lisp/language/viet-util.el

   1 ;;; viet-util.el --- utilities for Vietnamese  -*- coding: iso-2022-7bit; -*-
   2
   3 ;; Copyright (C) 1998, 2001-2012  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 ;; Copyright (C) 2003
   9 ;;   National Institute of Advanced Industrial Science and Technology (AIST)
  10 ;;   Registration Number H13PRO009
  11
  12 ;; Keywords: mule, multilingual, Vietnamese
  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 ;; Vietnamese uses ASCII characters and additional 134 unique
  32 ;; characters (these are Latin alphabets with various diacritical and
  33 ;; tone marks).  As far as I know, Vietnamese now has 5 different ways
  34 ;; for representing these characters: VISCII, TCVN-5712, VPS, VIQR,
  35 ;; and Unicode.  VISCII, TCVN-5712 and VPS are simple 1-byte code
  36 ;; which assigns 134 unique characters in control-code area
  37 ;; (0x00..0x1F) and right half area (0x80..0xFF).  VIQR is a mnemonic
  38 ;; encoding specification representing diacritical marks by following
  39 ;; ASCII characters.
  40
  41 ;;; Code:
  42
  43 (defvar viet-viscii-nonascii-translation-table)
  44
  45 ;;;###autoload
  46 (defun viet-encode-viscii-char (char)
  47   "Return VISCII character code of CHAR if appropriate."
  48   (encode-char char 'viscii))
  49
  50 ;; VIQR is a mnemonic encoding specification for Vietnamese.
  51 ;; It represents diacritical marks by ASCII characters as follows:
  52 ;; ------------+----------+--------
  53 ;;     mark    | mnemonic | example
  54 ;; ------------+----------+---------
  55 ;;    breve    |    (     | a( -> \e,1e\e(B
  56 ;;  circumflex |    ^     | a^ -> \e,1b\e(B
  57 ;;    horn     |    +     | o+ -> \e,1=\e(B
  58 ;; ------------+----------+---------
  59 ;;    acute    |    '     | a' -> \e,1a\e(B
  60 ;;    grave    |    `     | a` -> \e,1`\e(B
  61 ;;  hook above |    ?     | a? -> \e,1d\e(B
  62 ;;    tilde    |    ~     | a~ -> \e,1c\e(B
  63 ;;   dot below |    .     | a. -> \e,1U\e(B
  64 ;; ------------+----------+---------
  65 ;;    d bar    |   dd     | dd -> \e,1p\e(B
  66 ;; ------------+----------+---------
  67
  68 (defvar viet-viqr-alist
  69   '(;; lowercase
  70     (?\e,1!\e(B . "a('")                  ; 161
  71     (?\e,1"\e(B . "a(`")                  ; 162
  72     (?\e,1#\e(B . "a(.")                  ; 163
  73     (?\e,1$\e(B . "a^'")                  ; 164
  74     (?\e,1%\e(B . "a^`")                  ; 165
  75     (?\e,1&\e(B . "a^?")                  ; 166
  76     (?\e,1'\e(B . "a^.")                  ; 167
  77     (?\e,1(\e(B . "e~")                           ; 168
  78     (?\e,1)\e(B . "e.")                           ; 169
  79     (?\e,1*\e(B . "e^'")                  ; 170
  80     (?\e,1+\e(B . "e^`")                  ; 171
  81     (?\e,1,\e(B . "e^?")                  ; 172
  82     (?\e,1-\e(B . "e^~")                  ; 173
  83     (?\e,1.\e(B . "e^.")                  ; 174
  84     (?\e,1/\e(B . "o^'")                  ; 175
  85     (?\e,10\e(B . "o^`")                  ; 176
  86     (?\e,11\e(B . "o^?")                  ; 177
  87     (?\e,12\e(B . "o^~")                  ; 178
  88     (?\e,15\e(B . "o^.")                  ; 181
  89     (?\e,16\e(B . "o+`")                  ; 182
  90     (?\e,17\e(B . "o+?")                  ; 183
  91     (?\e,18\e(B . "i.")                           ; 184
  92     (?\e,1=\e(B . "o+")                           ; 189
  93     (?\e,1>\e(B . "o+'")                  ; 190
  94     (?\e,1F\e(B . "a(?")                  ; 198
  95     (?\e,1G\e(B . "a(~")                  ; 199
