1 ;;; tibet-util.el --- Support for inputting Tibetan characters
3 ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN.
4 ;; Licensed to the Free Software Foundation.
6 ;; Keywords: multilingual, Tibetan
8 ;; This file is part of GNU Emacs.
10 ;; GNU Emacs is free software; you can redistribute it and/or modify
11 ;; it under the terms of the GNU General Public License as published by
12 ;; the Free Software Foundation; either version 2, or (at your option)
15 ;; GNU Emacs is distributed in the hope that it will be useful,
16 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
17 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 ;; GNU General Public License for more details.
20 ;; You should have received a copy of the GNU General Public License
21 ;; along with GNU Emacs; see the file COPYING. If not, write to the
22 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
23 ;; Boston, MA 02111-1307, USA.
25 ;; Author: Toru TOMABECHI, <Toru.Tomabechi@orient.unil.ch>
27 ;; Created: Feb. 17. 1997
30 ;; 1997.03.13 Modification in treatment of text properties;
31 ;; Support for some special signs and punctuations.
36 (defun setup-tibetan-environment ()
38 (setup-english-environment)
39 (setq coding-category-iso-8-2 'tibetan)
41 (setq-default buffer-file-coding-system 'iso-2022-7bit)
43 (setq default-input-method "tibetan-wylie"))
45 ;;; This function makes a transcription string for
46 ;;; re-composing a character.
49 (defun tibetan-tibetan-to-transcription (ch)
50 "Return a transcription string of Tibetan character CH"
52 (l (append tibetan-consonant-transcription-alist
53 tibetan-vowel-transcription-alist
54 tibetan-precomposed-transcription-alist
55 tibetan-subjoined-transcription-alist))
56 decomp-l t-char trans str result)
57 (if (eq (char-charset char) 'composition)
58 (setq decomp-l (decompose-composite-char char 'list nil))
59 (setq decomp-l (cons char nil)))
62 (setq t-char (char-to-string (car decomp-l)))
63 (setq trans (car (rassoc t-char l)))
64 (setq str (concat str trans))
65 (setq decomp-l (cdr decomp-l)))
68 ;;; This function translates transcription string into a string of
69 ;;; Tibetan characters.
72 (defun tibetan-transcription-to-tibetan (transcription)
73 "Translate Roman transcription into a sequence of Tibetan components."
74 (let ((trans transcription)
75 (lp tibetan-precomposed-transcription-alist)
76 (l (append tibetan-consonant-transcription-alist
77 tibetan-vowel-transcription-alist
78 tibetan-subjoined-transcription-alist))
79 (case-fold-search nil)
80 substr t-char p-str t-str result)
84 (cond ((string-match tibetan-precomposed-regexp trans)
85 (setq substr (substring trans (match-beginning 0) (match-end 0)))
86 (setq trans (substring trans (match-end 0)))
87 (setq t-char (cdr (assoc substr lp)))
89 (while (string-match tibetan-regexp trans)
90 (setq substr (substring trans (match-beginning 0) (match-end 0)))
91 (setq trans (substring trans 0 (match-beginning 0)))
93 (cdr (assoc substr l)))
94 (setq t-str (concat t-char t-str)))
95 (setq result (concat p-str t-str))))
99 ;;; Functions for composing Tibetan character.
101 ;;; A Tibetan syllable is typically structured as follows:
103 ;;; [Prefix] C [C+] V [M] [Suffix [Post suffix]]
105 ;;; where C's are all vertically stacked, V appears below or above
106 ;;; consonant cluster and M is always put above the C[C+]V combination.
107 ;;; (Sanskrit visarga, though it is a vowel modifier, is considered
108 ;;; to be a punctuation.)
110 ;;; Here are examples of the words "bsgrubs" and "h'uM"
112 ;;;
\e$(7"7
\e2%q`"U
\e1"7"G
\e(B
\e2
\e$(7"H`#A`"U0"_
\e1
\e(B
120 ;;; Consonants ''', 'w', 'y', 'r' take special forms when they are used
121 ;;; as subjoined consonant. Consonant 'r' takes another special form
122 ;;; when used as superjoined as in "rka", and so on, while it does not
123 ;;; change its form when conjoined with subjoined ''', 'w' or 'y'
124 ;;; as in "rwa", "rya".
127 ;;; As a Tibetan input method should avoid using conversion key,
128 ;;; we use a "Tibetan glyph -> transcription -> Tibetan glyph"
129 ;;; translation at each key input.
131 ;;; 1st stage - Check the preceding char.
132 ;;; If the preceding char is Tibetan and composable, then
134 ;;; 2nd stage - Translate the preceding char into transcription
136 ;;; 3rd stage - Concatenate the transcription of preceding char
137 ;;; and the current input key.
