diff --git a/devel/phrases/Makefile b/devel/phrases/Makefile new file mode 100644 index 000000000..3d1b90362 --- /dev/null +++ b/devel/phrases/Makefile @@ -0,0 +1,71 @@ +GF_GRAMMAR_ABS = Travel +GF_GRAMMAR_ENG = $(GF_GRAMMAR_ABS)Eng +GF_GRAMMAR_THA = $(GF_GRAMMAR_ABS)Tha +GF_GRAMMAR_THP = $(GF_GRAMMAR_ABS)ThaiP +GF_GRAMMAR_CNC = $(GF_GRAMMAR_ABS)ThaiP $(GF_GRAMMAR_ABS)Tha $(GF_GRAMMAR_ABS)Eng +GF_GRAMMAR_CNC_FILES = $(addsuffix .gf, $(GF_GRAMMAR_CNC)) +GF_GRAMMAR_FILES = $(addsuffix .gf, $(GF_GRAMMAR_ABS)) $(GF_GRAMMAR_CNC_FILES) +GEN_FILES = $(addsuffix .grxml, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .gram, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .jsgf, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .jsgf, $(GF_GRAMMAR_THA)) \ + $(addsuffix .jsgf, $(GF_GRAMMAR_THP)) \ + $(addsuffix .vxml, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .vxml-generic, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .vxml-grxml, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .vxml-gram, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .vxml-jsgf, $(GF_GRAMMAR_ENG)) \ + $(addsuffix .js, $(GF_GRAMMAR_ABS)) +SRG_FORMAT = gram + +.PHONY: all clean + +all: $(GEN_FILES) + + + +%.grxml: %.gf + echo "pg -printer=srgs_xml_sisr_old | wf $@" | gf -s -nocpu -batch $^ + tidy -q -xml -i -wrap 200 -m $@ +# Work around tidy bug + perl -i -pe 's/ lang=/ xml:lang=/' $@ + +%.gram: %.gf + echo "pg -printer=srgs_abnf_sisr_old | wf $@" | gf -s -nocpu -batch $^ + +%.jsgf: %.gf + echo "pg -printer=jsgf_sisr_old | wf $@" | gf -s -nocpu -batch $^ + +%.vxml: %.vxml-$(SRG_FORMAT) + cp $^ $@ + +%.vxml-generic: %.gf + echo 'pg -printer=vxml | wf $@' | gf -s -nocpu -batch $^ + tidy -q -xml -i -wrap 200 -m $@ +# Work around tidy bug + perl -i -pe 's/ lang=/ xml:lang=/' $@ +# Work around Opera bug + perl -i -pe "s/ src=\"#/ src=\"$*.vxml#/" $@ + +%.vxml-grxml: %.vxml-generic + cp $^ $@ + +%.vxml-gram: %.vxml-generic + cp $^ $@ + perl -i -pe 's/\.grxml/\.gram/' $@ + +%.vxml-jsgf: %.vxml-generic + cp $^ $@ + perl -i -pe 's/\.grxml/\.jsgf/' $@ +# Work around Opera for Zaurus bug + perl -i -pe 's/ src="(.*\.jsgf)#(\w+)"/ src="$$1" root="$$2"/' $@ + +$(GF_GRAMMAR_ABS).js: $(GF_GRAMMAR_FILES) + echo "pm -printer=js | wf $@" | gf -s -nocpu -batch $(GF_GRAMMAR_CNC_FILES) + +gflib.js: $(GF_LIB_PATH)/javascript/gflib.js + cat $^ > $@ + +clean: + -rm -f $(GEN_FILES) + -rm -f *.gfc diff --git a/devel/phrases/Numeral.gf b/devel/phrases/Numeral.gf new file mode 100644 index 000000000..5c8f06e97 --- /dev/null +++ b/devel/phrases/Numeral.gf @@ -0,0 +1,50 @@ +--1 Numerals + +-- This grammar defines numerals from 1 to 999999. +-- The implementations are adapted from the +-- [numerals library http://www.cs.chalmers.se/~aarne/GF/examples/numerals/] +-- which defines numerals for 88 languages. +-- The resource grammar implementations add to this inflection (if needed) +-- and ordinal numbers. +-- +-- *Note* 1. Number 1 as defined +-- in the category $Numeral$ here should not be used in the formation of +-- noun phrases, and should therefore be removed. Instead, one should use +-- [Structural Structural.html]$.one_Quant$. This makes the grammar simpler +-- because we can assume that numbers form plural noun phrases. +-- +-- *Note* 2. The implementations introduce spaces between +-- parts of a numeral, which is often incorrect - more work on +-- (un)lexing is needed to solve this problem. + +abstract Numeral = { + +cat + Numeral ; + Digit ; -- 2..9 + Sub10 ; -- 1..9 + Sub100 ; -- 1..99 + Sub1000 ; -- 1..999 + Sub1000000 ; -- 1..999999 + +fun + num : Sub1000000 -> Numeral ; + + n2, n3, n4, n5, n6, n7, n8, n9 : Digit ; + + pot01 : Sub10 ; -- 1 + pot0 : Digit -> Sub10 ; -- d * 1 + pot110 : Sub100 ; -- 10 + pot111 : Sub100 ; -- 11 + pot1to19 : Digit -> Sub100 ; -- 10 + d + pot0as1 : Sub10 -> Sub100 ; -- coercion of 1..9 + pot1 : Digit -> Sub100 ; -- d * 10 + pot1plus : Digit -> Sub10 -> Sub100 ; -- d * 10 + n + pot1as2 : Sub100 -> Sub1000 ; -- coercion of 1..99 + pot2 : Sub10 -> Sub1000 ; -- m * 100 + pot2plus : Sub10 -> Sub100 -> Sub1000 ; -- m * 100 + n + pot2as3 : Sub1000 -> Sub1000000 ; -- coercion of 1..999 + pot3 : Sub1000 -> Sub1000000 ; -- m * 1000 + pot3plus : Sub1000 -> Sub1000 -> Sub1000000 ; -- m * 1000 + n + +} diff --git a/devel/phrases/NumeralEng.gf b/devel/phrases/NumeralEng.gf new file mode 100644 index 000000000..af84485a8 --- /dev/null +++ b/devel/phrases/NumeralEng.gf @@ -0,0 +1,77 @@ +--# -path=.:prelude + +concrete NumeralEng of Numeral = open Prelude in { + +lincat + Numeral = {s : Str} ; ---{s : CardOrd => Str ; n : Num} ; + Digit = {s : DForm => CardOrd => Str} ; + Sub10 = {s : DForm => CardOrd => Str ; n : Num} ; + Sub100 = {s : CardOrd => Str ; n : Num} ; + Sub1000 = {s : CardOrd => Str ; n : Num} ; + Sub1000000 = {s : CardOrd => Str ; n : Num} ; + +lin num x = {s = x.s ! NCard} ; ---- +lin n2 = let two = mkNum "two" "twelve" "twenty" "second" in + {s = \\f,c => case of { + => "twelfth" ; + _ => two.s ! f ! c + } + } ; + +lin n3 = mkNum "three" "thirteen" "thirty" "third" ; +lin n4 = mkNum "four" "fourteen" "forty" "fourth" ; +lin n5 = mkNum "five" "fifteen" "fifty" "fifth" ; +lin n6 = regNum "six" ; +lin n7 = regNum "seven" ; +lin n8 = mkNum "eight" "eighteen" "eighty" "eighth" ; +lin n9 = mkNum "nine" "nineteen" "ninety" "ninth" ; + +lin pot01 = mkNum "one" "eleven" "ten" "first" ** {n = Sg} ; +lin pot0 d = d ** {n = Pl} ; +lin pot110 = regCardOrd "ten" ** {n = Pl} ; +lin pot111 = regCardOrd "eleven" ** {n = Pl} ; +lin pot1to19 d = {s = d.s ! teen} ** {n = Pl} ; +lin pot0as1 n = {s = n.s ! unit} ** {n = n.n} ; +lin pot1 d = {s = d.s ! ten} ** {n = Pl} ; +lin pot1plus d e = { + s = \\c => d.s ! ten ! NCard ++ "-" ++ e.s ! unit ! c ; n = Pl} ; +lin pot1as2 n = n ; +lin pot2 d = {s = \\c => d.s ! unit ! NCard ++ mkCard c "hundred"} ** {n = Pl} ; +lin pot2plus d e = { + s = \\c => d.s ! unit ! NCard ++ "hundred" ++ "and" ++ e.s ! c ; n = Pl} ; +lin pot2as3 n = n ; +lin pot3 n = { + s = \\c => n.s ! NCard ++ mkCard c "thousand" ; n = Pl} ; +lin pot3plus n m = { + s = \\c => n.s ! NCard ++ "thousand" ++ m.s ! c ; n = Pl} ; + +oper + mkNum : Str -> Str -> Str -> Str -> {s : DForm => CardOrd => Str} = + \two, twelve, twenty, second -> + {s = table { + unit => table {NCard => two ; NOrd => second} ; + teen => \\c => mkCard c twelve ; + ten => \\c => mkCard c twenty + } + } ; + + regNum : Str -> {s : DForm => CardOrd => Str} = + \six -> mkNum six (six + "teen") (six + "ty") (regOrd six) ; + + regCardOrd : Str -> {s : CardOrd => Str} = \ten -> + {s = table {NCard => ten ; NOrd => regOrd ten}} ; + + mkCard : CardOrd -> Str -> Str = \c,ten -> + (regCardOrd ten).s ! c ; + + regOrd : Str -> Str = \ten -> + case last ten of { + "y" => init ten + "ieth" ; + _ => ten + "th" + } ; + +param Num = Sg | Pl ; + CardOrd = NCard | NOrd ; + DForm = unit | teen | ten ; + +} diff --git a/devel/phrases/NumeralTha.gf b/devel/phrases/NumeralTha.gf new file mode 100644 index 000000000..3b398c23d --- /dev/null +++ b/devel/phrases/NumeralTha.gf @@ -0,0 +1,70 @@ +--# -path=.:prelude:resource-1.0/thai + +concrete NumeralTha of Numeral = open StringsTha in { + +flags coding=utf8 ; unlexer=concat ; + +lincat + Numeral = {s : Str} ; + Digit = {s : DForm => Str} ; + Sub10 = {s : DForm => Str} ; + Sub100 = {s : NForm => Str} ; + Sub1000 = {s : NForm => Str} ; + Sub1000000 = {s : Str} ; + +lin + num x = x ; + + pot01 = mkNum nvg_s nvg_s et_s ; + + n2 = mkNum soog_s yii_s soog_s ; + n3 = regNum saam_s ; + n4 = regNum sii_s ; + n5 = regNum haa_s ; + n6 = regNum hok_s ; + n7 = regNum cet_s ; + n8 = regNum peet_s ; + n9 = regNum kaaw_s ; + + + pot0 d = d ; + + pot110 = {s = sip} ; + pot111 = {s = table { + Unit => sip_s ++ et_s ; + Thousand => nvg_s ++ mvvn_s ++ nvg_s ++ phan_s + } + } ; + pot1to19 d = {s = table { + Unit => sip_s ++ d.s ! After ; + Thousand => nvg_s ++ mvvn_s ++ d.s ! Indep ++ phan_s + } + } ; + pot0as1 d = {s = \\n => d.s ! Indep ++ phan ! n} ; + pot1 d = {s = \\n => d.s ! ModTen ++ sip ! n} ; + pot1plus d e = { + s = \\n => d.s ! ModTen ++ sip ! n ++ e.s ! After ++ phan ! n + } ; + pot1as2 n = n ; + pot2 d = {s = \\n => d.s ! Indep ++ roy ! n} ; + pot2plus d e = {s = \\n => d.s ! Indep ++ roy ! n ++ e.s ! n} ; + pot2as3 n = {s = n.s ! Unit} ; + pot3 n = {s = n.s ! Thousand} ; + pot3plus n m = {s = n.s ! Thousand ++ m.s ! Unit} ; + +param + DForm = Indep | ModTen | After ; + NForm = Unit | Thousand ; + +oper + mkNum : Str -> Str -> Str -> {s : DForm => Str} = \x,y,z -> + {s = table {Indep => x ; ModTen => y ; After => z}} ; + regNum : Str -> {s : DForm => Str} = \x -> + mkNum x x x ; + + + sip = table {Unit => sip_s ; Thousand => mvvn_s} ; + roy = table {Unit => rooy_s ; Thousand => seen_s} ; + phan = table {Unit => [] ; Thousand => phan_s} ; + +} diff --git a/devel/phrases/NumeralThaiP.gf b/devel/phrases/NumeralThaiP.gf new file mode 100644 index 000000000..fedb0fc1c --- /dev/null +++ b/devel/phrases/NumeralThaiP.gf @@ -0,0 +1,86 @@ +--# -path=.:prelude:resource-1.0/thai + +concrete NumeralThaiP of Numeral = { + +lincat + Numeral = {s : Str} ; + Digit = {s : DForm => Str} ; + Sub10 = {s : DForm => Str} ; + Sub100 = {s : NForm => Str} ; + Sub1000 = {s : NForm => Str} ; + Sub1000000 = {s : Str} ; + +lin + num x = x ; + + pot01 = mkNum nvg_s nvg_s et_s ; + + n2 = mkNum soog_s yii_s soog_s ; + n3 = regNum saam_s ; + n4 = regNum sii_s ; + n5 = regNum haa_s ; + n6 = regNum hok_s ; + n7 = regNum cet_s ; + n8 = regNum peet_s ; + n9 = regNum kaaw_s ; + + pot0 d = d ; + + pot110 = {s = sip} ; + pot111 = {s = table { + Unit => sip_s ++ et_s ; + Thousand => nvg_s ++ mvvn_s ++ nvg_s ++ phan_s + } + } ; + pot1to19 d = {s = table { + Unit => sip_s ++ d.s ! After ; + Thousand => nvg_s ++ mvvn_s ++ d.s ! Indep ++ phan_s + } + } ; + pot0as1 d = {s = \\n => d.s ! Indep ++ phan ! n} ; + pot1 d = {s = \\n => d.s ! ModTen ++ sip ! n} ; + pot1plus d e = { + s = \\n => d.s ! ModTen ++ sip ! n ++ e.s ! After ++ phan ! n + } ; + pot1as2 n = n ; + pot2 d = {s = \\n => d.s ! Indep ++ roy ! n} ; + pot2plus d e = {s = \\n => d.s ! Indep ++ roy ! n ++ e.s ! n} ; + pot2as3 n = {s = n.s ! Unit} ; + pot3 n = {s = n.s ! Thousand} ; + pot3plus n m = {s = n.s ! Thousand ++ m.s ! Unit} ; + +oper + phan_s = "pahn" ; + rooy_s = "rawy" ; + mvvn_s = "meun" ; + seen_s = "sain" ; + + nvg_s = "neung" ; + soog_s = "song" ; + saam_s = "sahm" ; + sii_s = "see" ; + haa_s = "hah" ; + hok_s = "hok" ; + cet_s = "jet" ; + peet_s = "baat" ; + kaaw_s = "gow" ; + sip_s = "sip" ; + yii_s = "yee" ; + et_s = "et" ; + +param + DForm = Indep | ModTen | After ; + NForm = Unit | Thousand ; + +oper + mkNum : Str -> Str -> Str -> {s : DForm => Str} = \x,y,z -> + {s = table {Indep => x ; ModTen => y ; After => z}} ; + regNum : Str -> {s : DForm => Str} = \x -> + mkNum x x x ; + + + sip = table {Unit => sip_s ; Thousand => mvvn_s} ; + roy = table {Unit => rooy_s ; Thousand => seen_s} ; + phan = table {Unit => [] ; Thousand => phan_s} ; + +} diff --git a/devel/phrases/Numerals.gf b/devel/phrases/Numerals.gf new file mode 100644 index 000000000..90aae3cb7 --- /dev/null +++ b/devel/phrases/Numerals.gf @@ -0,0 +1,32 @@ +-- numerals from 1 to 999999 in decimal notation + +flags startcat=Numeral ; + +cat + Numeral ; -- 0.. + Digit ; -- 2..9 + Sub10 ; -- 1..9 + Sub100 ; -- 1..99 + Sub1000 ; -- 1..999 + Sub1000000 ; -- 1..999999 + +fun + num : Sub1000000 -> Numeral ; + + n2, n3, n4, n5, n6, n7, n8, n9 : Digit ; + + pot01 : Sub10 ; -- 1 + pot0 : Digit -> Sub10 ; -- d * 1 + pot110 : Sub100 ; -- 10 + pot111 : Sub100 ; -- 11 + pot1to19 : Digit -> Sub100 ; -- 10 + d + pot0as1 : Sub10 -> Sub100 ; -- coercion of 1..9 + pot1 : Digit -> Sub100 ; -- d * 10 + pot1plus : Digit -> Sub10 -> Sub100 ; -- d * 10 + n + pot1as2 : Sub100 -> Sub1000 ; -- coercion of 1..99 + pot2 : Sub10 -> Sub1000 ; -- m * 100 + pot2plus : Sub10 -> Sub100 -> Sub1000 ; -- m * 100 + n + pot2as3 : Sub1000 -> Sub1000000 ; -- coercion of 1..999 + pot3 : Sub1000 -> Sub1000000 ; -- m * 1000 + pot3plus : Sub1000 -> Sub1000 -> Sub1000000 ; -- m * 1000 + n + diff --git a/devel/phrases/NumeralsEng.gf b/devel/phrases/NumeralsEng.gf new file mode 100644 index 000000000..270b0c9d6 --- /dev/null +++ b/devel/phrases/NumeralsEng.gf @@ -0,0 +1,43 @@ +include numerals.Abs.gf ; + +param DForm = unit | teen | ten ; + +lincat Numeral = { s : Str } ; +lincat Digit = {s : DForm => Str} ; +lincat Sub10 = {s : DForm => Str} ; +lincat Sub100 = { s : Str } ; +lincat Sub1000 = { s : Str } ; +lincat Sub1000000 = { s : Str } ; + +oper mkNum : Str -> Str -> Str -> Lin Digit = + \two -> \twelve -> \twenty -> + {s = table {unit => two ; teen => twelve ; ten => twenty}} ; +oper regNum : Str -> Lin Digit = + \six -> mkNum six (six + "teen") (six + "ty") ; +oper ss : Str -> {s : Str} = \s -> {s = s} ; + +lin num x = x ; +lin n2 = mkNum "two" "twelve" "twenty" ; +lin n3 = mkNum "three" "thirteen" "thirty" ; +lin n4 = mkNum "four" "fourteen" "forty" ; +lin n5 = mkNum "five" "fifteen" "fifty" ; +lin n6 = regNum "six" ; +lin n7 = regNum "seven" ; +lin n8 = mkNum "eight" "eighteen" "eighty" ; +lin n9 = regNum "nine" ; + +lin pot01 = {s = table {f => "one"}} ; +lin pot0 d = {s = table {f => d.s ! f}} ; +lin pot110 = ss "ten" ; +lin pot111 = ss "eleven" ; +lin pot1to19 d = {s = d.s ! teen} ; +lin pot0as1 n = {s = n.s ! unit} ; +lin pot1 d = {s = d.s ! ten} ; +lin pot1plus d e = {s = d.s ! ten ++ "-" ++ e.s ! unit} ; +lin pot1as2 n = n ; +lin pot2 d = {s = d.s ! unit ++ "hundred"} ; +lin pot2plus d e = {s = d.s ! unit ++ "hundred" ++ "and" ++ e.s} ; +lin pot2as3 n = n ; +lin pot3 n = {s = n.s ++ "thousand"} ; +lin pot3plus n m = {s = n.s ++ "thousand" ++ m.s} ; + diff --git a/devel/phrases/PizzaDraw.gf b/devel/phrases/PizzaDraw.gf new file mode 100644 index 000000000..7edca6bc7 --- /dev/null +++ b/devel/phrases/PizzaDraw.gf @@ -0,0 +1,4 @@ +concrete PizzaDraw of Pizza = { + + +} \ No newline at end of file diff --git a/devel/phrases/Travel.gf b/devel/phrases/Travel.gf new file mode 100644 index 000000000..f7fa73bb5 --- /dev/null +++ b/devel/phrases/Travel.gf @@ -0,0 +1,54 @@ +abstract Travel = Numeral ** { + +cat + Order ; +cat + Output ; + +fun + confirm : Order -> Number -> Output ; + + +-- the essential phrases from Lone Planet Thai Phrasebook + + order : Phrase -> Order ; + +cat + Phrase ; + Number ; + +fun + Hello : Phrase ; + Goodbye : Phrase ; + Please : Phrase ; + ThankYou : Phrase ; + YoureWelcome : Phrase ; + Yes : Phrase ; + No : Phrase ; + ExcuseAttention : Phrase ; + ExcuseGetPast : Phrase ; + Sorry : Phrase ; + IUnderstand : Phrase ; + IDontUnderstand : Phrase ; + Help : Phrase ; + WhereAreToilets : Phrase ; + + + SayNumber : Numeral -> Phrase ; + + One, Two, Three, Four, Five, Six, Seven, Eight, Nine, Ten : Number ; + + +cat + Product ; + Kind ; + +fun + HowMuchCost : Product -> Order ; + + This : Kind -> Product ; + + Beer : Kind ; + Shirt : Kind ; + +} diff --git a/devel/phrases/TravelEng.gf b/devel/phrases/TravelEng.gf new file mode 100644 index 000000000..e04e7fce4 --- /dev/null +++ b/devel/phrases/TravelEng.gf @@ -0,0 +1,86 @@ +--# -path=.:prelude:resource-1.0/thai + +concrete TravelEng of Travel = NumeralEng ** open Prelude in { + +flags startcat = Order; language = en_US; + +lincat + Order = { s : Str } ; + +printname cat + Order = "What would you like to say?" ; + +lin + order is = { s = is.s } ; + + + +lincat + Output = { s : Str } ; + +lin + confirm o t = { s = o.s} ; + +flags unlexer=unwords ; + +lincat + Phrase = SS ; + Number = SS ; + +lin + Hello = ss "hello" ; + Goodbye = ss "bye" ; + Please = ss "please" ; + ThankYou = ss "thanks" ; + YoureWelcome = ss ["you are welcome"] ; + Yes = ss "yes" ; + No = ss "no" ; + ExcuseAttention = ss ["excuse me"] ; + ExcuseGetPast = ss ["excuse me"] ; + Sorry = ss "sorry" ; + IUnderstand = ss ["I understand"] ; + IDontUnderstand = ss ["I do not understand"] ; + Help = ss "help" ; + WhereAreToilets = ss ["where are the toilets"] ; + + + SayNumber n = n ; + + One = ss "one" ; + Two = ss "two" ; + Three = ss "three" ; + Four = ss "four" ; + Five = ss "five" ; + Six = ss "six" ; + Seven = ss "seven" ; + Eight = ss "eight" ; + Nine = ss "nine" ; + Ten = ss "ten" ; + +lincat + Product = {s : Str} ; + Kind = {s : Str} ; + +printname cat + Product = "what product do you mean?" ; + Kind = "what kind of product do you mean?" ; + +lin + HowMuchCost p = {s = ["how much does"] ++ + variants { + p.s ; + -- no kind given + "this" ; + -- no product given at all + "it" + } ++ + "cost"} ; + + This k = {s = "this" ++ k.s} ; + + Beer = {s = "beer"} ; + Shirt = {s = "shirt"} ; + + + +} diff --git a/devel/phrases/TravelTha.gf b/devel/phrases/TravelTha.gf new file mode 100644 index 000000000..d697b0db5 --- /dev/null +++ b/devel/phrases/TravelTha.gf @@ -0,0 +1,80 @@ +--# -path=.:prelude:resource-1.0/thai + +concrete TravelTha of Travel = NumeralTha ** open Prelude, StringsTha in { + +flags startcat = Order; language = en_US; coding=utf8 ; + +lincat + Order = { s : Str } ; + +printname cat + Order = "What would you like to say?" ; + +lin + order is = { s = is.s } ; + + + +lincat + Output = { s : Str } ; + +lin + confirm o t = { s = o.s } ; + + + + +flags unlexer=concat ; + +lincat + Phrase = SS ; + Number = SS ; + +lin + Hello = ss (sawat_s ++ dii_s) ; + Goodbye = ss (laa_s ++ koon_s) ; + Please = ss (khoo_s) ; + ThankYou = ss (khoop_s ++ khun_s) ; + YoureWelcome = ss (yin_s ++ dii_s) ; + Yes = ss (chay_s) ; + No = ss (may_s) ; + ExcuseAttention = ss (khoo_s ++ thoot_s) ; + ExcuseGetPast = ss (khoo_s ++ aphai_s) ; + Sorry = ss (khoo_s ++ thoot_s) ; + IUnderstand = ss (phom_s ++ khow_s ++ jai_s) ; + IDontUnderstand = ss (phom_s ++ may_s ++ khow_s ++ jai_s) ; + Help = ss (chuay_s ++ duay_s) ; + WhereAreToilets = ss (hoog_s ++ nam_s ++ yuu_s ++ thii_s ++ nai_s) ; + + + SayNumber n = n ; + + One = ss (nvg_s) ; + Two = ss (soog_s) ; + Three = ss (saam_s) ; + Four = ss (sii_s) ; + Five = ss (haa_s) ; + Six = ss (hok_s) ; + Seven = ss (cet_s) ; + Eight = ss (peet_s) ; + Nine = ss (kaaw_s) ; + Ten = ss (sip_s) ; + +lincat + Product = {s : Str} ; + Kind = {s : Str} ; + +printname cat + Product = "what product do you mean?" ; + Kind = "what kind of product do you mean?" ; + +lin + HowMuchCost p = ss (p.s ++ thao_s ++ rai_s) ; + + This k = ss (k.s ++ nii_s) ; + + Beer = ss biar_s ; + Shirt = ss (seua_s ++ cheut_s) ; + + +} \ No newline at end of file diff --git a/devel/phrases/TravelThaiP.gf