code.delx.au - gnu-emacs/blob - lisp/language/lao-util.el

   1 ;;; lao-util.el --- utilities for Lao -*- coding: utf-8; -*-
   2
   3 ;; Copyright (C) 2001-2016 Free Software Foundation, Inc.
   4 ;; Copyright (C) 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006,
   5 ;;   2007, 2008, 2009, 2010, 2011
   6 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
   7 ;;   Registration Number H14PRO021
   8 ;; Copyright (C) 2003
   9 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
  10 ;;   Registration Number H13PRO009
  11
  12 ;; Keywords: multilingual, Lao, i18n
  13
  14 ;; This file is part of GNU Emacs.
  15
  16 ;; GNU Emacs is free software: you can redistribute it and/or modify
  17 ;; it under the terms of the GNU General Public License as published by
  18 ;; the Free Software Foundation, either version 3 of the License, or
  19 ;; (at your option) any later version.
  20
  21 ;; GNU Emacs is distributed in the hope that it will be useful,
  22 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  23 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  24 ;; GNU General Public License for more details.
  25
  26 ;; You should have received a copy of the GNU General Public License
  27 ;; along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.
  28
  29 ;;; Commentary:
  30
  31 ;;; Code:
  32
  33 ;; Setting information of Thai characters.
  34
  35 (defconst lao-category-table (make-category-table))
  36 (define-category ?c "Lao consonant" lao-category-table)
  37 (define-category ?s "Lao semi-vowel" lao-category-table)
  38 (define-category ?v "Lao upper/lower vowel" lao-category-table)
  39 (define-category ?t "Lao tone" lao-category-table)
  40
  41 (let ((l '((?ກ consonant "CHICKEN")
  42            (?ຂ consonant "EGG")
  43            (?຃ invalid)
  44            (?ຄ consonant "BUFFALO")
  45            (?຅ invalid)
  46            (?  invalid)
  47            (?ງ consonant "SNAKE")
  48            (?ຈ consonant "BUDDHIST NOVICE")
  49            (?ຉ invalid)
  50            (?ຊ consonant "ELEPHANT")
  51            (?຋ invalid)
  52            (?ຌ invalid)
  53            (?ຍ consonant "MOSQUITO")
  54            (?ຎ invalid)
  55            (?ຎ invalid)
  56            (?ຎ invalid)
  57            (?ຎ invalid)
  58            (?ຎ invalid)
  59            (?ຎ invalid)
  60            (?ດ consonant "NOSE")
  61            (?ຕ consonant "EYE")
  62            (?ຖ consonant "TO ASK,QUESTION")
  63            (?ທ consonant "FLAG")
  64            (?ຘ invalid)
  65            (?ນ consonant "BIRD")
  66            (?ບ consonant "FISHHOOK")
  67            (?ປ consonant "FISH")
  68            (?ຜ consonant "BEE")
  69            (?ຝ consonant "WALL")
  70            (?ພ consonant "MOUNTAIN")
  71            (?ຟ consonant "FIRE")
  72            (?ຠ invalid)
  73            (?ມ consonant "HORSE")
  74            (?ຢ consonant "MEDICINE")
  75            (?ຣ consonant "CAR")
  76            (?຤ invalid)
  77            (?ລ consonant "MONKEY")
  78            (?຦ invalid)
  79            (?ວ consonant "HAND FAN")
  80            (?ຨ invalid)
  81            (?ຩ invalid)
  82            (?ສ consonant "TIGER")
  83            (?ຫ consonant "JAR")
  84            (?ຬ invalid)
  85            (?ອ consonant "TAKE")
  86            (?ຮ consonant "BOAT")
  87            (?ຯ special)
  88            (?ະ vowel-base)
  89            (?ັ vowel-upper)
  90            (?າ vowel-base)
  91            (?ຳ vowel-base)
  92            (?ິ vowel-upper)
  93            (?ີ vowel-upper)
  94            (?ຶ vowel-upper)
  95            (?ື vowel-upper)
  96            (?ຸ vowel-lower)
  97            (?ູ vowel-lower)
  98            (?຺ invalid)
  99            (?ົ vowel-upper)
 100            (?ຼ semivowel-lower)
