]> code.delx.au - gnu-emacs/blob - lisp/international/skkdic-cnv.el
Fix comment. Don't bother to require format.
[gnu-emacs] / lisp / international / skkdic-cnv.el
1 ;;; skkdic-cnv.el --- Convert a SKK dictionary for `skkdic-utl'
2
3 ;; Copyright (C) 1995 Electrotechnical Laboratory, JAPAN.
4 ;; Licensed to the Free Software Foundation.
5
6 ;; Keywords: mule, multilingual, Japanese, SKK
7
8 ;; This file is part of GNU Emacs.
9
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)
13 ;; any later version.
14
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.
19
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.
24
25 ;;; Commentary:
26
27 ;; SKK is a Japanese input method running on Mule created by Masahiko
28 ;; Sato <masahiko@sato.riec.tohoku.ac.jp>. Here we provide utilities
29 ;; to handle a dictionary distributed with SKK so that a different
30 ;; input method (e.g. quail-japanese) can utilize the dictionary.
31
32 ;; The format of SKK dictionary is quite simple. Each line has the
33 ;; form "KANASTRING /CONV1/CONV2/.../" which means KANASTRING (\e$B2>L>J8\e(B
34 ;; \e$B;zNs\e(B) can be converted to one of CONVi. CONVi is a Kanji (\e$B4A;z\e(B)
35 ;; and Kana (\e$B2>L>\e(B) mixed string.
36 ;;
37 ;; KANASTRING may have a trailing ASCII letter for Okurigana (\e$BAw$j2>L>\e(B)
38 ;; information. For instance, the trailing letter `k' means that one
39 ;; of the following Okurigana is allowed: \e$B$+$-$/$1$3\e(B. So, in that
40 ;; case, the string "KANASTRING\e$B$/\e(B" can be converted to one of "CONV1\e$B$/\e(B",
41 ;; CONV2\e$B$/\e(B, ...
42
43 ;;; Code:
44
45 ;; Name of a file to generate from SKK dictionary.
46 (defvar skkdic-filename "skkdic.el")
47
48 ;; To make a generated skkdic.el smaller.
49 (make-coding-system
50 'iso-2022-7bit-short
51 2 ?J
52 "Like `iso-2022-7bit' but no ASCII designation before SPC."
53 '(ascii nil nil nil t t nil t))
54
55 (defun skkdic-convert-okuri-ari (skkbuf buf)
56 (message "Processing OKURI-ARI entries ...")
57 (goto-char (point-min))
58 (save-excursion
59 (set-buffer buf)
60 (insert ";; Setting okuri-ari entries.\n"
61 "(skkdic-set-okuri-ari\n"))
62 (while (not (eobp))
63 (let ((from (point))
64 to)
65 (end-of-line)
66 (setq to (point))
67
68 (save-excursion
69 (set-buffer buf)
70 (insert-buffer-substring skkbuf from to)
71 (beginning-of-line)
72 (insert "\"")
73 (search-forward " ")
74 (delete-char 1) ; delete the first '/'
75 (let ((p (point)))
76 (end-of-line)
77 (delete-char -1) ; delete the last '/'
78 (subst-char-in-region p (point) ?/ ? 'noundo))
79 (insert "\"\n"))
80
81 (forward-line 1)))
82 (save-excursion
83 (set-buffer buf)
84 (insert ")\n\n")))
85
86 (defconst skkdic-postfix-list '(skkdic-postfix-list))
87
88 (defconst skkdic-postfix-data
89 '(("\e$B$$$-\e(B" "\e$B9T\e(B")
90 ("\e$B$,$+$j\e(B" "\e$B78\e(B")
91 ("\e$B$,$/\e(B" "\e$B3X\e(B")
92 ("\e$B$,$o\e(B" "\e$B@n\e(B")
93 ("\e$B$7$c\e(B" "\e$B<R\e(B")
94 ("\e$B$7$e$&\e(B" "\e$B=8\e(B")
95 ("\e$B$7$g$&\e(B" "\e$B>^\e(B" "\e$B>k\e(B")
96 ("\e$B$8$g$&\e(B" "\e$B>k\e(B")
97 ("\e$B$;$s\e(B" "\e$B@~\e(B")
98 ("\e$B$@$1\e(B" "\e$B3Y\e(B")
99 ("\e$B$A$c$/\e(B" "\e$BCe\e(B")
100 ("\e$B$F$s\e(B" "\e$BE9\e(B")
101 ("\e$B$H$&$2\e(B" "\e$BF=\e(B")
102 ("\e$B$I$*$j\e(B" "\e$BDL$j\e(B")
103 ("\e$B$d$^\e(B" "\e$B;3\e(B")
104 ("\e$B$P$7\e(B" "\e$B66\e(B")
105 ("\e$B$O$D\e(B" "\e$BH/\e(B")
106 ("\e$B$b$/\e(B" "\e$BL\\e(B")
107 ("\e$B$f$-\e(B" "\e$B9T\e(B")))
108
109 (defun skkdic-convert-postfix (skkbuf buf)
110 (message "Processing POSTFIX entries ...")