b/devel/phrases/TravelThaiP.gf new file mode 100644 index 000000000..5462563ec --- /dev/null +++ b/devel/phrases/TravelThaiP.gf @@ -0,0 +1,79 @@ +--# -path=.:prelude:resource-1.0/thai + +concrete TravelThaiP of Travel = NumeralThaiP ** open Prelude, StringsTha in { + +flags startcat = Order; language = en_US ; + +lincat + Order = { s : Str } ; + +printname cat + Order = "What would you like to say?" ; + +lin + order is = { s = is.s } ; + + + +lincat + Output = { s : Str } ; + +lin + confirm o t = { s = o.s } ; + + + + +flags unlexer=unwords ; + +lincat + Phrase = SS ; + Number = SS ; + +lin + Hello = ss ["sah wut dee"] ; + Goodbye = ss ["lah gorn"] ; + Please = ss "kor" ; + ThankYou = ss ["kop koon"] ; + YoureWelcome = ss ["yin dee"] ; + Yes = ss "chai" ; + No = ss "mai" ; + ExcuseAttention = ss ["koh tort"] ; + ExcuseGetPast = ss ["koh ahpai"] ; + Sorry = ss ["koh tort"] ; + IUnderstand = ss ["pom kow jai"] ; + IDontUnderstand = ss ["pom mai kow jai"] ; + Help = ss ["chew wai dewai"] ; + WhereAreToilets = ss ["hong narm yoo tee nai"] ; + + + SayNumber n = n ; + + One = ss "neung" ; + Two = ss "song" ; + Three = ss "sahm" ; + Four = ss "see" ; + Five = ss "hah" ; + Six = ss "hok" ; + Seven = ss "jet" ; + Eight = ss "baat" ; + Nine = ss "gow" ; + Ten = ss "sip" ; + +lincat + Product = {s : Str} ; + Kind = {s : Str} ; + +printname cat + Product = "what product do you mean?" ; + Kind = "what kind of product do you mean?" ; + +lin + HowMuchCost p = ss (p.s ++ "tao" ++ "rai") ; + + This k = ss (k.s ++ "nee") ; + + Beer = ss "beea" ; + Shirt = ss ("seua" ++ "cheut") ; + +} \ No newline at end of file diff --git a/devel/phrases/flash-controls.js b/devel/phrases/flash-controls.js new file mode 100644 index 000000000..142ed3740 --- /dev/null +++ b/devel/phrases/flash-controls.js @@ -0,0 +1,18 @@ +function getFlashMovieObject(movieName) { + if (window.document[movieName]) { + return window.document[movieName]; + } + if (document.embeds && document.embeds[movieName]) { + return document.embeds[movieName]; + } else { + return document.getElementById(movieName); + } +} + +function flashPlay(movieName) { + getFlashMovieObject(movieName).Play(); +} + +function flashPause(movieName) { + getFlashMovieObject(movieName).StopPlay(); +} diff --git a/devel/phrases/gflib.js b/devel/phrases/gflib.js new file mode 100644 index 000000000..e3f30b42f --- /dev/null +++ b/devel/phrases/gflib.js @@ -0,0 +1,252 @@ +/* Abstract syntax trees */ +function Fun(name) { + this.name = name; + this.args = copy_arguments(arguments, 1); +} +Fun.prototype.print = function () { return this.show(0); } ; +Fun.prototype.show = function (prec) { + if (this.isMeta()) { + if (isUndefined(this.type)) { + return '?'; + } else { + var s = '?:' + this.type; + if (prec > 0) { + s = "(" + s + ")" ; + } + return s; + } + } else { + var s = this.name; + var cs = this.args; + for (var i in cs) { + s += " " + cs[i].show(1); + } + if (prec > 0 && cs.length > 0) { + s = "(" + s + ")" ; + } + return s; + } +}; +Fun.prototype.getArg = function (i) { + return this.args[i]; +}; +Fun.prototype.setArg = function (i,c) { + this.args[i] = c; +}; +Fun.prototype.isMeta = function() { + return this.name == '?'; +} ; +Fun.prototype.isComplete = function() { + if (this.isMeta()) { + return false; + } else { + for (var i in tree.args) { + if (!tree.args[i].isComplete()) { + return false; + } + } + return true; + } +} ; + +/* Concrete syntax terms */ + +function Arr() { this.arr = copy_arguments(arguments, 0); } +Arr.prototype.tokens = function() { return this.arr[0].tokens(); }; +Arr.prototype.sel = function(i) { return this.arr[i.toIndex()]; }; + +function Seq() { this.seq = copy_arguments(arguments, 0); } +Seq.prototype.tokens = function() { + var xs = new Array(); + for (var i in this.seq) { + var ys = this.seq[i].tokens(); + for (var j in ys) { + xs.push(ys[j]); + } + } + return xs; +}; + +function Variants() { this.variants = copy_arguments(arguments, 0); } +Variants.prototype.tokens = function() { return this.variants[0].tokens(); }; + +function Rp(index,value) { this.index = index; this.value = value; } +Rp.prototype.tokens = function() { return new Array(this.index); }; +Rp.prototype.toIndex = function() { return this.index.toIndex(); }; + +function Suffix(prefix,suffix) { this.prefix = prefix; this.suffix = suffix; }; +Suffix.prototype.tokens = function() { + var xs = this.suffix.tokens(); + for (var i in xs) { + xs[i] = this.prefix + xs[i]; + } + return xs; +}; +Suffix.prototype.sel = function(i) { return new Suffix(this.prefix, this.suffix.sel(i)); }; + +function