 101            (?ຽ vowel-base)
 102            (?຾ invalid)
 103            (?຿ invalid)
 104            (?ເ vowel-base)
 105            (?ແ vowel-base)
 106            (?ໂ vowel-base)
 107            (?ໃ vowel-base)
 108            (?ໄ vowel-base)
 109            (?໅ invalid)
 110            (?ໆ special)
 111            (?໇ invalid)
 112            (?່ tone)
 113            (?້ tone)
 114            (?໊ tone)
 115            (?໋ tone)
 116            (?໌ tone)
 117            (?ໍ vowel-upper)
 118            (?໎ invalid)
 119            (?໏ invalid)
 120            (?໐ special)
 121            (?໑ special)
 122            (?໒ special)
 123            (?໓ special)
 124            (?໔ special)
 125            (?໕ special)
 126            (?໖ special)
 127            (?໗ special)
 128            (?໘ special)
 129            (?໙ special)
 130            (?໚ invalid)
 131            (?໛ invalid)
 132            (?ໜ consonant "MOUSE")
 133            (?ໝ consonant "DOG")
 134            (?ໞ invalid)))
 135       elm)
 136   (while l
 137     (setq elm (car l) l (cdr l))
 138     (let ((char (car elm))
 139           (ptype (nth 1 elm)))
 140       (cond ((eq ptype 'consonant)
 141              (modify-category-entry char ?c lao-category-table))
 142             ((memq ptype '(vowel-upper vowel-lower))
 143              (modify-category-entry char ?v lao-category-table))
 144             ((eq ptype 'semivowel-lower)
 145              (modify-category-entry char ?s lao-category-table))
 146             ((eq ptype 'tone)
 147              (modify-category-entry char ?t lao-category-table)))
 148       (put-char-code-property char 'phonetic-type ptype)
 149       (put-char-code-property char 'meaning (nth 2 elm)))))
 150
 151 ;; The general composing rules are as follows:
 152 ;;
 153 ;;                          T
 154 ;;       V        T         V                  T
 155 ;; CV -> C, CT -> C, CVT -> C, Cv -> C, CvT -> C
 156 ;;                                   v         v
 157 ;;                             T
 158 ;;        V         T          V                    T
 159 ;; CsV -> C, CsT -> C, CsVT -> C, Csv -> C, CsvT -> C
 160 ;;        s         s          s         s          s
 161 ;;                                       v          v
 162
 163
 164 ;; where C: consonant, V: vowel upper, v: vowel lower,
 165 ;;       T: tone mark, s: semivowel lower
 166
 167 (defvar lao-composition-pattern
 168   "\\cc\\(\\ct\\|\\cv\\ct?\\|\\cs\\(\\ct\\|\\cv\\ct?\\)?\\)"
 169   "Regular expression matching a Lao composite sequence.")
 170
 171 ;;;###autoload
 172 (defun lao-compose-string (str)
 173   (with-category-table lao-category-table
 174    (let ((idx 0))
 175      (while (setq idx (string-match lao-composition-pattern str idx))
 176        (compose-string str idx (match-end 0))
 177        (setq idx (match-end 0))))
 178    str))
 179
 180 ;;; LRT: Lao <-> Roman Transcription
 181
 182 ;; Upper vowels and tone-marks are put on the letter.
 183 ;; Semi-vowel-sign-lo and lower vowels are put under the letter.
 184
 185 (defconst lao-transcription-consonant-alist
 186   (sort '(;; single consonants
 187           ("k" . "ກ")
 188           ("kh" . "ຂ")
 189           ("qh" . "ຄ")
 190           ("ng" . "ງ")
 191           ("j" . "ຈ")
 192           ("s" . "ສ")
 193           ("x" . "ຊ")
 194           ("y" . "ຍ")
 195           ("d" . "ດ")
 196           ("t" . "ຕ")
 197           ("th" . "ຖ")
 198           ("dh" . "ທ")
 199           ("n" . "ນ")
 200           ("b" . "ບ")
 201           ("p" . "ປ")
 202           ("hp" . "ຜ")
 203           ("fh" . "ຝ")
 204           ("ph" . "ພ")
 205           ("f" . "ຟ")
 206           ("m" . "ມ")
 207           ("gn" . "ຢ")
 208           ("l" . "ລ")
 209           ("r" . "ຣ")
