wip: noun morphology

.dir-locals.el

@@ -0,0 +1,6 @@
+((nil
+  . ((eval
+      . (progn (add-to-list 'load-path (project-root (project-current)))
+               (require 'helpers)
+               (add-hook 'before-save-hook
+                         #'hangul-convert-buffer-to-jamo t))))))

flake.nix

@@ -33,6 +33,7 @@
 		inputs.gf.packages.${system}.gf-with-rgl
 		pkgs.graphviz
 		gf-lsp.gf-lsp
+		pkgs.babashka
 	      ];
 	    };
 	  });

helpers.el

@@ -0,0 +1,56 @@
+(defun hangul-syllables-to-jamo (str)
+  "Convert HANGUL SYLLABLES characters in STR to their HANGUL JAMO
+equivalents."
+  (let ((result "")
+	(i 0))
+    (while (< i (length str))
+      (let ((char (aref str i)))
+	(if (and (>= char #xAC00) (<= char #xD7A3))
+	    ;; Hangul syllable
+	    (let* ((code (- char #xAC00))
+		   (lead (/ code (* 21 28)))
+		   (medial (/ (% code (* 21 28)) 28))
+		   (final (% code 28))
+		   (lead-jamo (+ #x1100 lead))
+		   (medial-jamo (+ #x1161 medial))
+		   (final-jamo (if (> final 0) (+ #x11A7 final) nil)))
+	      (setq result
+		    (concat result (char-to-string lead-jamo)
+			    (char-to-string medial-jamo)
+			    (if final-jamo (char-to-string final-jamo) ""))))
+	  ;; Not a Hangul syllable
+	  (setq result (concat result (char-to-string char)))))
+      (setq i (1+ i)))
+    result))
+
+(defun hangul-convert-region-to-jamo (beg end)
+  (interactive "r")
+  (replace-region-contents
+   beg end (lambda ()
+	     (hangul-syllables-to-jamo (buffer-substring (point-min)
+							 (point-max))))))
+
+(defun hangul-convert-buffer-to-jamo ()
+  (interactive)
+  (hangul-convert-region-to-jamo (point-min) (point-max))
+  (message "Converted Hangul Syllables in buffer to Jamo."))
+
+(require 'dash)
+
+(defconst gf-hangul/choseong
+  (cl-loop for i from #x1100 to #x1112 collect i))
+
+(defconst gf-hangul/jungseong
+  (cl-loop for i from #x1161 to #x1175 collect i))
+
+(defconst gf-hangul/batchim
+  (cl-loop for i from #x11a8 to #x11c2 collect i))
+
+(defun gf-hangul/make-pattern (name seq)
+  (format "'%s' : pattern Str = #(%s) ;"
+	  name (->> seq
+		    (--map (concat "\"" it "\""))
+		    (-interpose " | ")
+		    (apply #'concat))))
+
+(provide 'helpers)