139 ;;; 4th stage - Re-translate the concatenated transcription into
140 ;;; a sequence of Tibetan letters.
142 ;;; 5th stage - Convert leading consonants into one single precomposed char
145 ;;; 6th stage - Compose the consonants into one composite glyph.
147 ;;; (If the current input is a vowel sign or a vowel modifier,
148 ;;; then it is composed with preceding char without checking
149 ;;; except when the preceding char is a punctuation or a digit.)
153 ;;; This function is used to avoid composition
154 ;;; between Tibetan and non-Tibetan chars.
157 (defun tibetan-char-examin (ch)
158 "Check if char CH is Tibetan character.
159 Returns non-nil if CH is Tibetan. Otherwise, returns nil."
161 (if (eq (char-charset chr) 'composition)
162 (string-match "\\cq+" (decompose-composite-char chr))
163 (string-match "\\cq" (char-to-string chr)))))
165 ;;; This is used to avoid composition between digits, signs, punctuations
166 ;;; and word constituents.
169 (defun tibetan-composable-examin (ch)
170 "Check if Tibetan char CH is composable.
171 Returns t if CH is a composable char \(i.e. neither punctuation nor digit)."
174 (if (eq (char-charset chr) 'composition)
175 (setq chstr (decompose-composite-char chr))
176 (setq chstr (char-to-string chr)))
177 (not (string-match "[
\e$(7!1
\e(B-
\e$(7!o"f
\e$(8!;!=!?!@!A!D"`
\e(B]" chstr))))
180 ;;; This checks if a character to be composed contains already
181 ;;; one or more vowels / vowel modifiers. If the character contains
182 ;;; them, then no more consonant should be added.
185 (defun tibetan-complete-char-examin (ch)
186 "Check if composite char CH contains one or more vowel/vowel modifiers.
187 Returns non-nil, if CH contains vowel/vowel modifiers."
190 (if (eq (char-charset chr) 'composition)
191 (setq chstr (decompose-composite-char chr))
192 (setq chstr (char-to-string chr)))
193 (string-match "[
\e$(7!g!e"Q
\e(B-
\e$(7"^"_
\e(B-
\e$(7"l
\e(B]" chstr)))
195 ;;; This function makes a composite character consisting of two characters
196 ;;; vertically stacked.
199 (defun tibetan-vertical-stacking (first second upward)
200 "Return a vertically stacked composite char consisting of FIRST and SECOND.
201 If UPWARD is non-nil, then SECOND is put above FIRST."
203 (compose-chars first '(tc . bc) second)
204 (compose-chars first '(bc . tc) second)))
206 ;;; This function makes a composite char from a string.
207 ;;; Note that this function returns a string, not a char.
210 (defun tibetan-compose-string (str)
211 "Compose a sequence of Tibetan character components into a composite character.
212 Returns a string containing a composite character."
214 f-str s-str f-ch s-ch rest composed result)
215 ;;Make sure no redundant vowel sign is present.
217 "^\\(.+\\)\\(
\e$(7"Q
\e(B\\)\\([
\e$(7!I!g!e"Q
\e(B-
\e$(7"^"_
\e(B-
\e$(7"l
\e(B]\\)" t-str)
219 (match-string 1 t-str)
220 (match-string 3 t-str))))
222 "^\\(.+\\)\\([
\e$(7!I!g!e"Q
\e(B-
\e$(7"^"_
\e(B-
\e$(7"l
\e(B]\\)\\(
\e$(7"Q
\e(B\\)" t-str)
224 (match-string 1 t-str)
225 (match-string 2 t-str))))
228 ;; Consecutive base/precomposed consonants are reduced to the last one.
229 (while (string-match "^\\([
\e$(7"!
\e(B-
\e$(7"J$!
\e(B-
\e$(7%u
\e(B]\\)\\([
\e$(7"!
\e(B-
\e$(7"@"B
\e(B-
\e$(7"J$!
\e(B-
\e$(7%u
\e(B].*\\)" t-str)
230 (setq result (concat result (match-string 1 t-str)))
231 (setq t-str (match-string 2 t-str)))
232 ;; Vowel/vowel modifier, subjoined consonants are added one by one
233 ;; to the preceding element.
235 (string-match "^\\(.\\)\\([
\e$(7"A#!
\e(B-
\e$(7#J!I!g!e"Q
\e(B-
\e$(7"^"_
\e(B-
\e$(7"l
\e(B]\\)\\(.*\\)" t-str)
236 (setq f-str (match-string 1 t-str))
237 (setq f-ch (string-to-char f-str))
238 (setq s-str (match-string 2 t-str))
239 ;;Special treatment for 'a chung.
240 ;;If 'a follows a consonant, then turned into its subjoined form.