  96     (?\e,1O\e(B . "y`")                           ; 207
  97     (?\e,1Q\e(B . "u+'")                  ; 209
  98     (?\e,1U\e(B . "a.")                           ; 213
  99     (?\e,1V\e(B . "y?")                           ; 214
 100     (?\e,1W\e(B . "u+`")                  ; 215
 101     (?\e,1X\e(B . "u+?")                  ; 216
 102     (?\e,1[\e(B . "y~")                           ; 219
 103     (?\e,1\\e(B . "y.")                           ; 220
 104     (?\e,1^\e(B . "o+~")                  ; 222
 105     (?\e,1_\e(B . "u+")                           ; 223
 106     (?\e,1`\e(B . "a`")                           ; 224
 107     (?\e,1a\e(B . "a'")                           ; 225
 108     (?\e,1b\e(B . "a^")                           ; 226
 109     (?\e,1c\e(B . "a~")                           ; 227
 110     (?\e,1d\e(B . "a?")                           ; 228
 111     (?\e,1e\e(B . "a(")                           ; 229
 112     (?\e,1f\e(B . "u+~")                  ; 230
 113     (?\e,1g\e(B . "a^~")                  ; 231
 114     (?\e,1h\e(B . "e`")                           ; 232
 115     (?\e,1i\e(B . "e'")                           ; 233
 116     (?\e,1j\e(B . "e^")                           ; 234
 117     (?\e,1k\e(B . "e?")                           ; 235
 118     (?\e,1l\e(B . "i`")                           ; 236
 119     (?\e,1m\e(B . "i'")                           ; 237
 120     (?\e,1n\e(B . "i~")                           ; 238
 121     (?\e,1o\e(B . "i?")                           ; 239
 122     (?\e,1p\e(B . "dd")                           ; 240
 123     (?\e,1q\e(B . "u+.")                  ; 241
 124     (?\e,1r\e(B . "o`")                           ; 242
 125     (?\e,1s\e(B . "o'")                           ; 243
 126     (?\e,1t\e(B . "o^")                           ; 244
 127     (?\e,1u\e(B . "o~")                           ; 245
 128     (?\e,1v\e(B . "o?")                           ; 246
 129     (?\e,1w\e(B . "o.")                           ; 247
 130     (?\e,1x\e(B . "u.")                           ; 248
 131     (?\e,1y\e(B . "u`")                           ; 249
 132     (?\e,1z\e(B . "u'")                           ; 250
 133     (?\e,1{\e(B . "u~")                           ; 251
 134     (?\e,1|\e(B . "u?")                           ; 252
 135     (?\e,1}\e(B . "y'")                           ; 253
 136     (?\e,1~\e(B . "o+.")                  ; 254
 137
 138     ;; upper case
 139     (?\e,2!\e(B . "A('")                  ; 161
 140     (?\e,2"\e(B . "A(`")                  ; 162
 141     (?\e,2#\e(B . "A(.")                  ; 163
 142     (?\e,2$\e(B . "A^'")                  ; 164
 143     (?\e,2%\e(B . "A^`")                  ; 165
 144     (?\e,2&\e(B . "A^?")                  ; 166
 145     (?\e,2'\e(B . "A^.")                  ; 167
 146     (?\e,2(\e(B . "E~")                           ; 168
 147     (?\e,2)\e(B . "E.")                           ; 169
 148     (?\e,2*\e(B . "E^'")                  ; 170
 149     (?\e,2+\e(B . "E^`")                  ; 171
 150     (?\e,2,\e(B . "E^?")                  ; 172
 151     (?\e,2-\e(B . "E^~")                  ; 173
 152     (?\e,2.\e(B . "E^.")                  ; 174
 153     (?\e,2/\e(B . "O^'")                  ; 175
 154     (?\e,20\e(B . "O^`")                  ; 176
 155     (?\e,21\e(B . "O^?")                  ; 177
 156     (?\e,22\e(B . "O^~")                  ; 178
 157     (?\e,25\e(B . "O^.")                  ; 181
 158     (?\e,26\e(B . "O+`")                  ; 182
 159     (?\e,27\e(B . "O+?")                  ; 183