lab1/grammar/korean/HangulJamo.gf

@@ -0,0 +1,72 @@
+resource HangulJamo = open Prelude in {
+  flags coding=utf8 ;
+  oper
+    lemmaToStem : (lemma : Str) -> Str
+      = \lemma -> case lemma of {
+        stem + "다" => stem ;
+        _ => Predef.error ("lemmaToStem was applied to a non-lemma," ++ lemma)
+      } ;
+
+    infinitive : (stem : Str) -> Str
+      = \stem -> case stem of {
+	-- 하다 ⇒ 해
+	init + #ha => init + "해" ;
+	-- 찾다 ⇒ 찾아, 좁다 ⇒ 좁아
+        _ + #a_o + #batchim => stem + "아" ;
+	-- 가다 ⇒ 가, 
+        _ + (#a | #eo) => stem ;
+	-- 오다 ⇒ 와
+        init + #o => init + "ᅪ" ;
+	-- 따르다 ⇒ 따러
+        init + #eu => init + "ᅥ" ;
+	-- 기다리다 ⇒ 기다려
+        init + #i => init + "ᅧ" ;
+	--
+        init + #u => init + "ᅯ" ;
+	-- 맛있다 ⇒ 맛있어
+        _ => stem + "어"
+      } ;
+
+    present_haeyo : (stem : Str) -> Str
+      = \stem -> infinitive stem + "요" ;
+
+    past_haeyo : (stem : Str) -> Str
+      = \stem -> infinitive stem + "ᆻ어요" ;
+
+    vc_allomorph : (s,vowel,consonant : Str) -> Str
+      = \s,v,c -> case s of {
+	_ + #vowel => v ;
+	_ => c
+      } ;
+
+  oper
+    a_o : pattern Str = #("ᅡ" | "ᅩ") ;
+    a : pattern Str = #"ᅡ" ;
+    o : pattern Str = #"ᅩ" ;
+    eo : pattern Str = #"ᅥ" ;
+    eu : pattern Str = #"ᅳ" ;
+    i : pattern Str = #"ᅵ" ;
+    u : pattern Str = #"ᅮ" ;
+    ha : pattern Str = #"하" ;
+
+  oper
+    consonant : pattern Str =
+      #("ᄀ" | "ᄁ" | "ᄂ" | "ᄃ" | "ᄄ" | "ᄅ" | "ᄆ" | "ᄇ"
+          | "ᄈ" | "ᄉ" | "ᄊ" | "ᄋ" | "ᄌ" | "ᄍ" | "ᄎ" | "ᄏ"
+          | "ᄐ" | "ᄑ" | "ᄒ" | "ᆨ" | "ᆩ" | "ᆪ" | "ᆫ" | "ᆬ"
+          | "ᆭ" | "ᆮ" | "ᆯ" | "ᆰ" | "ᆱ" | "ᆲ" | "ᆳ" | "ᆴ"
+          | "ᆵ" | "ᆶ" | "ᆷ" | "ᆸ" | "ᆹ" | "ᆺ" | "ᆻ" | "ᆼ"
+          | "ᆽ" | "ᆾ" | "ᆿ" | "ᇀ" | "ᇁ" | "ᇂ") ;
+    batchim : pattern Str =
+      #("ᆨ" | "ᆩ" | "ᆪ" | "ᆫ" | "ᆬ" | "ᆭ" | "ᆮ" | "ᆯ" | "ᆰ"
+          | "ᆱ" | "ᆲ" | "ᆳ" | "ᆴ" | "ᆵ" | "ᆶ" | "ᆷ" | "ᆸ" | "ᆹ"
+          | "ᆺ" | "ᆻ" | "ᆼ" | "ᆽ" | "ᆾ" | "ᆿ" | "ᇀ" | "ᇁ" | "ᇂ") ;
+    choseong : pattern Str =
+      #("ᄀ" | "ᄁ" | "ᄂ" | "ᄃ" | "ᄄ" | "ᄅ" | "ᄆ" | "ᄇ" | "ᄈ"
+          | "ᄉ" | "ᄊ" | "ᄋ" | "ᄌ" | "ᄍ" | "ᄎ" | "ᄏ" | "ᄐ" | "ᄑ"
+          | "ᄒ") ;
+    vowel : pattern Str =
+      #("ᅡ" | "ᅢ" | "ᅣ" | "ᅤ" | "ᅥ" | "ᅦ" | "ᅧ" | "ᅨ" | "ᅩ"
+          | "ᅪ" | "ᅫ" | "ᅬ" | "ᅭ" | "ᅮ" | "ᅯ" | "ᅰ" | "ᅱ" | "ᅲ"
+          | "ᅳ" | "ᅴ" | "ᅵ") ;
+}

lab1/grammar/korean/MicroLangKor.gf

@@ -1,6 +1,8 @@
 --# -path=.:../abstract
 concrete MicroLangKor of MicroLang = open MicroResKor, Prelude in {
 
+  flags coding=utf8 ;
+
 -----------------------------------------------------
 ---------------- Grammar part -----------------------
 -----------------------------------------------------
@@ -9,11 +11,11 @@ concrete MicroLangKor of MicroLang = open MicroResKor, Prelude in {
     Utt = {s : Str} ;
 