 210           ("v" . "ວ")
 211           ("w" . "ວ")
 212           ("hh" . "ຫ")
 213           ("O" . "ອ")
 214           ("h" . "ຮ")
 215           ("nh" . "ໜ")
 216           ("mh" . "ໝ")
 217           ("lh" . ["ຫຼ"])
 218           ;; double consonants
 219           ("ngh" . ["ຫງ"])
 220           ("yh" . ["ຫຽ"])
 221           ("wh" . ["ຫວ"])
 222           ("hl" . ["ຫລ"])
 223           ("hy" . ["ຫຍ"])
 224           ("hn" . ["ຫນ"])
 225           ("hm" . ["ຫມ"])
 226           )
 227         (function (lambda (x y) (> (length (car x)) (length (car y)))))))
 228
 229 (defconst lao-transcription-semi-vowel-alist
 230   '(("r" . "ຼ")))
 231
 232 (defconst lao-transcription-vowel-alist
 233   (sort '(("a" . "ະ")
 234           ("ar" . "າ")
 235           ("i" . "ິ")
 236           ("ii" . "ີ")
 237           ("eu" . "ຶ")
 238           ("ur" . "ື")
 239           ("u" . "ຸ")
 240           ("uu" . "ູ")
 241           ("e" . ["ເະ"])
 242           ("ee" . "ເ")
 243           ("ae" . ["ແະ"])
 244           ("aa" . "ແ")
 245           ("o" . ["ໂະ"])
 246           ("oo" . "ໂ")
 247           ("oe" . ["ເາະ"])
 248           ("or" . "ໍ")
 249           ("er" . ["ເິ"])
 250           ("ir" . ["ເີ"])
 251           ("ua" . ["ົວະ"])
 252           ("uaa" . ["ົວ"])
 253           ("ie" . ["ເັຽະ"])
 254           ("ia" . ["ເັຽ"])
 255           ("ea" . ["ເຶອ"])
 256           ("eaa" . ["ເືອ"])
 257           ("ai" . "ໄ")
 258           ("ei" . "ໃ")
 259           ("ao" . ["ເົາ"])
 260           ("aM" . "ຳ"))
 261         (function (lambda (x y) (> (length (car x)) (length (car y)))))))
 262
 263 ;; Maa-sakod is put at the tail.
 264 (defconst lao-transcription-maa-sakod-alist
 265   '(("k" . "ກ")
 266     ("g" . "ງ")
 267     ("y" . "ຍ")
 268     ("d" . "ດ")
 269     ("n" . "ນ")
 270     ("b" . "ບ")
 271     ("m" . "ມ")
 272     ("v" . "ວ")
 273     ("w" . "ວ")
 274     ))
 275
 276 (defconst lao-transcription-tone-alist
 277   '(("'" . "່")
 278     ("\"" . "້")
 279     ("^" . "໊")
 280     ("+" . "໋")
 281     ("~" . "໌")))
 282
 283 (defconst lao-transcription-punctuation-alist
 284   '(("\\0" . "໐")
 285     ("\\1" . "໑")
 286     ("\\2" . "໒")
 287     ("\\3" . "໓")
 288     ("\\4" . "໔")
 289     ("\\5" . "໕")
 290     ("\\6" . "໖")
 291     ("\\7" . "໗")
 292     ("\\8" . "໘")
 293     ("\\9" . "໙")
 294     ("\\\\" . "ໆ")
 295     ("\\$" . "ຯ")))
 296
 297 (defconst lao-transcription-pattern
 298   (concat
 299    "\\("
 300    (mapconcat 'car lao-transcription-consonant-alist "\\|")
 301    "\\)\\("
 302    (mapconcat 'car lao-transcription-semi-vowel-alist "\\|")
 303    "\\)?\\(\\("
 304    (mapconcat 'car lao-transcription-vowel-alist "\\|")
 305    "\\)\\("
 306    (mapconcat 'car lao-transcription-maa-sakod-alist "\\|")
 307    "\\)?\\("
 308    (mapconcat (lambda (x) (regexp-quote (car x)))
 309               lao-transcription-tone-alist "\\|")
 310    "\\)?\\)?\\|"
 311    (mapconcat (lambda (x) (regexp-quote (car x)))
 312               lao-transcription-punctuation-alist "\\|")
 313    )
 314   "Regexp of Roman transcription pattern for one Lao syllable.")
 315
 316 (defconst lao-transcription-pattern
 317   (concat
 318    "\\("
 319    (regexp-opt (mapcar 'car lao-transcription-consonant-alist))
 320    "\\)\\("
 321    (regexp-opt (mapcar 'car lao-transcription-semi-vowel-alist))
 322    "\\)?\\(\\("
 323    (regexp-opt (mapcar 'car lao-transcription-vowel-alist))
 324    "\\)\\("
 325    (regexp-opt (mapcar 'car lao-transcription-maa-sakod-alist))
 326    "\\)?\\("
 327    (regexp-opt (mapcar 'car lao-transcription-tone-alist))
 328    "\\)?\\)?\\|"
 329    (regexp-opt (mapcar 'car lao-transcription-punctuation-alist))
 330    )
 331   "Regexp of Roman transcription pattern for one Lao syllable.")
