_A2 = mkA2 (mkA "evli") (mkPrep "ile" Nom) ;
--Verbs
add_V3 = mkV3 (mkV "eklemek") ;
answer_V2S = mkV2S (mkV "yanıtlamak") ;

View File

@@ -17,6 +17,9 @@ concrete NounTur of Noun = CatTur ** open ResTur, SuffixTur, HarmonyTur, Prelude
UsePron p = p ;
-- TODO: look further into how correct this is.
UsePN pn = { s = \\c => pn.s ! Sg ! c; a = {n = Sg; p = P1}} ;
PossPron p = {s = []; useGen = YesGen p.a} ;
DetQuant quant num = {

File diff suppressed because it is too large Load Diff

30
src/turkish/PhraseTur.gf Normal file
View File

@@ -0,0 +1,30 @@
concrete PhraseTur of Phrase = CatTur ** open Prelude, ResTur in {
lin
PhrUtt pconj utt voc = {s = pconj.s ++ utt.s ++ voc.s} ;
-- The following are utterly untested.
-- Currently, they should be treated as just implementation stubs.
UttQS qs = {s = qs.s} ;
UttImpSg pol imp = {s = imp.s} ;
UttImpPl pol imp = {s = imp.s} ;
UttImpPol pol imp = {s = imp.s} ;
UttIP ip = {s = ip.s} ;
UttIAdv iadv = iadv ;
UttCard n = {s = n.s ! Sg ! Nom} ;
UttInterj i = i ;
-- The following have been somewhat tested and seem to be working fine
-- to some extent.
UttNP np = {s = np.s ! Nom} ;
UttVP vp = {s = vp.s ! VInfinitive} ;
UttAP ap = {s = ap.s ! Sg ! Nom} ;
UttCN n = {s = n.s ! Sg ! Nom} ;
UttS s = s ;
UttAdv adv = adv ;
NoPConj = {s = []} ;
PConjConj conj = {s = conj.s} ;
NoVoc = {s = []} ;
VocNP np = {s = np.s ! Nom} ;
}

View File

@@ -0,0 +1,4 @@
concrete QuestionTur of Question = CatTur ** open ResTur, Prelude in {
}

View File

@@ -0,0 +1,4 @@
concrete RelativeTur of Relative = CatTur ** open ResTur in {
}

View File

@@ -14,7 +14,11 @@ resource ResTur = ParamX ** open Prelude, Predef, HarmonyTur in {
oper
Agr = {n : Number ; p : Person} ;
Noun = {s : Number => Case => Str; gen : Number => Agr => Str; harmony : Harmony} ;
Noun = {
s : Number => Case => Str ;
gen : Number => Agr => Str ;
harmony : Harmony
} ;
Pron = {s : Case => Str; a : Agr} ;
agrP3 : Number -> Agr ;
@@ -32,6 +36,7 @@ resource ResTur = ParamX ** open Prelude, Predef, HarmonyTur in {
| VAorist Agr
| VImperative
| VInfinitive
| Gerund Number Case
;
UseGen = NoGen | YesGen Agr | UseIndef ;
@@ -70,4 +75,14 @@ resource ResTur = ParamX ** open Prelude, Predef, HarmonyTur in {
mkClause : Str -> Agr -> Verb -> {s : Str} =
\np, a, v -> ss (np ++ v.s ! VProg a) ;
attachMe : Verb -> {s : Str} =
\v ->
let
s : Str = v.s ! VImperative
in
case s of {
(_ + #vowel + _ )* + (_ + #frontVowel + _) => ss (s ++ "me") ;
(_ + #vowel + _)* + (_ + #backVowel + _) => ss (s ++ "ma")
} ;
}

View File

@@ -4,4 +4,7 @@ concrete SentenceTur of Sentence = CatTur ** open Prelude, ResTur in {
PredVP np vp = mkClause (np.s ! Nom) np.a vp ;
-- TODO: Check how correct this is.
EmbedVP vp = {s = (vp.s ! Gerund Sg Acc)} ;
}

View File

@@ -87,4 +87,10 @@ concrete StructuralTur of Structural = CatTur **
but_PConj = ss "ama" ;
at_most_AdN = ss "en fazla" ;
at_least_AdN = ss "en az" ;
as_CAdv = {s = "kadar"; p = "kadar"} ;
}

View File

@@ -278,27 +278,30 @@ resource SuffixTur = open Prelude, Predef, ResTur, HarmonyTur in {
_ => wordC.p1
} ;
-- An auxiallary oper that fills in SuffixTable, used to avoid copy-paste
makeH4Table : {p1 : Str ; p2 : Str ; p3 : Str ; p4 : Str ;} -> {p1 : Str ; p2 : Str ; p3 : Str ; p4 : Str ; p5 : Str ; p6 : Str ; p7 : Str ; p8 : Str ;} -> SuffixTable =
\lirV,lirC ->
table {
SVow => table {
I_Har => lirV.p1 ;
Ih_Har => lirV.p2 ;
U_Har => lirV.p3 ;
Uh_Har => lirV.p4
} ;
SCon Soft => table {
I_Har => lirC.p1 ;
Ih_Har => lirC.p2 ;
U_Har => lirC.p3 ;
Uh_Har => lirC.p4
} ;
SCon Hard => table {
I_Har => lirC.p5 ;
Ih_Har => lirC.p6 ;
U_Har => lirC.p7 ;
Uh_Har => lirC.p8
}
} ;
-- An auxiliary oper that fills in SuffixTable, used to avoid copy-paste
makeH4Table : {p1 : Str ; p2 : Str ; p3 : Str ; p4 : Str ;}
-> {p1 : Str ; p2 : Str ; p3 : Str ; p4 : Str ;
p5 : Str ; p6 : Str ; p7 : Str ; p8 : Str }
-> SuffixTable =
\lirV,lirC ->
table {
SVow => table {
I_Har => lirV.p1 ;
Ih_Har => lirV.p2 ;
U_Har => lirV.p3 ;
Uh_Har => lirV.p4
} ;
SCon Soft => table {
I_Har => lirC.p1 ;
Ih_Har => lirC.p2 ;
U_Har => lirC.p3 ;
Uh_Har => lirC.p4
} ;
SCon Hard => table {
I_Har => lirC.p5 ;
Ih_Har => lirC.p6 ;
U_Har => lirC.p7 ;
Uh_Har => lirC.p8
}
} ;
}