lisp/language/viet-util.el

   1 ;;; viet-util.el --- utilities for Vietnamese  -*- coding: utf-8; -*-
   2
   3 ;; Copyright (C) 1998, 2001-2013 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( -> ă
  56 ;;  circumflex |    ^     | a^ -> â
  57 ;;    horn     |    +     | o+ -> ơ
  58 ;; ------------+----------+---------
  59 ;;    acute    |    '     | a' -> á
  60 ;;    grave    |    `     | a` -> à
  61 ;;  hook above |    ?     | a? -> ả
  62 ;;    tilde    |    ~     | a~ -> ã
  63 ;;   dot below |    .     | a. -> ạ
  64 ;; ------------+----------+---------
  65 ;;    d bar    |   dd     | dd -> đ
  66 ;; ------------+----------+---------
  67
  68 (defvar viet-viqr-alist
  69   '(;; lowercase
  70     (?ắ . "a('")                      ; 161
  71     (?ằ . "a(`")                      ; 162
  72     (?ặ . "a(.")                      ; 163
  73     (?ấ . "a^'")                      ; 164
  74     (?ầ . "a^`")                      ; 165
  75     (?ẩ . "a^?")                      ; 166
  76     (?ậ . "a^.")                      ; 167
  77     (?ẽ . "e~")                               ; 168
  78     (?ẹ . "e.")                               ; 169
  79     (?ế . "e^'")                      ; 170
  80     (?ề . "e^`")                      ; 171
  81     (?ể . "e^?")                      ; 172
  82     (?ễ . "e^~")                      ; 173
  83     (?ệ . "e^.")                      ; 174
  84     (?ố . "o^'")                      ; 175
  85     (?ồ . "o^`")                      ; 176
  86     (?ổ . "o^?")                      ; 177
  87     (?ỗ . "o^~")                      ; 178
  88     (?ộ . "o^.")                      ; 181
  89     (?ờ . "o+`")                      ; 182
  90     (?ở . "o+?")                      ; 183
  91     (?ị . "i.")                               ; 184
  92     (?ơ . "o+")                                ; 189
  93     (?ớ . "o+'")                      ; 190
  94     (?ẳ . "a(?")                      ; 198
  95     (?ẵ . "a(~")                      ; 199
  96     (?ỳ . "y`")                               ; 207
  97     (?ứ . "u+'")                      ; 209
  98     (?ạ . "a.")                               ; 213
  99     (?ỷ . "y?")                               ; 214
 100     (?ừ . "u+`")                      ; 215
 101     (?ử . "u+?")                      ; 216
 102     (?ỹ . "y~")                               ; 219
 103     (?ỵ . "y.")                               ; 220
 104     (?ỡ . "o+~")                      ; 222
 105     (?ư . "u+")                                ; 223
 106     (?à . "a`")                                ; 224
 107     (?á . "a'")                                ; 225
 108     (?â . "a^")                                ; 226
 109     (?ã . "a~")                                ; 227
 110     (?ả . "a?")                               ; 228
 111     (?ă . "a(")                                ; 229
 112     (?ữ . "u+~")                      ; 230
 113     (?ẫ . "a^~")                      ; 231
 114     (?è . "e`")                                ; 232
 115     (?é . "e'")                                ; 233
 116     (?ê . "e^")                                ; 234
 117     (?ẻ . "e?")                               ; 235
 118     (?ì . "i`")                                ; 236
 119     (?í . "i'")                                ; 237
 120     (?ĩ . "i~")                                ; 238
 121     (?ỉ . "i?")                               ; 239
 122     (?đ . "dd")                                ; 240
 123     (?ự . "u+.")                      ; 241
 124     (?ò . "o`")                                ; 242
 125     (?ó . "o'")                                ; 243
 126     (?ô . "o^")                                ; 244
 127     (?õ . "o~")                                ; 245
 128     (?ỏ . "o?")                               ; 246
 129     (?ọ . "o.")                               ; 247
 130     (?ụ . "u.")                               ; 248
 131     (?ù . "u`")                                ; 249