 332
 333 (defconst lao-vowel-reordering-rule
 334   '(("ະ" (0 ?ະ) (0 ?ັ))
 335     ("າ" (0 ?າ))
 336     ("ິ" (0 ?ີ))
 337     ("ີ" (0 ?ີ))
 338     ("ຶ" (0 ?ຶ))
 339     ("ື" (0 ?ື))
 340     ("ຸ" (0 ?ຸ))
 341     ("ູ" (0 ?ູ))
 342     ("ເະ" (?ເ 0 ?ະ) (?ເ 0 ?ັ))
 343     ("ເ" (?ເ 0))
 344     ("ແະ" (?ແ 0 ?ະ) (?ແ 0 ?ັ))
 345     ("ແ" (?ແ 0))
 346     ("ໂະ" (?ໂ 0 ?ະ) (0 ?ົ) (?ຍ ?ໂ 0 ?ັ) (?ວ ?ໂ 0 ?ັ))
 347     ("ໂ" (?ໂ 0))
 348     ("ເາະ" (?ເ 0 ?າ ?ະ) (0 ?ັ ?ອ))
 349     ("ໍ" (0 ?ໍ) (0 ?ອ))
 350     ("ເິ" (?ເ 0 ?ິ))
 351     ("ເີ" (?ເ 0 ?ີ))
 352     ("ົວະ" (0 ?ົ ?ວ ?ະ) (0 ?ັ ?ວ))
 353     ("ົວ" (0 ?ົ ?ວ) (0 ?ວ))
 354     ("ເັຽະ" (?ເ 0 ?ັ ?ຽ ?ະ) (0 ?ັ ?ຽ))
 355     ("ເັຽ" (?ເ 0 ?ັ ?ຽ) (0 ?ຽ))
 356     ("ເຶອ" (?ເ 0 ?ຶ ?ອ))
 357     ("ເືອ" (?ເ 0 ?ື ?ອ))
 358     ("ໄ" (?ໄ 0))
 359     ("ໃ" (?ໃ 0))
 360     ("ເົາ" (?ເ 0 ?ົ ?າ))
 361     ("ຳ" (0 ?ຳ)))
 362   "Alist of Lao vowel string vs the corresponding re-ordering rule.
 363 Each element has this form:
 364         (VOWEL NO-MAA-SAKOD-RULE WITH-MAA-SAKOD-RULE (MAA-SAKOD-0 RULE-0) ...)
 365
 366 VOWEL is a vowel string (e.g. \"ເັຽະ\").
 367
 368 NO-MAA-SAKOD-RULE is a rule to re-order and modify VOWEL following a
 369 consonant.  It is a list vowel characters or 0.  The element 0
 370 indicate the place to embed a consonant.
 371
 372 Optional WITH-MAA-SAKOD-RULE is a rule to re-order and modify VOWEL
 373 following a consonant and preceding a maa-sakod character.  If it is
 374 nil, NO-MAA-SAKOD-RULE is used.  The maa-sakod character is always
 375 appended at the tail.
 376
 377 For instance, rule `(\"ເືອ\" (?ເ t ?ື ?ອ))' tells that this vowel
 378 string following a consonant `ກ' should be re-ordered as \"ເກືອ\".
 379
 380 Optional (MAA-SAKOD-n RULE-n) are rules specially applied to maa-sakod
 381 character MAA-SAKOD-n.")
 382
 383 ;;;###autoload
 384 (defun lao-transcribe-single-roman-syllable-to-lao (from to &optional str)
 385   "Transcribe a Romanized Lao syllable in the region FROM and TO to Lao string.
 386 Only the first syllable is transcribed.
 387 The value has the form: (START END LAO-STRING), where
 388 START and END are the beginning and end positions of the Roman Lao syllable,
 389 LAO-STRING is the Lao character transcription of it.
 390
 391 Optional 3rd arg STR, if non-nil, is a string to search for Roman Lao
 392 syllable.  In that case, FROM and TO are indexes to STR."
 393   (if str
 394       (if (setq from (string-match lao-transcription-pattern str from))
 395           (progn
 396             (if (>= from to)
 397                 (setq from nil)
 398               (setq to (match-end 0)))))
 399     (save-excursion
 400       (goto-char from)
 401       (if (setq to (re-search-forward lao-transcription-pattern to t))
 402           (setq from (match-beginning 0))
 403         (setq from nil))))
 404   (if from
 405       (let* ((consonant (match-string 1 str))
 406              (semivowel (match-string 3 str))
 407              (vowel (match-string 5 str))
 408              (maa-sakod (match-string 8 str))
 409              (tone (match-string 9 str))
 410              lao-consonant lao-semivowel lao-vowel lao-maa-sakod lao-tone
 411              clen cidx)
 412         (setq to (match-end 0))
 413         (if (not consonant)
 414             (setq str (cdr (assoc (match-string 0 str)
 415                                   lao-transcription-punctuation-alist)))
 416           (setq lao-consonant
 417                 (cdr (assoc consonant lao-transcription-consonant-alist)))
 418           (if (vectorp lao-consonant)
 419               (setq lao-consonant (aref lao-consonant 0)))
 420           (setq clen (length lao-consonant))
 421           (if semivowel
 422               ;; Include semivowel in STR.