Meta() { } +Meta.prototype.tokens = function() { return new Array("?"); }; +Meta.prototype.toIndex = function() { return 0; }; +Meta.prototype.sel = function(i) { return this; }; + +function Str(value) { this.value = value; } +Str.prototype.tokens = function() { return new Array(this.value); }; + +function Int(value) { this.value = value; } +Int.prototype.tokens = function() { return new Array(this.value.toString()); }; +Int.prototype.toIndex = function() { return this.value; }; + +/* Type annotation */ + +function Abstract() { + this.types = new Array(); +} +Abstract.prototype.addType = function(fun, args, cat) { + this.types[fun] = new Type(args, cat); +} ; +Abstract.prototype.annotate = function(tree, type) { + if (tree.name == '?') { + tree.type = type; + } else { + var typ = this.types[tree.name]; + for (var i in tree.args) { + this.annotate(tree.args[i], typ.args[i]); + } + } + return tree; +} ; +/* Hack to get around the fact that our SISR doesn't build real Fun objects. */ +Abstract.prototype.copyTree = function(x) { + var t = new Fun(x.name); + if (!isUndefined(x.type)) { + t.type = x.type; + } + var cs = x.args; + if (!isUndefined(cs)) { + for (var i in cs) { + t.setArg(i, this.copyTree(cs[i])); + } + } + return t; +} ; +Abstract.prototype.parseTree = function(str, type) { + return this.annotate(this.parseTree_(str.match(/[\w\']+|\(|\)|\?/g), 0), type); +} ; +Abstract.prototype.parseTree_ = function(tokens, prec) { + if (tokens.length == 0 || tokens[0] == ")") { return null; } + var t = tokens.shift(); + if (t == "(") { + var tree = this.parseTree_(tokens, 0); + tokens.shift(); + return tree; + } else if (t == '?') { + return new Fun('?'); + } else { + var tree = new Fun(t); + if (prec == 0) { + var c, i; + for (i = 0; (c = this.parseTree_(tokens, 1)) !== null; i++) { + tree.setArg(i,c); + } + } + return tree; + } +} ; + +function Type(args, cat) { + this.args = args; + this.cat = cat; +} + +/* Linearization */ + +function Concrete(abstr) { + this.abstr = abstr; + this.rules = new Array(); +} +Concrete.prototype.rule = function (name, cs) { return this.rules[name](cs); }; +Concrete.prototype.addRule = function (name, f) { this.rules[name] = f; }; +Concrete.prototype.lindef = function (cat, v) { return this.rules["_d"+cat]([new Str(v)]); } ; +Concrete.prototype.linearize = function (tree) { + return this.unlex(this.linearizeToTerm(tree).tokens()); +}; +Concrete.prototype.linearizeToTerm = function (tree) { + if (tree.isMeta()) { + if (isUndefined(tree.type)) { + return new Meta(); + } else { + return this.lindef(tree.type, tree.name); + } + } else { + var cs = new Array(); + for (var i in tree.args) { + cs.push(this.linearizeToTerm(tree.args[i])); + } + return this.rule(tree.name, cs); + } +}; +Concrete.prototype.unlex = function (ts) { + if (ts.length == 0) { + return ""; + } + + var noSpaceAfter = /^[\(\-\[]/; + var noSpaceBefore = /^[\.\,\?\!\)\:\;\-\]]/; + + var s = ""; + for (var i = 0; i < ts.length; i++) { + var t = ts[i]; + var after = i < ts.length-1 ? ts[i+1] : null; + s += t; + if (after != null && !t.match(noSpaceAfter) + && !after.match(noSpaceBefore)) { + s += " "; + } + } + return s; +}; + + +/* Utilities */ + +/* from Remedial JavaScript by Douglas Crockford, http://javascript.crockford.com/remedial.html */ +function isString(a) { return typeof a == 'string'; } +function isArray(a) { return a && typeof a == 'object' && a.constructor == Array; } +function isUndefined(a) { return typeof a == 'undefined'; } +function isBoolean(a) { return typeof a == 'boolean'; } +function isNumber(a) { return typeof a == 'number' && isFinite(a); } +function isFunction(a) { return typeof a == 'function'; } + +function dumpObject (obj) { + if (isUndefined(obj)) { + return "undefined"; + } else if (isString(obj)) { + return '"' + obj.toString() + '"'; // FIXME: escape + } else if (isBoolean(obj) || isNumber(obj)) { + return obj.toString(); + } else if (isArray(obj)) { + var x = "["; + for (var i in obj) { + x += dumpObject(obj[i]); + if (i < obj.length-1) { + x += ","; + } + } + return x + "]"; + } else { + var x = "{"; + for (var y in obj) { + x += y + "=" + dumpObject(obj[y]) + ";" ; + } + return x + "}"; + } +} + + +function copy_arguments(args, start) { + var arr = new Array(); + for (var i = 0; i < args.length - start; i++) { + arr[i] = args[i + start]; + } + return arr; +} diff --git a/devel/phrases/index.html b/devel/phrases/index.html new file mode 100644 index 000000000..539887d60 --- /dev/null +++ b/devel/phrases/index.html @@ -0,0 +1,79 @@ + + + + English-Thai Phrase Translator + + + + + +

