lisp/language/thai-util.el

   1 ;;; thai-util.el --- utilities for Thai -*- coding: utf-8; -*-
   2
   3 ;; Copyright (C) 2000-2011  Free Software Foundation, Inc.
   4 ;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
   5 ;;   2005, 2006, 2007, 2008, 2009, 2010, 2011
   6 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
   7 ;;   Registration Number H14PRO021
   8
   9 ;; Keywords: mule, multilingual, Thai, i18n
  10
  11 ;; This file is part of GNU Emacs.
  12
  13 ;; GNU Emacs is free software: you can redistribute it and/or modify
  14 ;; it under the terms of the GNU General Public License as published by
  15 ;; the Free Software Foundation, either version 3 of the License, or
  16 ;; (at your option) any later version.
  17
  18 ;; GNU Emacs is distributed in the hope that it will be useful,
  19 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  20 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  21 ;; GNU General Public License for more details.
  22
  23 ;; You should have received a copy of the GNU General Public License
  24 ;; along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.
  25
  26 ;;; Commentary:
  27
  28 ;;; Code:
  29
  30 (defvar thai-auto-composition-mode)
  31
  32 ;; Setting information of Thai characters.
  33
  34 (defconst thai-category-table (make-category-table))
  35 (define-category ?c "Thai consonant" thai-category-table)
  36 (define-category ?v "Thai upper/lower vowel" thai-category-table)
  37 (define-category ?t "Thai tone mark" thai-category-table)
  38 (define-category ?u "Thai tone mark and upper sign" thai-category-table)
  39 (define-category ?I "THAI CHARACTER SARA I" thai-category-table)
  40 (define-category ?U "THAI CHARACTER THANTHAKHAT" thai-category-table)
  41
  42 ;; The general composing rules are as follows:
  43 ;;
  44 ;;                          T
  45 ;;       V        U         V                  U
  46 ;; CV -> C, CU -> C, CVT -> C, Cv -> C, CvU -> C
  47 ;;                                   v         v
  48 ;;
  49 ;; where C: consonant, V: vowel upper, v: vowel lower,
  50 ;;       T: tone mark, U: tone mark and upper sign.
  51 ;; Special rule: The sign `์' can be put on the vowel `ิ'.
  52
  53
  54 (defvar thai-composition-pattern
  55   "\\cc\\(\\cu\\|\\cI\\cU\\|\\cv\\ct?\\)\\|\\cv\\ct\\|\\cI\\cU"
  56   "Regular expression matching a Thai composite sequence.")
  57
  58 (let ((l '((?ก consonant "LETTER KO KAI")                             ; 0xA1
  59            (?ข consonant "LETTER KHO KHAI")                           ; 0xA2
  60            (?ฃ consonant "LETTER KHO KHUAT")                          ; 0xA3
  61            (?ค consonant "LETTER KHO KHWAI")                          ; 0xA4
  62            (?ฅ consonant "LETTER KHO KHON")                           ; 0xA5
  63            (?ฆ consonant "LETTER KHO RAKHANG")                                ; 0xA6
  64            (?ง consonant "LETTER NGO NGU")                            ; 0xA7
  65            (?จ consonant "LETTER CHO CHAN")                           ; 0xA8
  66            (?ฉ consonant "LETTER CHO CHING")                          ; 0xA9
  67            (?ช consonant "LETTER CHO CHANG")                          ; 0xAA
  68            (?ซ consonant "LETTER SO SO")                              ; 0xAB
  69            (?ฌ consonant "LETTER CHO CHOE")                           ; 0xAC
  70            (?ญ consonant "LETTER YO YING")                            ; 0xAD
  71            (?ฎ consonant "LETTER DO CHADA")                           ; 0xAE
  72            (?ฏ consonant "LETTER TO PATAK")                           ; 0xAF
  73            (?ฐ consonant "LETTER THO THAN")                           ; 0xB0
  74            (?ฑ consonant "LETTER THO NANGMONTHO")                     ; 0xB1
  75            (?ฒ consonant "LETTER THO PHUTHAO")                                ; 0xB2
  76            (?ณ consonant "LETTER NO NEN")                             ; 0xB3
  77            (?ด consonant "LETTER DO DEK")                             ; 0xB4