241 (if (and (string-match "
\e$(7"A
\e(B" s-str)
242 (not (tibetan-complete-char-examin f-ch)))
243 (setq s-str "
\e$(7#A
\e(B"))
244 (setq s-ch (string-to-char s-str))
245 (setq rest (match-string 3 t-str))
246 (cond ((string-match "\\c2" s-str);; upper vowel sign
248 (tibetan-vertical-stacking f-ch s-ch t)))
249 ((string-match "\\c3" s-str);; lower vowel sign
251 (tibetan-vertical-stacking f-ch s-ch nil)))
252 ;;Automatic conversion of ra-mgo (superscribed r).
253 ;;'r' is converted if followed by a subjoined consonant
254 ;;other than w, ', y, r.
255 ((and (string-match "
\e$(7"C
\e(B" f-str)
256 (not (string-match "[
\e$(7#>#A#B#C
\e(B]" s-str)))
257 (setq f-ch ?
\e$(7#P
\e(B)
259 (tibetan-vertical-stacking f-ch s-ch nil)))
260 ((not (tibetan-complete-char-examin f-ch))
261 ;;Initial base consonant is tranformed, if followed by
262 ;;a subjoined consonant, except when it is followed
263 ;;by a subscribed 'a.
264 (if (and (string-match "[
\e$(7"!
\e(B-
\e$(7"="?"@"D
\e(B-
\e$(7"J
\e(B]" f-str)
265 (not (string-match "
\e$(7#A
\e(B" s-str)))
268 (cdr (assoc f-str tibetan-base-to-subjoined-alist)))))
270 (tibetan-vertical-stacking f-ch s-ch nil)))
272 (setq composed s-str)
273 (setq result (concat result f-str))))
274 (setq t-str (concat composed rest)))
275 (setq result (concat result t-str))))
277 ;;; quail <-> conversion interface.
280 (defun tibetan-composition (pc key)
281 "Interface to quail input method.
282 Takes two arguments: char PC and string KEY, where PC is the preceding
283 character to be composed with current input KEY.
284 Returns a string which is the result of composition."
285 (let (trans cur-ch t-str result)
286 ;; Make a tibetan character corresponding to current input key.
287 (setq cur-ch (tibetan-transcription-to-tibetan key))
288 ;; Check if the preceding character is Tibetan and composable.
289 (cond ((and (tibetan-char-examin pc)
290 (tibetan-composable-examin pc))
291 ;;If Tibetan char corresponding to the current input key exists,
294 ;; Convert the preceding character into transcription,
295 ;; and concatenate it with the current input key,
296 (setq trans (tibetan-tibetan-to-transcription pc))
297 (setq trans (concat trans key))
298 ;; Concatenated transcription is converted to
299 ;; a sequence of Tibetan characters,
300 (setq t-str (tibetan-transcription-to-tibetan trans))
301 ;; And it is composed into a composite character.
302 (setq result (tibetan-compose-string t-str)))
305 ;; Simply concatenate the preceding character and
306 ;; the current input key.
307 (setq result (char-to-string pc))
308 (setq result (concat result key)))))
309 ;; If the preceding char is not Tibetan or not composable,
311 ;; pc = 0 means the point is at the beginning of buffer.
313 (setq result (char-to-string pc)))
315 (setq result (concat result cur-ch))
316 (setq result (concat result key))))
321 (defun tibetan-decompose-region (beg end)
322 "Decompose Tibetan characters in the region BEG END into their components.
323 Components are: base and subjoined consonants, vowel signs, vowel modifiers.
324 One column punctuations are converted to their 2 column equivalents."
326 (let (ch-str ch-beg ch-end)
329 (narrow-to-region beg end)
330 (goto-char (point-min))
331 ;; \\cq = Tibetan character
332 (while (re-search-forward "\\cq" nil t)
333 (setq ch-str (buffer-substring-no-properties
334 (match-beginning 0) (match-end 0)))
335 ;; Save the points. Maybe, using save-match-data is preferable.
336 ;; But in order not to lose the trace(because the body is too long),
337 ;; we save the points in variables.
338 (setq ch-beg (match-beginning 0))
339 (setq ch-end (match-end 0))
340 ;; Here starts the decomposition.
342 ;; 1 column punctuations -> 2 column equivalent
343 ((string-match "[
\e$(8!D!;!=!?!@!A"`
\e(B]" ch-str)
345 (car (rassoc ch-str tibetan-precomposition-rule-alist))))
346 ;; Decomposition of composite character.
347 ((eq (char-charset (string-to-char ch-str)) 'composition)
348 ;; Make a string which consists of a sequence of
350 (setq ch-str (decompose-composite-char (string-to-char ch-str)))
351 ;; Converts nyi zla into base elements.
