lisp/language/viet-util.el

   1 ;;; viet-util.el --- utilities for Vietnamese  -*- coding: iso-2022-7bit; -*-
   2
   3 ;; Copyright (C) 1998, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010
   4 ;;   Free Software Foundation, Inc.
   5 ;; Copyright (C) 1995, 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004,
   6 ;;   2005, 2006, 2007, 2008, 2009, 2010
   7 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
   8 ;;   Registration Number H14PRO021
   9 ;; Copyright (C) 2003
  10 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
  11 ;;   Registration Number H13PRO009
  12
  13 ;; Keywords: mule, multilingual, Vietnamese
  14
  15 ;; This file is part of GNU Emacs.
  16
  17 ;; GNU Emacs is free software: you can redistribute it and/or modify
  18 ;; it under the terms of the GNU General Public License as published by
  19 ;; the Free Software Foundation, either version 3 of the License, or
  20 ;; (at your option) any later version.
  21
  22 ;; GNU Emacs is distributed in the hope that it will be useful,
  23 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  24 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  25 ;; GNU General Public License for more details.
  26
  27 ;; You should have received a copy of the GNU General Public License
  28 ;; along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.
  29
  30 ;;; Commentary:
  31
  32 ;; Vietnamese uses ASCII characters and additional 134 unique
  33 ;; characters (these are Latin alphabets with various diacritical and
  34 ;; tone marks).  As far as I know, Vietnamese now has 5 different ways
  35 ;; for representing these characters: VISCII, TCVN-5712, VPS, VIQR,
  36 ;; and Unicode.  VISCII, TCVN-5712 and VPS are simple 1-byte code
  37 ;; which assigns 134 unique characters in control-code area
  38 ;; (0x00..0x1F) and right half area (0x80..0xFF).  VIQR is a menmonic
  39 ;; encoding specification representing diacritical marks by following
  40 ;; ASCII characters.
  41
  42 ;;; Code:
  43
  44 (defvar viet-viscii-nonascii-translation-table)
  45
  46 ;;;###autoload
  47 (defun viet-encode-viscii-char (char)
  48   "Return VISCII character code of CHAR if appropriate."
  49   (encode-char char 'viscii))
  50
  51 ;; VIQR is a menmonic encoding specification for Vietnamese.
  52 ;; It represents diacritical marks by ASCII characters as follows:
  53 ;; ------------+----------+--------
  54 ;;     mark    | mnemonic | example
  55 ;; ------------+----------+---------
  56 ;;    breve    |    (     | a( -> \e,1e\e(B
  57 ;;  circumflex |    ^     | a^ -> \e,1b\e(B
  58 ;;    horn     |    +     | o+ -> \e,1=\e(B
  59 ;; ------------+----------+---------
  60 ;;    acute    |    '     | a' -> \e,1a\e(B
  61 ;;    grave    |    `     | a` -> \e,1`\e(B
  62 ;;  hook above |    ?     | a? -> \e,1d\e(B
  63 ;;    tilde    |    ~     | a~ -> \e,1c\e(B
  64 ;;   dot below |    .     | a. -> \e,1U\e(B
  65 ;; ------------+----------+---------
  66 ;;    d bar    |   dd     | dd -> \e,1p\e(B
  67 ;; ------------+----------+---------
  68
  69 (defvar viet-viqr-alist
  70   '(;; lowercase
  71     (?\e,1!\e(B . "a('")                  ; 161
  72     (?\e,1"\e(B . "a(`")                  ; 162
  73     (?\e,1#\e(B . "a(.")                  ; 163
  74     (?\e,1$\e(B . "a^'")                  ; 164
  75     (?\e,1%\e(B . "a^`")                  ; 165
  76     (?\e,1&\e(B . "a^?")                  ; 166
  77     (?\e,1'\e(B . "a^.")                  ; 167
  78     (?\e,1(\e(B . "e~")                           ; 168
  79     (?\e,1)\e(B . "e.")                           ; 169
  80     (?\e,1*\e(B . "e^'")                  ; 170
  81     (?\e,1+\e(B . "e^`")                  ; 171
  82     (?\e,1,\e(B . "e^?")                  ; 172
  83     (?\e,1-\e(B . "e^~")                  ; 173
  84     (?\e,1.\e(B . "e^.")                  ; 174
  85     (?\e,1/\e(B . "o^'")                  ; 175