111 (goto-char (point-min))
112 (save-excursion
113 (set-buffer buf)
114 (insert ";; Setting postfix entries.\n"
115 "(skkdic-set-postfix\n"))
116
117 ;; Initialize SKKDIC-POSTFIX-LIST by predefined data
118 ;; SKKDIC-POSTFIX-DATA.
119 (save-excursion
120 (set-buffer buf)
121 (let ((l skkdic-postfix-data)
122 kana candidates entry)
123 (while l
124 (setq kana (car (car l)) candidates (cdr (car l)))
125 (insert "\"" kana)
126 (while candidates
127 (insert " " (car candidates))
128 (setq entry (lookup-nested-alist (car candidates)
129 skkdic-postfix-list nil nil t))
130 (if (consp (car entry))
131 (setcar entry (cons kana (car entry)))
132 (set-nested-alist (car candidates) (list kana)
133 skkdic-postfix-list))
134 (setq candidates (cdr candidates)))
135 (insert "\"\n")
136 (setq l (cdr l)))))
137
138 ;; Search postfix entries.
139 (while (re-search-forward "^[#<>?]\\(\\cH+\\) " nil t)
140 (let ((kana (match-string 1))
141 str candidates)
142 (while (looking-at "/[#0-9 ]*\\([^/\n]*\\)/")
143 (setq str (match-string 1))
144 (if (not (member str candidates))
145 (setq candidates (cons str candidates)))
146 (goto-char (match-end 1)))
147 (save-excursion
148 (set-buffer buf)
149 (insert "\"" kana)
150 (while candidates
151 (insert " " (car candidates))
152 (let ((entry (lookup-nested-alist (car candidates)
153 skkdic-postfix-list nil nil t)))
154 (if (consp (car entry))
155 (if (not (member kana (car entry)))
156 (setcar entry (cons kana (car entry))))
157 (set-nested-alist (car candidates) (list kana)
158 skkdic-postfix-list)))
159 (setq candidates (cdr candidates)))
160 (insert "\"\n"))))
161 (save-excursion
162 (set-buffer buf)
163 (insert ")\n\n")))
164
165 (defconst skkdic-prefix-list '(skkdic-prefix-list))
166
167 (defun skkdic-convert-prefix (skkbuf buf)
168 (message "Processing PREFIX entries ...")
169 (goto-char (point-min))
170 (save-excursion
171 (set-buffer buf)
172 (insert ";; Setting prefix entries.\n"
173 "(skkdic-set-prefix\n"))
174 (save-excursion
175 (while (re-search-forward "^\\(\\cH+\\)[<>?] " nil t)
176 (let ((kana (match-string 1))
177 str candidates)
178 (while (looking-at "/\\([^/\n]+\\)/")
179 (setq str (match-string 1))
180 (if (not (member str candidates))
181 (setq candidates (cons str candidates)))
182 (goto-char (match-end 1)))
183 (save-excursion
184 (set-buffer buf)
185 (insert "\"" kana)
186 (while candidates
187 (insert " " (car candidates))
188 (set-nested-alist (car candidates) kana skkdic-prefix-list)
189 (setq candidates (cdr candidates)))
190 (insert "\"\n")))))
191 (save-excursion
192 (set-buffer buf)
193 (insert ")\n\n")))
194
195 ;; FROM and TO point the head and tail of "/J../J../.../".
196 (defun skkdic-get-candidate-list (from to)
197 (let (candidates)
198 (goto-char from)
199 (while (re-search-forward "/\\cj+" to t)
200 (setq candidates (cons (buffer-substring (1+ (match-beginning 0))
201 (match-end 0))
202 candidates)))
203 candidates))
204
205 ;; Return entry for STR from nested alist ALIST.
206 (defsubst skkdic-get-entry (str alist)
207 (car (lookup-nested-alist str alist nil nil t)))
208
209
210 (defconst skkdic-word-list '(skkdic-word-list))
211
212 ;; Return t if substring of STR (between FROM and TO) can be broken up
213 ;; to chunks all of which can be derived from another entry in SKK
214 ;; dictionary. SKKBUF is the buffer where the original SKK dictionary
215 ;; is visited, KANA is the current entry for STR. FIRST is t iff this
216 ;; is called at top level.