 160     (?\e,28\e(B . "I.")                           ; 184
 161     (?\e,2=\e(B . "O+")                           ; 189
 162     (?\e,2>\e(B . "O+'")                  ; 190
 163     (?\e,2F\e(B . "A(?")                  ; 198
 164     (?\e,2G\e(B . "A(~")                  ; 199
 165     (?\e,2O\e(B . "Y`")                           ; 207
 166     (?\e,2Q\e(B . "U+'")                  ; 209
 167     (?\e,2U\e(B . "A.")                           ; 213
 168     (?\e,2V\e(B . "Y?")                           ; 214
 169     (?\e,2W\e(B . "U+`")                  ; 215
 170     (?\e,2X\e(B . "U+?")                  ; 216
 171     (?\e,2[\e(B . "Y~")                           ; 219
 172     (?\e,2\\e(B . "Y.")                           ; 220
 173     (?\e,2^\e(B . "O+~")                  ; 222
 174     (?\e,2_\e(B . "U+")                           ; 223
 175     (?\e,2`\e(B . "A`")                           ; 224
 176     (?\e,2a\e(B . "A'")                           ; 225
 177     (?\e,2b\e(B . "A^")                           ; 226
 178     (?\e,2c\e(B . "A~")                           ; 227
 179     (?\e,2d\e(B . "A?")                           ; 228
 180     (?\e,2e\e(B . "A(")                           ; 229
 181     (?\e,2f\e(B . "U+~")                  ; 230
 182     (?\e,2g\e(B . "A^~")                  ; 231
 183     (?\e,2h\e(B . "E`")                           ; 232
 184     (?\e,2i\e(B . "E'")                           ; 233
 185     (?\e,2j\e(B . "E^")                           ; 234
 186     (?\e,2k\e(B . "E?")                           ; 235
 187     (?\e,2l\e(B . "I`")                           ; 236
 188     (?\e,2m\e(B . "I'")                           ; 237
 189     (?\e,2n\e(B . "I~")                           ; 238
 190     (?\e,2o\e(B . "I?")                           ; 239
 191     (?\e,2p\e(B . "DD")                           ; 240
 192     (?\e,2p\e(B . "dD")                           ; 240
 193     (?\e,2p\e(B . "Dd")                           ; 240
 194     (?\e,2q\e(B . "U+.")                  ; 241
 195     (?\e,2r\e(B . "O`")                           ; 242
 196     (?\e,2s\e(B . "O'")                           ; 243
 197     (?\e,2t\e(B . "O^")                           ; 244
 198     (?\e,2u\e(B . "O~")                           ; 245
 199     (?\e,2v\e(B . "O?")                           ; 246
 200     (?\e,2w\e(B . "O.")                           ; 247
 201     (?\e,2x\e(B . "U.")                           ; 248
 202     (?\e,2y\e(B . "U`")                           ; 249
 203     (?\e,2z\e(B . "U'")                           ; 250
 204     (?\e,2{\e(B . "U~")                           ; 251
 205     (?\e,2|\e(B . "U?")                           ; 252
 206     (?\e,2}\e(B . "Y'")                           ; 253
 207     (?\e,2~\e(B . "O+.")                  ; 254
 208
 209     ;; escape from composition
 210     (?\( . "\\(")                       ; breve (left parenthesis)
 211     (?^ . "\\^")                        ; circumflex (caret)
 212     (?+ . "\\+")                        ; horn (plus sign)
 213     (?' . "\\'")                        ; acute (apostrophe)
 214     (?` . "\\`")                        ; grave (backquote)
 215     (?? . "\\?")                        ; hook above (question mark)
 216     (?~ . "\\~")                        ; tilde (tilde)
 217     (?. . "\\.")                        ; dot below (period)
 218     (?d . "\\d")                        ; d-bar (d)
 219     (?\\ . "\\\\")                      ; literal backslash
 220     )
 221   "Alist of Vietnamese characters vs corresponding `VIQR' string.")