  78            (?ต consonant "LETTER TO TAO")                             ; 0xB5
  79            (?ถ consonant "LETTER THO THUNG")                          ; 0xB6
  80            (?ท consonant "LETTER THO THAHAN")                         ; 0xB7
  81            (?ธ consonant "LETTER THO THONG")                          ; 0xB8
  82            (?น consonant "LETTER NO NU")                              ; 0xB9
  83            (?บ consonant "LETTER BO BAIMAI")                          ; 0xBA
  84            (?ป consonant "LETTER PO PLA")                             ; 0xBB
  85            (?ผ consonant "LETTER PHO PHUNG")                          ; 0xBC
  86            (?ฝ consonant "LETTER FO FA")                              ; 0xBD
  87            (?พ consonant "LETTER PHO PHAN")                           ; 0xBE
  88            (?ฟ consonant "LETTER FO FAN")                             ; 0xBF
  89            (?ภ consonant "LETTER PHO SAMPHAO")                                ; 0xC0
  90            (?ม consonant "LETTER MO MA")                              ; 0xC1
  91            (?ย consonant "LETTER YO YAK")                             ; 0xC2
  92            (?ร consonant "LETTER RO RUA")                             ; 0xC3
  93            (?ฤ vowel-base "LETTER RU (Pali vowel letter)")            ; 0xC4
  94            (?ล consonant "LETTER LO LING")                            ; 0xC5
  95            (?ฦ vowel-base "LETTER LU (Pali vowel letter)")            ; 0xC6
  96            (?ว consonant "LETTER WO WAEN")                            ; 0xC7
  97            (?ศ consonant "LETTER SO SALA")                            ; 0xC8
  98            (?ษ consonant "LETTER SO RUSI")                            ; 0xC9
  99            (?ส consonant "LETTER SO SUA")                             ; 0xCA
 100            (?ห consonant "LETTER HO HIP")                             ; 0xCB
 101            (?ฬ consonant "LETTER LO CHULA")                           ; 0xCC
 102            (?อ consonant "LETTER O ANG")                              ; 0xCD
 103            (?ฮ consonant "LETTER HO NOK HUK")                         ; 0xCE
 104            (?ฯ special "PAI YAN NOI (abbreviation)")                  ; 0xCF
 105            (?ะ vowel-base "VOWEL SIGN SARA A")                                ; 0xD0
 106            (?ั vowel-upper "VOWEL SIGN MAI HAN-AKAT N/S-T")           ; 0xD1
 107            (?า vowel-base "VOWEL SIGN SARA AA")                               ; 0xD2
 108            (?ำ vowel-base "VOWEL SIGN SARA AM")                               ; 0xD3
 109            (?ิ vowel-upper "VOWEL SIGN SARA I N/S-T")                 ; 0xD4
 110            (?ี vowel-upper "VOWEL SIGN SARA II N/S-T")                        ; 0xD5
 111            (?ึ vowel-upper "VOWEL SIGN SARA UE N/S-T")                        ; 0xD6
 112            (?ื vowel-upper "VOWEL SIGN SARA UEE N/S-T")                       ; 0xD7
 113            (?ุ vowel-lower "VOWEL SIGN SARA U N/S-B")                 ; 0xD8
 114            (?ู vowel-lower "VOWEL SIGN SARA UU N/S-B")                        ; 0xD9
 115            (?ฺ vowel-lower "VOWEL SIGN PHINTHU N/S-B (Pali virama)")  ; 0xDA
 116            (?฻ invalid nil)                                           ; 0xDA
 117            (?฼ invalid nil)                                           ; 0xDC
 118            (?฽ invalid nil)                                           ; 0xDC
 119            (?฾ invalid nil)                                           ; 0xDC
 120            (?฿ special "BAHT SIGN (currency symbol)")                 ; 0xDF
 121            (?เ vowel-base "VOWEL SIGN SARA E")                                ; 0xE0
 122            (?แ vowel-base "VOWEL SIGN SARA AE")                               ; 0xE1
 123            (?โ vowel-base "VOWEL SIGN SARA O")                                ; 0xE2
 124            (?ใ vowel-base "VOWEL SIGN SARA MAI MUAN")                 ; 0xE3
 125            (?ไ vowel-base "VOWEL SIGN SARA MAI MALAI")                        ; 0xE4
 126            (?ๅ vowel-base "LAK KHANG YAO")                            ; 0xE5