 423               (setq lao-semivowel
 424                     (cdr (assoc semivowel lao-transcription-semi-vowel-alist))
 425                     str (if (= clen 1)
 426                             (concat lao-consonant lao-semivowel)
 427                           (concat (substring lao-consonant 0 1) lao-semivowel
 428                                   (substring lao-consonant 1))))
 429             (setq str lao-consonant))
 430           (if vowel
 431               (let (rule)
 432                 (setq lao-vowel
 433                       (cdr (assoc vowel lao-transcription-vowel-alist)))
 434                 (if (vectorp lao-vowel)
 435                     (setq lao-vowel (aref lao-vowel 0)))
 436                 (setq rule (assoc lao-vowel lao-vowel-reordering-rule))
 437                 (if (null maa-sakod)
 438                     (setq rule (nth 1 rule))
 439                   (setq lao-maa-sakod
 440                         (cdr (assoc maa-sakod lao-transcription-maa-sakod-alist))
 441                         rule
 442                         (or (cdr (assq (aref lao-maa-sakod 0) (nthcdr 2 rule)))
 443                             (nth 2 rule)
 444                             (nth 1 rule))))
 445                 (or rule
 446                     (error "Lao vowel %S has no re-ordering rule" lao-vowel))
 447                 (setq lao-consonant str str "")
 448                 (while rule
 449                   (if (= (car rule) 0)
 450                       (setq str (concat str lao-consonant)
 451                             cidx (length str))
 452                     (setq str (concat str (list (car rule)))))
 453                   (setq rule (cdr rule)))
 454                 (or cidx
 455                     (error "Lao vowel %S has malformed re-ordering rule" vowel))
 456                 ;; Set CIDX to after upper or lower vowel if any.
 457                 (let ((len (length str)))
 458                   (while (and (< cidx len)
 459                               (memq (get-char-code-property (aref str cidx)
 460                                                             'phonetic-type)
 461                                     '(vowel-lower vowel-upper)))
 462                     (setq cidx (1+ cidx))))
 463                 (if lao-maa-sakod
 464                     (setq str (concat str lao-maa-sakod)))
 465                 (if tone
 466                     (setq lao-tone
 467                           (cdr (assoc tone lao-transcription-tone-alist))
 468                           str (concat (substring str 0 cidx) lao-tone
 469                                       (substring str cidx)))))))
 470         (list from to (lao-compose-string str)))))
 471
 472 ;;;###autoload
 473 (defun lao-transcribe-roman-to-lao-string (str)
 474   "Transcribe Romanized Lao string STR to Lao character string."
 475   (let ((from 0)
 476         (to (length str))
 477         (lao-str "")
 478         val)
 479     (while (setq val (lao-transcribe-single-roman-syllable-to-lao from to str))
 480       (let ((start (car val))
 481             (end (nth 1 val))
 482             (lao (nth 2 val)))
 483         (if (> start from)
 484             (setq lao-str (concat lao-str (substring str from start) lao))
 485           (setq lao-str (concat lao-str lao)))
 486         (setq from end)))
 487     (if (< from to)
 488         (concat lao-str (substring str from to))
 489       lao-str)))
 490
 491 ;;;###autoload
 492 (defun lao-composition-function (gstring)
 493   (if (= (lgstring-char-len gstring) 1)
 494       (compose-gstring-for-graphic gstring)
 495     (or (font-shape-gstring gstring)
 496         (let ((glyph-len (lgstring-glyph-len gstring))
 497               (i 0)
 498               glyph)
 499           (while (and (< i glyph-len)
 500                       (setq glyph (lgstring-glyph gstring i)))
 501             (setq i (1+ i)))
 502           (compose-glyph-string-relative gstring 0 i 0.1)))))
 503
 504 ;;;###autoload
 505 (defun lao-compose-region (from to)
 506   (interactive "r")
 507   (save-restriction
 508     (narrow-to-region from to)
 509     (goto-char (point-min))
 510     (with-category-table lao-category-table
 511       (while (re-search-forward lao-composition-pattern nil t)
 512         (compose-region (match-beginning 0) (point))))))
 513
 514 ;;
 515 (provide 'lao-util)
 516
 517 ;;; lao-util.el ends here