 132     (?ú . "u'")                                ; 250
 133     (?ũ . "u~")                                ; 251
 134     (?ủ . "u?")                               ; 252
 135     (?ý . "y'")                                ; 253
 136     (?ợ . "o+.")                      ; 254
 137
 138     ;; upper case
 139     (?Ắ . "A('")                      ; 161
 140     (?Ằ . "A(`")                      ; 162
 141     (?Ặ . "A(.")                      ; 163
 142     (?Ấ . "A^'")                      ; 164
 143     (?Ầ . "A^`")                      ; 165
 144     (?Ẩ . "A^?")                      ; 166
 145     (?Ậ . "A^.")                      ; 167
 146     (?Ẽ . "E~")                               ; 168
 147     (?Ẹ . "E.")                               ; 169
 148     (?Ế . "E^'")                      ; 170
 149     (?Ề . "E^`")                      ; 171
 150     (?Ể . "E^?")                      ; 172
 151     (?Ễ . "E^~")                      ; 173
 152     (?Ệ . "E^.")                      ; 174
 153     (?Ố . "O^'")                      ; 175
 154     (?Ồ . "O^`")                      ; 176
 155     (?Ổ . "O^?")                      ; 177
 156     (?Ỗ . "O^~")                      ; 178
 157     (?Ộ . "O^.")                      ; 181
 158     (?Ờ . "O+`")                      ; 182
 159     (?Ở . "O+?")                      ; 183
 160     (?Ị . "I.")                               ; 184
 161     (?Ơ . "O+")                                ; 189
 162     (?Ớ . "O+'")                      ; 190
 163     (?Ẳ . "A(?")                      ; 198
 164     (?Ẵ . "A(~")                      ; 199
 165     (?Ỳ . "Y`")                               ; 207
 166     (?Ứ . "U+'")                      ; 209
 167     (?Ạ . "A.")                               ; 213
 168     (?Ỷ . "Y?")                               ; 214
 169     (?Ừ . "U+`")                      ; 215
 170     (?Ử . "U+?")                      ; 216
 171     (?Ỹ . "Y~")                               ; 219
 172     (?Ỵ . "Y.")                               ; 220
 173     (?Ỡ . "O+~")                      ; 222
 174     (?Ư . "U+")                                ; 223
 175     (?À . "A`")                                ; 224
 176     (?Á . "A'")                                ; 225
 177     (?Â . "A^")                                ; 226
 178     (?Ã . "A~")                                ; 227
 179     (?Ả . "A?")                               ; 228
 180     (?Ă . "A(")                                ; 229
 181     (?Ữ . "U+~")                      ; 230
 182     (?Ẫ . "A^~")                      ; 231
 183     (?È . "E`")                                ; 232
 184     (?É . "E'")                                ; 233
 185     (?Ê . "E^")                                ; 234
 186     (?Ẻ . "E?")                               ; 235
 187     (?Ì . "I`")                                ; 236
 188     (?Í . "I'")                                ; 237
 189     (?Ĩ . "I~")                                ; 238
 190     (?Ỉ . "I?")                               ; 239
 191     (?Đ . "DD")                                ; 240
 192     (?Đ . "dD")                                ; 240
 193     (?Đ . "Dd")                                ; 240
 194     (?Ự . "U+.")                      ; 241
 195     (?Ò . "O`")                                ; 242
 196     (?Ó . "O'")                                ; 243
 197     (?Ô . "O^")                                ; 244
 198     (?Õ . "O~")                                ; 245
 199     (?Ỏ . "O?")                               ; 246
 200     (?Ọ . "O.")                               ; 247
 201     (?Ụ . "U.")                               ; 248
 202     (?Ù . "U`")                                ; 249
 203     (?Ú . "U'")                                ; 250
 204     (?Ũ . "U~")                                ; 251
 205     (?Ủ . "U?")                               ; 252
 206     (?Ý . "Y'")                                ; 253
 207     (?Ợ . "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