  86     (?\e,10\e(B . "o^`")                  ; 176
  87     (?\e,11\e(B . "o^?")                  ; 177
  88     (?\e,12\e(B . "o^~")                  ; 178
  89     (?\e,15\e(B . "o^.")                  ; 181
  90     (?\e,16\e(B . "o+`")                  ; 182
  91     (?\e,17\e(B . "o+?")                  ; 183
  92     (?\e,18\e(B . "i.")                           ; 184
  93     (?\e,1=\e(B . "o+")                           ; 189
  94     (?\e,1>\e(B . "o+'")                  ; 190
  95     (?\e,1F\e(B . "a(?")                  ; 198
  96     (?\e,1G\e(B . "a(~")                  ; 199
  97     (?\e,1O\e(B . "y`")                           ; 207
  98     (?\e,1Q\e(B . "u+'")                  ; 209
  99     (?\e,1U\e(B . "a.")                           ; 213
 100     (?\e,1V\e(B . "y?")                           ; 214
 101     (?\e,1W\e(B . "u+`")                  ; 215
 102     (?\e,1X\e(B . "u+?")                  ; 216
 103     (?\e,1[\e(B . "y~")                           ; 219
 104     (?\e,1\\e(B . "y.")                           ; 220
 105     (?\e,1^\e(B . "o+~")                  ; 222
 106     (?\e,1_\e(B . "u+")                           ; 223
 107     (?\e,1`\e(B . "a`")                           ; 224
 108     (?\e,1a\e(B . "a'")                           ; 225
 109     (?\e,1b\e(B . "a^")                           ; 226
 110     (?\e,1c\e(B . "a~")                           ; 227
 111     (?\e,1d\e(B . "a?")                           ; 228
 112     (?\e,1e\e(B . "a(")                           ; 229
 113     (?\e,1f\e(B . "u+~")                  ; 230
 114     (?\e,1g\e(B . "a^~")                  ; 231
 115     (?\e,1h\e(B . "e`")                           ; 232
 116     (?\e,1i\e(B . "e'")                           ; 233
 117     (?\e,1j\e(B . "e^")                           ; 234
 118     (?\e,1k\e(B . "e?")                           ; 235
 119     (?\e,1l\e(B . "i`")                           ; 236
 120     (?\e,1m\e(B . "i'")                           ; 237
 121     (?\e,1n\e(B . "i~")                           ; 238
 122     (?\e,1o\e(B . "i?")                           ; 239
 123     (?\e,1p\e(B . "dd")                           ; 240
 124     (?\e,1q\e(B . "u+.")                  ; 241
 125     (?\e,1r\e(B . "o`")                           ; 242
 126     (?\e,1s\e(B . "o'")                           ; 243
 127     (?\e,1t\e(B . "o^")                           ; 244
 128     (?\e,1u\e(B . "o~")                           ; 245
 129     (?\e,1v\e(B . "o?")                           ; 246
 130     (?\e,1w\e(B . "o.")                           ; 247
 131     (?\e,1x\e(B . "u.")                           ; 248
 132     (?\e,1y\e(B . "u`")                           ; 249
 133     (?\e,1z\e(B . "u'")                           ; 250
 134     (?\e,1{\e(B . "u~")                           ; 251
 135     (?\e,1|\e(B . "u?")                           ; 252
 136     (?\e,1}\e(B . "y'")                           ; 253
 137     (?\e,1~\e(B . "o+.")                  ; 254
 138
 139     ;; upper case
 140     (?\e,2!\e(B . "A('")                  ; 161
 141     (?\e,2"\e(B . "A(`")                  ; 162
 142     (?\e,2#\e(B . "A(.")                  ; 163
 143     (?\e,2$\e(B . "A^'")                  ; 164
 144     (?\e,2%\e(B . "A^`")                  ; 165
 145     (?\e,2&\e(B . "A^?")                  ; 166
 146     (?\e,2'\e(B . "A^.")                  ; 167
 147     (?\e,2(\e(B . "E~")                           ; 168
 148     (?\e,2)\e(B . "E.")                           ; 169
 149     (?\e,2*\e(B . "E^'")                  ; 170
 150     (?\e,2+\e(B . "E^`")                  ; 171
 151     (?\e,2,\e(B . "E^?")                  ; 172
 152     (?\e,2-\e(B . "E^~")                  ; 173
 153     (?\e,2.\e(B . "E^.")                  ; 174
 154     (?\e,2/\e(B . "O^'")                  ; 175