English-Thai Phrase Translator

+ +

This document describes the + GF XHTML+Voice English-Thai Phrase Translator, + a demonstration of + XHTML+Voice + dialog system generated from a + Grammatical Framework grammar. + For a more detailed explanation of how this generation is done, + see the article Generating Dialog Systems from Grammars. + This demo was built by reusing code and ideas from Björn Bringert's + Pizza Demo. +

+ + + +

Try the demo

+ +

You can try the demo if you + have a web browser which can handle + XHTML+Voice + and + SVG. + Currently this only includes + Opera for Windows (when voice controlled + browsing is enabled). See + Using Opera with Voice + for more information.

+ +

There is a slightly simpler version of the demo which also works + on the Opera multimodal browser for the Sharp Zaurus. It will be added here shortly.

+ + +

Functionality

+ + +

References

+ + + +
Aarne Ranta, + aarne@cs.chalmers.se.
+ + + diff --git a/devel/phrases/order-simple.js b/devel/phrases/order-simple.js new file mode 100644 index 000000000..4243ad67e --- /dev/null +++ b/devel/phrases/order-simple.js @@ -0,0 +1,49 @@ +var currentOrder = new Fun("?"); +var talkText; + + +function say(text) { + talkText = text; + activateForm("talker"); +} + +function newOrder() { + currentOrder = new Fun("?"); + + document.getElementById("top_abs").value = ""; + document.getElementById("top_img").value = ""; + + document.getElementById("ordertext").value = ""; + + return getOrder(); +} + +function getOrder() { + activateForm("getorder"); + return true; +} + + +function done(input) { + currentOrder = Pizza.copyTree(input, "Order"); + document.getElementById("top_abs").value = currentOrder.print(); + + sayOrder(); +} + +function sayOrder() { + var eng = PizzaEng.linearize(currentOrder); + document.getElementById("ordertext").value = eng; + say("You have ordered " + eng); +} + + + +/* XHTML+Voice Utilities */ + +function activateForm(formid) { + var form = document.getElementById(formid); + var e = document.createEvent("UIEvents"); + e.initEvent("DOMActivate","true","true"); + form.dispatchEvent(e); +} diff --git a/devel/phrases/order.js b/devel/phrases/order.js new file mode 100644 index 000000000..b1f10548d --- /dev/null +++ b/devel/phrases/order.js @@ -0,0 +1,70 @@ +var svgNS = "http://www.w3.org/2000/svg"; + +var currentOrder = new Fun("?"); + +var talkText; + +function say(text) { + talkText = text; + activateForm("talker"); +} + +function newOrder() { + currentOrder = new Fun("?"); + + document.getElementById("in_abs").value = ""; + + setText(document.getElementById("ordertext"), ""); + setText(document.getElementById("ordertextf"), ""); + setText(document.getElementById("ordertextt"), ""); + + return getOrder(); +} + +function getOrder() { + activateForm("getorder"); + return true; +} + +function done(input) { + currentOrder = Travel.copyTree(input); + document.getElementById("in_abs").value = currentOrder.print(); + + sayOrder(); +} + +function sayOrder() { + var output = currentOrder; + var eng = TravelEng.linearize(output); + setText(document.getElementById("ordertext"), eng); + + var fin = TravelTha.linearize(output).replace(/ /g,""); + setText(document.getElementById("ordertextf"), fin); + var tha = TravelThaiP.linearize(output); + setText(document.getElementById("ordertextt"), tha); + say(tha); +} + + + +/* XHTML+Voice Utilities */ + +function activateForm(formid) { + var form = document.getElementById(formid); + var e = document.createEvent("UIEvents"); + e.initEvent("DOMActivate","true","true"); + form.dispatchEvent(e); +} + +/* DOM utilities */ + +function removeChildren(node) { + while (node.hasChildNodes()) { + node.removeChild(node.firstChild); + } + } + +function setText(node, text) { + removeChildren(node); + node.appendChild(document.createTextNode(text)); +} diff --git a/devel/phrases/pizza-movie-large.html b/devel/phrases/pizza-movie-large.html new file mode 100644 index 000000000..906eb7169 --- /dev/null +++ b/devel/phrases/pizza-movie-large.html @@ -0,0 +1,39 @@ + + + + Pizza Demo - GF XHTML+Voice + + + + + +

Pizza Demo - GF XHTML+Voice

+ +

This is a demo of a dialog system built with GF and XHTML+Voice. +There is more information about this demo here.

+ + +
+

+ + +

+
+ +

+ + + + + + + +

+ + + diff --git a/devel/phrases/pizza-simple.html b/devel/phrases/pizza-simple.html new file mode 100644 index 000000000..772c6f808 --- /dev/null +++ b/devel/phrases/pizza-simple.html @@ -0,0 +1,62 @@ + + + + + +Order a Pizza + + + + + + + + + + + + + + + + + + + + + + + + +

+ +
+
+ +
+
+
+
+ +
+
+ +
+
+ +
+
+

+Current order state
+
+ +

+
+
+ + + diff --git a/devel/phrases/pizza.xml b/devel/phrases/pizza.xml new file mode 100644 index 000000000..1d1f60731 --- /dev/null +++ b/devel/phrases/pizza.xml @@ -0,0 +1,73 @@ + + + + + Say Phrases in Thai + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
+ +

This demo requires an XHTML+Voice browser. + More information about this demo.

+ +

+ +
+ +

+

+

+ +

+ +
+ + +
+
+ + +
+
+ + + + + + + + + + + + + diff --git a/devel/phrases/travel.xml b/devel/phrases/travel.xml new file mode 100644 index 000000000..29e8b51e0 --- /dev/null +++ b/devel/phrases/travel.xml @@ -0,0 +1,73 @@ + + + + + Say Phrases in Thai + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
ภาษาไทยแปลว่าอะไร
+ +

This demo requires an XHTML+Voice browser. + More information about this demo.