     S  = {s : Str} ;
-    VP = {verb : Verb ; compl : Str} ; ---s special case of Mini
+    VP = {verb : Verb ; compl : N} ; ---s special case of Mini
     Comp = Noun ;
     AP = Adjective ;
     CN = Noun ;
-    NP = {s : Str} ;
+    NP = Noun ;
     Pron = {s : Str} ;
     Det = {s : Str} ;
     Prep = {s : Str} ;
@@ -24,10 +26,10 @@ concrete MicroLangKor of MicroLang = open MicroResKor, Prelude in {
     Adv = {s : Str} ;
 
   lin
-    UttS s = s ;
+    UttS s = {s = s.s} ;
-    UttNP np = {s = np.s} ;
+    UttNP np = {s = np.s ! NBare} ;
-    -- PredVPS np vp = {s = np.s ++ vp.compl ++ vp.verb.s} ;
+    PredVPS np vp = {s = np.s ! NSubject ++ vp.compl ! NObject ++ vp.verb.s ! VPresent} ;
-    UseV v = {verb = v ; compl = []} ;
+    UseV v = {verb = v ; compl = {s = \\_ => []}} ;
     ComplV2 v2 np = {verb = v2 ; compl = np.s};
     UseN n = n ;
     DetCN det cn = {s = det.s ++ cn.s} ;
@@ -35,101 +37,103 @@ concrete MicroLangKor of MicroLang = open MicroResKor, Prelude in {
     aPl_Det = {s = []} ;
     the_Det = {s = []} ;
     thePl_Det = {s = []} ;
-    this_Det = {s = "이"} ;
+    this_Det = {s = "이"} ;
-    thisPl_Det = {s = "이"} ;
+    thisPl_Det = {s = "이"} ;
-    that_Det = {s = "그"} ;
+    that_Det = {s = "그"} ;
-    thatPl_Det = {s = "그"} ;
+    thatPl_Det = {s = "그"} ;
+    PositA a = a ;
+    AdjCN ap cn = {s = ap.s ! VAdnomial ++ cn.s} ;
 
 -----------------------------------------------------
 ---------------- Lexicon part -----------------------
 -----------------------------------------------------
 