 155     (?\e,20\e(B . "O^`")                  ; 176
 156     (?\e,21\e(B . "O^?")                  ; 177
 157     (?\e,22\e(B . "O^~")                  ; 178
 158     (?\e,25\e(B . "O^.")                  ; 181
 159     (?\e,26\e(B . "O+`")                  ; 182
 160     (?\e,27\e(B . "O+?")                  ; 183
 161     (?\e,28\e(B . "I.")                           ; 184
 162     (?\e,2=\e(B . "O+")                           ; 189
 163     (?\e,2>\e(B . "O+'")                  ; 190
 164     (?\e,2F\e(B . "A(?")                  ; 198
 165     (?\e,2G\e(B . "A(~")                  ; 199
 166     (?\e,2O\e(B . "Y`")                           ; 207
 167     (?\e,2Q\e(B . "U+'")                  ; 209
 168     (?\e,2U\e(B . "A.")                           ; 213
 169     (?\e,2V\e(B . "Y?")                           ; 214
 170     (?\e,2W\e(B . "U+`")                  ; 215
 171     (?\e,2X\e(B . "U+?")                  ; 216
 172     (?\e,2[\e(B . "Y~")                           ; 219
 173     (?\e,2\\e(B . "Y.")                           ; 220
 174     (?\e,2^\e(B . "O+~")                  ; 222
 175     (?\e,2_\e(B . "U+")                           ; 223
 176     (?\e,2`\e(B . "A`")                           ; 224
 177     (?\e,2a\e(B . "A'")                           ; 225
 178     (?\e,2b\e(B . "A^")                           ; 226
 179     (?\e,2c\e(B . "A~")                           ; 227
 180     (?\e,2d\e(B . "A?")                           ; 228
 181     (?\e,2e\e(B . "A(")                           ; 229
 182     (?\e,2f\e(B . "U+~")                  ; 230
 183     (?\e,2g\e(B . "A^~")                  ; 231
 184     (?\e,2h\e(B . "E`")                           ; 232
 185     (?\e,2i\e(B . "E'")                           ; 233
 186     (?\e,2j\e(B . "E^")                           ; 234
 187     (?\e,2k\e(B . "E?")                           ; 235
 188     (?\e,2l\e(B . "I`")                           ; 236
 189     (?\e,2m\e(B . "I'")                           ; 237
 190     (?\e,2n\e(B . "I~")                           ; 238
 191     (?\e,2o\e(B . "I?")                           ; 239
 192     (?\e,2p\e(B . "DD")                           ; 240
 193     (?\e,2p\e(B . "dD")                           ; 240
 194     (?\e,2p\e(B . "Dd")                           ; 240
 195     (?\e,2q\e(B . "U+.")                  ; 241
 196     (?\e,2r\e(B . "O`")                           ; 242
 197     (?\e,2s\e(B . "O'")                           ; 243
 198     (?\e,2t\e(B . "O^")                           ; 244
 199     (?\e,2u\e(B . "O~")                           ; 245
 200     (?\e,2v\e(B . "O?")                           ; 246
 201     (?\e,2w\e(B . "O.")                           ; 247
 202     (?\e,2x\e(B . "U.")                           ; 248
 203     (?\e,2y\e(B . "U`")                           ; 249
 204     (?\e,2z\e(B . "U'")                           ; 250
 205     (?\e,2{\e(B . "U~")                           ; 251
 206     (?\e,2|\e(B . "U?")                           ; 252
 207     (?\e,2}\e(B . "Y'")                           ; 253
 208     (?\e,2~\e(B . "O+.")                  ; 254
 209
 210     ;; escape from composition
 211     (?\( . "\\(")                       ; breve (left parenthesis)
 212     (?^ . "\\^")                        ; circumflex (caret)
 213     (?+ . "\\+")                        ; horn (plus sign)
 214     (?' . "\\'")                        ; acute (apostrophe)
 215     (?` . "\\`")                        ; grave (backquote)
 216     (?? . "\\?")                        ; hook above (question mark)
 217     (?~ . "\\~")                        ; tilde (tilde)
 218     (?. . "\\.")                        ; dot below (period)
 219     (?d . "\\d")                        ; d-bar (d)
 220     (?\\ . "\\\\")                      ; literal backslash
 221     )
 222   "Alist of Vietnamese characters vs corresponding `VIQR' string.")