 127            (?ๆ special "MAI YAMOK (repetition)")                      ; 0xE6
 128            (?็ sign-upper "VOWEL SIGN MAI TAI KHU N/S-T")             ; 0xE7
 129            (?่ tone "TONE MAI EK N/S-T")                              ; 0xE8
 130            (?้ tone "TONE MAI THO N/S-T")                             ; 0xE9
 131            (?๊ tone "TONE MAI TRI N/S-T")                             ; 0xEA
 132            (?๋ tone "TONE MAI CHATTAWA N/S-T")                                ; 0xEB
 133            (?์ sign-upper "THANTHAKHAT N/S-T (cancellation mark)")    ; 0xEC
 134            (?ํ sign-upper "NIKKHAHIT N/S-T (final nasal)")            ; 0xED
 135            (?๎ sign-upper "YAMAKKAN N/S-T")                           ; 0xEE
 136            (?๏ special "FONRMAN")                                     ; 0xEF
 137            (?๐ special "DIGIT ZERO")                                  ; 0xF0
 138            (?๑ special "DIGIT ONE")                                   ; 0xF1
 139            (?๒ special "DIGIT TWO")                                   ; 0xF2
 140            (?๓ special "DIGIT THREE")                                 ; 0xF3
 141            (?๔ special "DIGIT FOUR")                                  ; 0xF4
 142            (?๕ special "DIGIT FIVE")                                  ; 0xF5
 143            (?๖ special "DIGIT SIX")                                   ; 0xF6
 144            (?๗ special "DIGIT SEVEN")                                 ; 0xF7
 145            (?๘ special "DIGIT EIGHT")                                 ; 0xF8
 146            (?๙ special "DIGIT NINE")                                  ; 0xF9
 147            (?๚ special "ANGKHANKHU (ellipsis)")                               ; 0xFA
 148            (?๛ special "KHOMUT (beginning of religious texts)")               ; 0xFB
 149            (?๜ invalid nil)                                           ; 0xFC
 150            (?๝ invalid nil)                                           ; 0xFD
 151            (?๞ invalid nil)                                           ; 0xFE
 152            ))
 153       elm)
 154   (while l
 155     (setq elm (car l) l (cdr l))
 156     (let ((char (car elm))
 157           (ptype (nth 1 elm)))
 158       (put-char-code-property char 'phonetic-type ptype)
 159       (cond ((eq ptype 'consonant)
 160              (modify-category-entry char ?c thai-category-table))
 161             ((memq ptype '(vowel-upper vowel-lower))
 162              (modify-category-entry char ?v thai-category-table)
 163              (if (= char ?ิ)
 164                  ;; Give category `I' to "SARA I".
 165                  (modify-category-entry char ?I thai-category-table)))
 166             ((eq ptype 'tone)
 167              (modify-category-entry char ?t thai-category-table)
 168              (modify-category-entry char ?u thai-category-table))
 169             ((eq ptype 'sign-upper)
 170              (modify-category-entry char ?u thai-category-table)
 171              (if (= char ?์)
 172                  ;; Give category `U' to "THANTHAKHAT".
 173                  (modify-category-entry char ?U thai-category-table))))
 174       (put-char-code-property char 'name (nth 2 elm)))))
 175
 176 (defun thai-compose-syllable (beg end &optional category-set string)
 177   (or category-set
 178       (setq category-set
 179             (char-category-set (if string (aref string beg) (char-after beg)))))
 180   (if (aref category-set ?c)
 181       ;; Starting with a consonant.  We do relative composition.
 182       (if string
 183           (compose-string string beg end)
 184         (compose-region beg end))
 185     ;; Vowel tone sequence.
 186     (if string
 187         (compose-string string beg end (list (aref string beg) '(Bc . Bc)
 188                                              (aref string (1+ beg))))
 189       (compose-region beg end (list (char-after beg) '(Bc . Bc)
 190                                     (char-after (1+ beg))))))
 191   (- end beg))
 192
 193 ;;;###autoload
 194 (defun thai-compose-region (beg end)
 195   "Compose Thai characters in the region.