--- lin already_Adv = mkAdv "벌써" ;
+-- lin already_Adv = mkAdv "벌써" ;
-lin animal_N = mkN "동물" ;
+lin animal_N = mkN "동물" ;
-lin apple_N = mkN "사과" ;
+lin apple_N = mkN "사과" ;
-lin baby_N = mkN "아기" ;
+lin baby_N = mkN "아기" ;
--- lin bad_A = mkA "나쁜" ;
+lin bad_A = mkA "나쁘다" ;
-lin beer_N = mkN "beer" ;
+-- lin beer_N = mkN "beer" ;
--- lin big_A = mkA "큰" ;
+lin big_A = mkA "크다" ;
-lin bike_N = mkN "bike" ;
+-- lin bike_N = mkN "bike" ;
-lin bird_N = mkN "bird" ;
+-- lin bird_N = mkN "bird" ;
 -- lin black_A = mkA "black" ;
-lin blood_N = mkN "피" ;
+lin blood_N = mkN "피" ;
 -- lin blue_A = mkA "blue" ;
-lin boat_N = mkN "boat" ;
+-- lin boat_N = mkN "boat" ;
-lin book_N = mkN "책" ;
+lin book_N = mkN "책" ;
-lin boy_N = mkN "소녁" ;
+lin boy_N = mkN "소녁" ;
-lin bread_N = mkN "bread" ;
+lin bread_N = mkN "빵" ;
 -- lin break_V2 = mkV2 (mkV "break" "broke" "broken") ;
--- lin buy_V2 = mkV2 (mkV "buy" "bought" "bought") ;
+lin buy_V2 = mkV2 "사다" ;
-lin car_N = mkN "자동차" ;
+lin car_N = mkN "자동차" ;
-lin cat_N = mkN "고양이" ;
+lin cat_N = mkN "고양이" ;
-lin child_N = mkN "어린이" ;
+lin child_N = mkN "어린이" ;
-lin city_N = mkN "city" ;
+-- lin city_N = mkN "city" ;
--- lin clean_A = mkA "정소한" ;
+lin clean_A = mkA "정소하다" ;
--- lin clever_A = mkA "똑똑한" ;
+lin clever_A = mkA "똑똑하다" ;
-lin cloud_N = mkN "cloud" ;
+-- lin cloud_N = mkN "cloud" ;
--- lin cold_A = mkA "차가운" ;
+-- lin cold_A = mkA "차가운" ;
-lin come_V = regVerb_a_o "오다" ;
+lin come_V = mkV "오다" ;
-lin computer_N = mkN "컴퓨터" ;
+lin computer_N = mkN "컴퓨터" ;
-lin cow_N = mkN "cow" ;
+-- lin cow_N = mkN "cow" ;
--- lin dirty_A = mkA "더러운" ;
+-- lin dirty_A = mkA "더러운" ;
-lin dog_N = mkN "개" ;
+lin dog_N = mkN "개" ;
--- lin drink_V2 = mkV2 (mkV "drink" "drank" "drunk") ;
+lin drink_V2 = mkV2 "마시다" ;
--- lin eat_V2 = mkV2 (mkV "eat" "ate" "eaten") ;
+lin eat_V2 = mkV2 "먹다" ;
--- lin find_V2 = mkV2 (mkV "find" "found" "found") ;
+lin find_V2 = mkV2 "찾다" ;
-lin fire_N = mkN "fire" ;
+-- lin fire_N = mkN "fire" ;
-lin fish_N = mkN "생선" ;
+lin fish_N = mkN "생선" ;
-lin flower_N = mkN "flower" ;
+-- lin flower_N = mkN "flower" ;
-lin friend_N = mkN "진구" ;
+lin friend_N = mkN "진구" ;
-lin girl_N = mkN "소녀" ;
+lin girl_N = mkN "소녀" ;
--- lin good_A = mkA "좋은" ;
+-- lin good_A = mkA "좋은" ;
--- lin go_V = mkV "go" "went" "gone" ;
+lin go_V = mkV "가다" ;
-lin grammar_N = mkN "grammar" ;
+-- lin grammar_N = mkN "grammar" ;
 -- lin green_A = mkA "green" ;
 -- lin heavy_A = mkA "heavy" ;
-lin horse_N = mkN "horse" ;
+-- lin horse_N = mkN "horse" ;
 -- lin hot_A = mkA "hot" ;
-lin house_N = mkN "집" ;
+lin house_N = mkN "집" ;
 -- lin john_PN = mkPN "John" ;
 -- lin jump_V = mkV "jump" ;
--- lin kill_V2 = mkV2 "kill" ;
+lin kill_V2 = mkV2 "죽다" ;
--- lin know_VS = mkVS (mkV "know" "knew" "known") ;
+-- lin know_V2 = mkV "알다" ;
-lin language_N = mkN "언어" ;
+lin language_N = mkN "언어" ;
 -- lin live_V = mkV "live" ;
 -- lin love_V2 = mkV2 (mkV "love") ;
-lin man_N = mkN "남자" ;
+lin man_N = mkN "남자" ;
-lin milk_N = mkN "우유" ;
+lin milk_N = mkN "우유" ;
-lin music_N = mkN "음악" ;
+lin music_N = mkN "음악" ;
 -- lin new_A = mkA "new" ;
--- lin now_Adv = mkAdv "지금" ;
+lin now_Adv = mkAdv "지금" ;
--- lin old_A = mkA "낡안" ;
+-- lin old_A = mkA "낡안" ;
 -- lin paris_PN = mkPN "Paris" ;
--- lin play_V = mkV "놀" ;
+lin play_V = mkV "놀다" ;
--- lin read_V2 = mkV2 (mkV "read" "read" "read") ;
+lin read_V2 = mkV2 "읽다" ;
 -- lin ready_A = mkA "ready" ;
 -- lin red_A = mkA "red" ;
-lin river_N = mkN "river" ;
+lin river_N = mkN "강" ;
 -- lin run_V = mkV "run" "ran" "run" ;
-lin sea_N = mkN "바다" ;
+lin sea_N = mkN "바다" ;
--- lin see_V2 = mkV2 (mkV "see" "saw" "seen") ;
+lin see_V2 = mkV2 "보다" ;
-lin ship_N = mkN "ship" ;
+-- lin ship_N = mkN "ship" ;
-lin sleep_V = regVerb_a_o "자다" ;
+lin sleep_V = mkV "자다" ;
--- lin small_A = mkA "작은" ;
+-- lin small_A = mkA "작은" ;
-lin star_N = mkN "별" ;
+lin star_N = mkN "별" ;
--- lin swim_V = mkV "swim" "swam" "swum" ;
+lin swim_V = mkV "수영하다" ;
--- lin teach_V2 = mkV2 (mkV "teach" "taught" "taught") ;
+lin teach_V2 = mkV2 "가르치다" ;
-lin train_N = mkN "train" ;
+-- lin train_N = mkN "train" ;
 -- lin travel_V = mkV "travel" ;
-lin tree_N = mkN "tree" ;
+-- lin tree_N = mkN "tree" ;
 -- lin understand_V2 = mkV2 (mkV "understand" "understood" "understood") ;
--- lin wait_V2 = mkV2 "wait" "for" ;
+lin wait_V2 = mkV2 "기다리다" ;
--- lin walk_V = mkV "walk" ;
+lin walk_V = mkV "걷다" ;
--- lin warm_A = mkA "따뜻한" ;
+-- lin warm_A = mkA "따뜻한" ;
-lin water_N = mkN "물" ;
+lin water_N = mkN "물" ;
--- lin white_A = mkA "하얗은" ;
+-- lin white_A = mkA "하얗은" ;
-lin wine_N = mkN "wine" ;
+-- lin wine_N = mkN "wine" ;
-lin woman_N = mkN "여자" ;
+lin woman_N = mkN "여자" ;
 -- lin yellow_A = mkA "yellow" ;
 -- lin young_A = mkA "young" ;
 