 223
 224 ;; Regular expression matching single Vietnamese character represented
 225 ;; by VIQR.
 226 (defconst viqr-regexp
 227   "[aeiouyAEIOUY]\\([(^+]?['`?~.]\\|[(^+]\\)\\|[Dd][Dd]")
 228
 229 ;;;###autoload
 230 (defun viet-decode-viqr-region (from to)
 231   "Convert `VIQR' mnemonics of the current region to Vietnamese characters.
 232 When called from a program, expects two arguments,
 233 positions (integers or markers) specifying the stretch of the region."
 234   (interactive "r")
 235   (save-restriction
 236     (narrow-to-region from to)
 237     (goto-char (point-min))
 238     (while (re-search-forward viqr-regexp nil t)
 239       (let* ((viqr (buffer-substring (match-beginning 0) (match-end 0)))
 240              (ch (car (rassoc viqr viet-viqr-alist))))
 241         (if ch
 242             (progn
 243               (delete-region (match-beginning 0) (match-end 0))
 244               (insert ch)))))))
 245
 246 ;;;###autoload
 247 (defun viet-decode-viqr-buffer ()
 248   "Convert `VIQR' mnemonics of the current buffer to Vietnamese characters."
 249   (interactive)
 250   (viet-decode-viqr-region (point-min) (point-max)))
 251
 252 ;;;###autoload
 253 (defun viet-encode-viqr-region (from to)
 254   "Convert Vietnamese characters of the current region to `VIQR' mnemonics.
 255 When called from a program, expects two arguments,
 256 positions (integers or markers) specifying the stretch of the region."
 257   (interactive "r")
 258   (save-restriction
 259     (narrow-to-region from to)
 260     (goto-char (point-min))
 261     (while (re-search-forward "\\cv" nil t)
 262       (let* ((ch (preceding-char))
 263              (viqr (cdr (assq ch viet-viqr-alist))))
 264         (if viqr
 265             (progn
 266               (delete-char -1)
 267               (insert viqr)))))))
 268
 269 ;;;###autoload
 270 (defun viet-encode-viqr-buffer ()
 271   "Convert Vietnamese characters of the current buffer to `VIQR' mnemonics."
 272   (interactive)
 273   (viet-encode-viqr-region (point-min) (point-max)))
 274
 275 ;;;###autoload
 276 (defun viqr-post-read-conversion (len)
 277   (save-excursion
 278     (save-restriction
 279       (narrow-to-region (point) (+ (point) len))
 280       (let ((buffer-modified-p (buffer-modified-p)))
 281         (viet-decode-viqr-region (point-min) (point-max))
 282         (set-buffer-modified-p buffer-modified-p)
 283         (- (point-max) (point-min))))))
 284
 285 ;;;###autoload
 286 (defun viqr-pre-write-conversion (from to)
 287   (let ((old-buf (current-buffer)))
 288     (set-buffer (generate-new-buffer " *temp*"))
 289     (if (stringp from)
 290         (insert from)
 291       (insert-buffer-substring old-buf from to))
 292     (viet-encode-viqr-region (point-min) (point-max))
 293     ;; Should return nil as annotations.
 294     nil))
 295
 296 ;;;
 297 (provide 'viet-util)
 298
 299 ;; arch-tag: 082a4d3b-168f-45b4-b3e1-82bfa1b5a194
 300 ;;; viet-util.el ends here