 196 When called from a program, expects two arguments,
 197 positions (integers or markers) specifying the region."
 198   (interactive "r")
 199   (let ((pos (point)))
 200     (save-restriction
 201       (narrow-to-region beg end)
 202       (goto-char (point-min))
 203       (with-category-table thai-category-table
 204         (while (re-search-forward thai-composition-pattern nil t)
 205           (setq beg (match-beginning 0) end (match-end 0))
 206           (if (and (> pos beg) (< pos end))
 207               (setq pos end))
 208           (thai-compose-syllable beg end
 209                                  (char-category-set (char-after beg))))))
 210     (goto-char pos)))
 211
 212 ;;;###autoload
 213 (defun thai-compose-string (string)
 214   "Compose Thai characters in STRING and return the resulting string."
 215   (with-category-table thai-category-table
 216     (let ((idx 0))
 217       (while (setq idx (string-match thai-composition-pattern string idx))
 218         (thai-compose-syllable idx (match-end 0) nil string)
 219         (setq idx (match-end 0)))))
 220   string)
 221
 222 ;;;###autoload
 223 (defun thai-compose-buffer ()
 224   "Compose Thai characters in the current buffer."
 225   (interactive)
 226   (thai-compose-region (point-min) (point-max)))
 227
 228 ;;;###autoload
 229 (defun thai-composition-function (gstring)
 230   (if (= (lgstring-char-len gstring) 1)
 231       (compose-gstring-for-graphic gstring)
 232     (or (font-shape-gstring gstring)
 233         (let ((glyph-len (lgstring-glyph-len gstring))
 234               (last-char (lgstring-char gstring
 235                                         (1- (lgstring-char-len gstring))))
 236               (i 0)
 237               glyph)
 238           (while (and (< i glyph-len)
 239                       (setq glyph (lgstring-glyph gstring i)))
 240             (setq i (1+ i)))
 241           (if (= last-char ?ำ)
 242               (setq i (1- i)))
 243           (compose-glyph-string-relative gstring 0 i 0.1)))))
 244
 245 ;; Thai-word-mode requires functions in the feature `thai-word'.
 246 (require 'thai-word)
 247
 248 (defvar thai-word-mode-map
 249   (let ((map (make-sparse-keymap)))
 250     (define-key map [remap forward-word] 'thai-forward-word)
 251     (define-key map [remap backward-word] 'thai-backward-word)
 252     (define-key map [remap kill-word] 'thai-kill-word)
 253     (define-key map [remap backward-kill-word] 'thai-backward-kill-word)
 254     (define-key map [remap transpose-words] 'thai-transpose-words)
 255     map)
 256   "Keymap for `thai-word-mode'.")
 257
 258 (define-minor-mode thai-word-mode
 259   "Minor mode to make word-oriented commands aware of Thai words.
 260 The commands affected are \\[forward-word], \\[backward-word], \\[kill-word], \\[backward-kill-word], \\[transpose-words], and \\[fill-paragraph]."
 261   :global t :group 'mule
 262   (cond (thai-word-mode
 263          ;; This enables linebreak between Thai characters.
 264          (modify-category-entry (make-char 'thai-tis620) ?|)
 265          ;; This enables linebreak at a Thai word boundary.
 266          (put-charset-property 'thai-tis620 'fill-find-break-point-function
 267                                'thai-fill-find-break-point))
 268         (t
 269          (modify-category-entry (make-char 'thai-tis620) ?| nil t)
 270          (put-charset-property 'thai-tis620 'fill-find-break-point-function
 271                                nil))))
 272
 273 ;; Function to call on entering the Thai language environment.
 274 (defun setup-thai-language-environment-internal ()
 275   (thai-word-mode 1))
 276
 277 ;; Function to call on exiting the Thai language environment.
 278 (defun exit-thai-language-environment-internal ()
 279   (thai-word-mode -1))
 280
 281 ;;
 282 (provide 'thai-util)
 283
 284 ;;; thai-util.el ends here