@@ -139,14 +143,23 @@ lin woman_N = mkN "여자" ;
 
 oper
   mkN : Str -> Noun
-    = \s -> lin N {s = s} ;
+    = \s -> lin N (noun s) ;
 
-  -- mkA : Str -> A
+  mkV : Str -> V = \lemma -> lin V (regVerb lemma) ;
-  --   = \stem -> lin A (mkV stem) ;
+
+  mkV2 = overload {
+    mkV2 : Str -> V2
+      = \lemma -> lin V2 (mkV lemma) ;
+    mkV2 : V -> V2
+      = \v -> lin V2 v ;
+    } ;
 
   mkAdv : Str -> Adv
     = \s -> lin Adv {s = s} ;
 
   mkPrep : Str -> Prep
     = \s -> lin Prep {s = s} ;
+
+  mkA : Str -> A
+    = \lemma -> lin A (regAdjective lemma) ;
 }

lab1/grammar/korean/MicroResKor.gf

@@ -1,44 +1,47 @@
-resource MicroResKor = open Prelude, Hangul in {
+resource MicroResKor = open Prelude, HangulJamo in {
   param
-    VForm = Lemma | VPresent | VPast ;
+    VForm = VLemma | VPresent | VPast | VAdnomial ;
+    NForm = NBare | NTopic | NSubject | NObject | NPluralSubject | NPluralTopic ;
 
   oper
-    Noun : Type = {s : Str} ;
+    Noun : Type = {s : NForm => Str} ;
     Verb : Type = {s : VForm => Str} ;
     Verb2 : Type = Verb ;
     Adjective : Type = Verb ;
 
     lemmaToStem : (lemma : Str) -> Str
       = \lemma -> case lemma of {
-	stem + "다" => stem ;
+	stem + "다" => stem ;
 	_ => Predef.error ("lemmaToStem was applied to a non-lemma," ++ lemma)
       } ;
 