 222
 223 ;; Regular expression matching single Vietnamese character represented
 224 ;; by VIQR.
 225 (defconst viqr-regexp
 226   "[aeiouyAEIOUY]\\([(^+]?['`?~.]\\|[(^+]\\)\\|[Dd][Dd]")
 227
 228 ;;;###autoload
 229 (defun viet-decode-viqr-region (from to)
 230   "Convert `VIQR' mnemonics of the current region to Vietnamese characters.
 231 When called from a program, expects two arguments,
 232 positions (integers or markers) specifying the stretch of the region."
 233   (interactive "r")
 234   (save-restriction
 235     (narrow-to-region from to)
 236     (goto-char (point-min))
 237     (while (re-search-forward viqr-regexp nil t)
 238       (let* ((viqr (buffer-substring (match-beginning 0) (match-end 0)))
 239              (ch (car (rassoc viqr viet-viqr-alist))))
 240         (if ch
 241             (progn
 242               (delete-region (match-beginning 0) (match-end 0))
 243               (insert ch)))))))
 244
 245 ;;;###autoload
 246 (defun viet-decode-viqr-buffer ()
 247   "Convert `VIQR' mnemonics of the current buffer to Vietnamese characters."
 248   (interactive)
 249   (viet-decode-viqr-region (point-min) (point-max)))
 250
 251 ;;;###autoload
 252 (defun viet-encode-viqr-region (from to)
 253   "Convert Vietnamese characters of the current region to `VIQR' mnemonics.
 254 When called from a program, expects two arguments,
 255 positions (integers or markers) specifying the stretch of the region."
 256   (interactive "r")
 257   (save-restriction
 258     (narrow-to-region from to)
 259     (goto-char (point-min))
 260     (while (re-search-forward "\\cv" nil t)
 261       (let* ((ch (preceding-char))
 262              (viqr (cdr (assq ch viet-viqr-alist))))
 263         (if viqr
 264             (progn
 265               (delete-char -1)
 266               (insert viqr)))))))
 267
 268 ;;;###autoload
 269 (defun viet-encode-viqr-buffer ()
 270   "Convert Vietnamese characters of the current buffer to `VIQR' mnemonics."
 271   (interactive)
 272   (viet-encode-viqr-region (point-min) (point-max)))
 273
 274 ;;;###autoload
 275 (defun viqr-post-read-conversion (len)
 276   (save-excursion
 277     (save-restriction
 278       (narrow-to-region (point) (+ (point) len))
 279       (let ((buffer-modified-p (buffer-modified-p)))
 280         (viet-decode-viqr-region (point-min) (point-max))
 281         (set-buffer-modified-p buffer-modified-p)
 282         (- (point-max) (point-min))))))
 283
 284 ;;;###autoload
 285 (defun viqr-pre-write-conversion (from to)
 286   (let ((old-buf (current-buffer)))
 287     (set-buffer (generate-new-buffer " *temp*"))
 288     (if (stringp from)
 289         (insert from)
 290       (insert-buffer-substring old-buf from to))
 291     (viet-encode-viqr-region (point-min) (point-max))
 292     ;; Should return nil as annotations.
 293     nil))
 294
 295 ;;;
 296 (provide 'viet-util)
 297
 298 ;;; viet-util.el ends here