352 (cond ((string= ch-str "
\e$(7#R#S#S#S
\e(B")
353 (setq ch-str "
\e$(7!4!5!5
\e(B"))
354 ((string= ch-str "
\e$(7#R#S#S
\e(B")
355 (setq ch-str "
\e$(7!4!5
\e(B"))
356 ((string= ch-str "
\e$(7#R#S!I
\e(B")
357 (setq ch-str "
\e$(7!6
\e(B"))
358 ((string= ch-str "
\e$(7#R#S
\e(B")
359 (setq ch-str "
\e$(7!4
\e(B")))))
360 ;; If the sequence of components starts with a subjoined consonants,
361 (if (string-match "^\\([
\e$(7#!
\e(B-
\e$(7#J
\e(B]\\)\\(.*\\)$" ch-str)
362 ;; then the first components is converted to its base form.
364 (concat (car (rassoc (match-string 1 ch-str)
365 tibetan-base-to-subjoined-alist))
366 (match-string 2 ch-str))))
367 ;; If the sequence of components starts with a precomposed character,
368 (if (string-match "^\\([
\e$(7$!
\e(B-
\e$(7%u
\e(B]\\)\\(.*\\)$" ch-str)
369 ;; then it is converted into a sequence of components.
371 (concat (car (rassoc (match-string 1 ch-str)
372 tibetan-precomposition-rule-alist))
373 (match-string 2 ch-str))))
374 ;; Special treatment for superscribed r.
375 (if (string-match "^
\e$(7#P
\e(B\\(.*\\)$" ch-str)
376 (setq ch-str (concat "
\e$(7"C
\e(B" (match-string 1 ch-str))))
377 ;; Finally, the result of decomposition is inserted, and
378 ;; the composite character is deleted.
379 (insert-and-inherit ch-str)
380 (delete-region ch-beg ch-end))))))
383 (defun tibetan-compose-region (beg end)
384 "Make composite chars from Tibetan character components in the region BEG END.
385 Two column punctuations are converted to their 1 column equivalents."
390 (narrow-to-region beg end)
391 (goto-char (point-min))
392 ;; First, sequence of components which has a precomposed equivalent
394 (while (re-search-forward
395 tibetan-precomposition-rule-regexp nil t)
396 (setq str (buffer-substring-no-properties
397 (match-beginning 0) (match-end 0)))
400 (cdr (assoc str tibetan-precomposition-rule-alist))))
401 (delete-region (match-beginning 0) (match-end 0)))
402 (goto-char (point-min))
403 ;; Then, composable elements are put into a composite character.
404 (while (re-search-forward
405 "[
\e$(7"!
\e(B-
\e$(7"J$!
\e(B-
\e$(7%u
\e(B]+[
\e$(7#!
\e(B-
\e$(7#J!I!g!e"Q
\e(B-
\e$(7"^"_
\e(B-
\e$(7"l
\e(B]+"
407 (setq str (buffer-substring-no-properties
408 (match-beginning 0) (match-end 0)))
410 (setq result (tibetan-compose-string str))
411 (insert-and-inherit result))
412 (delete-region (match-beginning 0) (match-end 0)))))))
415 ;;; This variable is used to avoid repeated decomposition.
417 (setq-default tibetan-decomposed nil)
420 (defun tibetan-decompose-buffer ()
421 "Decomposes Tibetan characters in the buffer into their components.
422 See also docstring of the function tibetan-decompose-region."
424 (make-local-variable 'tibetan-decomposed)
425 (cond ((not tibetan-decomposed)
426 (tibetan-decompose-region (point-min) (point-max))
427 (setq tibetan-decomposed t))))
430 (defun tibetan-compose-buffer ()
431 "Composes Tibetan character components in the buffer.
432 See also docstring of the function tibetan-compose-region."
434 (make-local-variable 'tibetan-decomposed)
435 (tibetan-compose-region (point-min) (point-max))
436 (setq tibetan-decomposed nil))
439 (defun tibetan-post-read-conversion (len)
442 (let ((buffer-modified-p (buffer-modified-p)))
443 (narrow-to-region (point) (+ (point) len))
444 (tibetan-compose-region (point-min) (point-max))
445 (set-buffer-modified-p buffer-modified-p)
447 (make-local-variable 'tibetan-decomposed)
448 (setq tibetan-decomposed nil))
452 (defun tibetan-pre-write-conversion (from to)
453 (setq tibetan-decomposed-temp tibetan-decomposed)
454 (let ((old-buf (current-buffer))
455 (work-buf (get-buffer-create " *tibetan-work*")))
456 (set-buffer work-buf)
460 (insert-buffer-substring old-buf from to))
461 (if (not tibetan-decomposed-temp)
462 (tibetan-decompose-region (point-min) (point-max)))
463 ;; Should return nil as annotations.
466 (provide 'tibet-util)
468 ;;; language/tibet-util.el ends here.