-    regVerb_eo : (lemma : Str) -> Verb
+    reg : (descriptive : Bool) -> (lemma : Str) -> Verb
-      = \lemma ->
+      = \descriptive,lemma ->
       let stem = lemmaToStem lemma ;
       in {
 	s = table {
-	  Lemma => lemma ;
+	  VLemma => lemma ;
-	  VPresent => stem + "어요" ;
+	  VPresent => present_haeyo stem ;
-	  VPast => stem + "었어요"
+	  VPast => past_haeyo stem ;
+	  VAdnomial =>
+	    if_then_else Str descriptive
+	      (stem + vc_allomorph stem "ᆫ" "은")
+	      (stem + "는")
 	  }
       } ;
 
-    regVerb_a_o : (lemma : Str) -> Verb
+    regVerb : (lemma : Str) -> Verb = reg False ;
-      = \lemma ->
+    regAdjective : (lemma : Str) -> Adjective = reg True ;
-      let stem = lemmaToStem lemma ;
+
-      in {
+    noun : (bare : Str) -> Noun
+      = \bare -> {
 	s = table {
-	  Lemma => lemma ;
+	  NBare => bare ;
-	  VPresent => stem + "아요" ;
+	  NTopic => bare + vc_allomorph bare "는" "은" ;
-	  VPast => stem + "았어요"
+	  NSubject => bare + vc_allomorph bare "이" "가" ;
+	  NObject => bare + vc_allomorph bare "를" "을" ;
+	  NPluralTopic => bare + "들은" ;
+	  NPluralSubject => bare + "들이"
 	  }
       } ;
-
-    blah : Str -> Str
-      = \s -> case s of {
-	_ + "ᅡ" => "아-final" ;
-	_ => "not 아-final"
-      } ;
 }

scripts/search-hangul

@@ -0,0 +1,12 @@
+#!/usr/bin/env bb
+
+(require '[babashka.fs :as fs]
+         '[clojure.string :as str])
+
+(doseq [f (fs/glob "lab1/grammar/korean" "**.gf")
+        [line-number line] (map-indexed (fn [i x] [i x])
+                                        (-> f str slurp str/split-lines))]
+  (when (re-find #"\p{block=HangulJamo}" line)
+    (printf "JAMO: %s:%d: %s\n" (str f) line-number line))
+  (when (re-find #"\p{block=HangulSyllables}" line)
+    (printf "SYLLABLE: %s:%d: %s\n" (str f) line-number line)))

scripts/syllables-to-jamo

@@ -0,0 +1,44 @@
+#!/usr/bin/env bb
+
+(require '[babashka.fs :as fs]
+         '[clojure.string :as str])
+
+(defn syllables->jamo [s]
+  "Convert Hangul syllables in string S to their jamo components."
+  (->> s
+       (map int)
+       (mapcat
+        (fn [c]
+          (if (<= 0xAC00 c 0xD7A3)
+            ;; Hangul syllable - decompose
+            (let [code (- c 0xAC00)
+                  lead (quot code (* 21 28))
+                  medial (quot (mod code (* 21 28)) 28)
+                  final (mod code 28)
+                  lead-jamo (+ 0x1100 lead)
+                  medial-jamo (+ 0x1161 medial)
+                  final-jamo (if (> final 0) (+ 0x11A7 final) nil)]
+              (remove nil? [lead-jamo medial-jamo final-jamo]))
+            ;; Not a Hangul syllable
+            [c])))
+       (map char)
+       (apply str)))
+
+
+(defn -main [& args]
+  (if (seq args)
+    (doseq [f args]
+      (let [x (-> (slurp f)
+                  (str/replace #"\p{block=HangulSyllables}+"
+                               syllables->jamo))]
+        (spit f x)))
+    (loop [line (read-line)]
+      (when line
+        (-> line
+            (str/replace #"\p{block=HangulSyllables}+"
+                         syllables->jamo)
+            println)
+        (recur (read-line))))))
+
+(when (= *file* (System/getProperty "babashka.file"))
+  (apply -main *command-line-args*))