+ +

+ +
+ +

+ +

+

+ +

+ +
+ + +
+
+ +
+
+ + + + + + + + + + + + + diff --git a/src/GF/Text/Thai.hs b/src/GF/Text/Thai.hs index 2b9456f06..1b186cb3a 100644 --- a/src/GF/Text/Thai.hs +++ b/src/GF/Text/Thai.hs @@ -11,7 +11,9 @@ -- AR 27/12/2006. Execute test2 to see the transliteration table. -module GF.Text.Thai (mkThai,mkThaiWord,mkThaiPron,thaiFile,thaiPronFile) where +module GF.Text.Thai ( + mkThai,mkThaiWord,mkThaiPron,mkThaiFake,thaiFile,thaiPronFile,thaiFakeFile + ) where import qualified Data.Map as Map import Data.Char @@ -26,6 +28,7 @@ import Debug.Trace mkThai :: String -> String mkThai = concat . map mkThaiWord . words mkThaiPron = unwords . map mkPronSyllable . words +mkThaiFake = unwords . map (fakeEnglish . mkPronSyllable) . words type ThaiChar = Char @@ -78,7 +81,7 @@ allThaiTrans :: [String] allThaiTrans = words $ "- k k1 - k2 - k3 g c c1 c2 s' c3 y' d' t' " ++ "t1 t2 t3 n' d t t4 t5 t6 n b p p1 f p2 f' " ++ - "p3 m y r - l - w s- r' s h l' O h' - " ++ + "p3 m y r - l - w s- s. s h l' O h' - " ++ "a. a a: a+ i i: v v: u u: - - - - - - " ++ "e e' o: a% a& L R S T1 T2 T3 T4 K - - - " ++ "N0 N1 N2 N3 N4 N5 N6 N7 N8 N9 - - - - - - " @@ -91,6 +94,42 @@ allThaiCodes = [0x0e00 .. 0x0e7f] -- heuristic pronunciation of codes --------------------- +-- fake English for TTS, a la Teach Yourself Thai + +fakeEnglish :: String -> String +fakeEnglish s = case s of + 'a':'a':cs -> "ah" ++ fakeEnglish cs + 'a':'y':cs -> "ai" ++ fakeEnglish cs + 'a' :cs -> "ah" ++ fakeEnglish cs + 'c':'h':cs -> "ch" ++ fakeEnglish cs + 'c' :cs -> "j" ++ fakeEnglish cs + 'e':'e':cs -> "aih" ++ fakeEnglish cs + 'g' :cs -> "ng" ++ fakeEnglish cs + 'i':'i':cs -> "ee" ++ fakeEnglish cs + 'k':'h':cs -> "k" ++ fakeEnglish cs + 'k' :cs -> "g" ++ fakeEnglish cs + 'O':'O':cs -> "or" ++ fakeEnglish cs + 'O' :cs -> "or" ++ fakeEnglish cs + 'o':'o':cs -> "or" ++ fakeEnglish cs + 'p':'h':cs -> "p" ++ fakeEnglish cs + 'p' :cs -> "b" ++ fakeEnglish cs + 't':'h':cs -> "t" ++ fakeEnglish cs + 't' :cs -> "d" ++ fakeEnglish cs + 'u':'u':cs -> "oo" ++ fakeEnglish cs + 'u' :cs -> "oo" ++ fakeEnglish cs + 'v':'v':cs -> "eu" ++ fakeEnglish cs + 'v' :cs -> "eu" ++ fakeEnglish cs + '\228':'\228':cs -> "air" ++ fakeEnglish cs + '\228' :cs -> "a" ++ fakeEnglish cs + '\246':'\246':cs -> "er" ++ fakeEnglish cs + '\246' :cs -> "er" ++ fakeEnglish cs + c:cs | isTone c -> fakeEnglish cs + c:cs -> c : fakeEnglish cs + _ -> s + where + isTone = flip elem "'`^~" + + -- this works for one syllable mkPronSyllable s = case fst $ pronAndOrth s of @@ -124,17 +163,17 @@ pronSyllable s = ([0x0e40],[0x0e35],_,[0x0e22],_,_) -> "ia" -- e-i:y ([0x0e40],[0x0e2d,0x0e35],_,_,_,_) -> "va" -- e-i:O ([0x0e40],[0x0e30,0x0e35],_,[0x0e22],_,_) -> "ia" -- e-i:ya. - ([0x0e40],[0x0e30,0x0e2d],_,_,_,_) -> "ö" -- e-Oa. + ([0x0e40],[0x0e30,0x0e2d],_,_,_,_) -> "\246" -- e-Oa. ([0x0e40],[0x0e30,0x0e32],_,_,_,_) -> "O" -- e-a:a. -- open o - ([0x0e40],[0x0e2d],_,_,_,_) -> "öö" -- e-O - ([0x0e40],[0x0e34],_,_,_,_) -> "öö" -- e-i + ([0x0e40],[0x0e2d],_,_,_,_) -> "\246\246" -- e-O + ([0x0e40],[0x0e34],_,_,_,_) -> "\246\246" -- e-i ([0x0e40],[0x0e30],_,_,_,_) -> "e" -- e-a. ([0x0e40],[0x0e32],_,_,_,_) -> "aw" -- e-a: - ([0x0e40],[],[],[0x0e22],_,_) -> "ööy" -- e-y + ([0x0e40],[],[],[0x0e22],_,_) -> "\246\246y" -- e-y ([0x0e40],[],[],_,True,_) -> "e" - ([0x0e41],[0x0e30],_,_,_,_) -> "ä" -- ä-a. - ([0x0e41],[],[],_,True,_) -> "ä" + ([0x0e41],[0x0e30],_,_,_,_) -> "\228" -- ä-a. + ([0x0e41],[],[],_,True,_) -> "\228" ([0x0e42],[0x0e30],_,_,_,_) -> "o" -- o:-a. @@ -245,6 +284,12 @@ thaiPronFile f mo = do let put = maybe putStr writeFile mo put $ encodeUTF8 $ thaiPronStrings s +thaiFakeFile :: FilePath -> Maybe FilePath -> IO () +thaiFakeFile f mo = do + s <- readFile f + let put = maybe putStr writeFile mo + put $ encodeUTF8 $ (convStrings mkThaiFake) s + finalThai c = maybe "" return (Map.lookup c thaiFinalMap) thaiFinalMap = Map.fromList $ zip allThaiCodes finals @@ -303,7 +348,7 @@ pronThai s = case s of | p==':' -> c:[c] | elem p "%&" -> c:"y" | p=='+' -> c:"m" - | s == "e'" -> "ää" + | s == "e'" -> "\228\228" | otherwise -> [c] "O" -> "O" "e" -> "ee"