217
218 (defun skkdic-breakup-string (skkbuf kana str from to &optional first)
219 (let ((len (- to from)))
220 (or (and (>= len 2)
221 (let ((min-idx (+ from 2))
222 (idx (if first (1- to ) to))
223 (found nil))
224 (while (and (not found) (>= idx min-idx))
225 (let ((kana2-list (skkdic-get-entry
226 (substring str from idx)
227 skkdic-word-list)))
228 (if (or (and (consp kana2-list)
229 (let ((kana-len (length kana))
230 kana2)
231 (catch 'skkdic-tag
232 (while kana2-list
233 (setq kana2 (car kana2-list))
234 (if (string-match kana2 kana)
235 (throw 'skkdic-tag t))
236 (setq kana2-list (cdr kana2-list)))))
237 (or (= idx to)
238 (skkdic-breakup-string skkbuf kana str
239 idx to)))
240 (and (stringp kana2-list)
241 (string-match kana2-list kana)))
242 (setq found t)
243 (setq idx (1- idx)))))
244 found))
245 (and first
246 (> len 2)
247 (let ((kana2 (skkdic-get-entry
248 (substring str from (1+ from))
249 skkdic-prefix-list)))
250 (and (stringp kana2)
251 (eq (string-match kana2 kana) 0)))
252 (skkdic-breakup-string skkbuf kana str (1+ from) to))
253 (and (not first)
254 (>= len 1)
255 (let ((kana2-list (skkdic-get-entry
256 (substring str from to)
257 skkdic-postfix-list)))
258 (and (consp kana2-list)
259 (let (kana2)
260 (catch 'skkdic-tag
261 (while kana2-list
262 (setq kana2 (car kana2-list))
263 (if (string= kana2
264 (substring kana (- (length kana2))))
265 (throw 'skkdic-tag t))
266 (setq kana2-list (cdr kana2-list)))))))))))
267
268 ;; Return list of candidates which excludes some from CANDIDATES.
269 ;; Excluded candidates can be derived from another entry.
270
271 (defun skkdic-reduced-candidates (skkbuf kana candidates)
272 (let (elt l)
273 (while candidates
274 (setq elt (car candidates))
275 (if (or (= (length elt) 1)
276 (and (string-match "^\\cj" elt)
277 (not (skkdic-breakup-string skkbuf kana elt 0 (length elt)
278 'first))))
279 (setq l (cons elt l)))
280 (setq candidates (cdr candidates)))
281 (nreverse l)))
282
283 (defconst skkdic-okuri-nasi-entries (list nil))
284 (defconst skkdic-okuri-nasi-entries-count 0)
285
286 (defun skkdic-collect-okuri-nasi ()
287 (message "Collecting OKURI-NASI entries ...")
288 (save-excursion
289 (let ((prev-ratio 0)
290 ratio)
291 (while (re-search-forward "^\\(\\cH+\\) \\(/\\cj.*\\)/$" nil t)
292 (let ((kana (match-string 1))
293 (candidates (skkdic-get-candidate-list (match-beginning 2)
294 (match-end 2))))
295 (setq skkdic-okuri-nasi-entries
296 (cons (cons kana candidates) skkdic-okuri-nasi-entries)
297 skkdic-okuri-nasi-entries-count
298 (1+ skkdic-okuri-nasi-entries-count))
299 (setq ratio (floor (/ (* (point) 100.0) (point-max))))
300 (if (/= ratio prev-ratio)
301 (progn
302 (message "collected %2d%% %s ..." ratio kana)
303 (setq prev-ratio ratio)))
304 (while candidates
305 (let ((entry (lookup-nested-alist (car candidates)
306 skkdic-word-list nil nil t)))
307 (if (consp (car entry))
308 (setcar entry (cons kana (car entry)))
309 (set-nested-alist (car candidates) (list kana)
310 skkdic-word-list)))
311 (setq candidates (cdr candidates))))))))
312
313 (defun skkdic-convert-okuri-nasi (skkbuf buf)
314 (message "Processing OKURI-NASI entries ...")
315 (save-excursion
316 (set-buffer buf)
317 (insert ";; Setting okuri-nasi entries.\n"
318 "(skkdic-set-okuri-nasi\n")
319 (let ((l (nreverse skkdic-okuri-nasi-entries))
320 (count 0)
321 (prev-ratio 0)
322 ratio)
323 (while l
324 (let ((kana (car (car l)))
325 (candidates (cdr (car l))))
326 (setq ratio (/ (* count 1000) skkdic-okuri-nasi-entries-count)
327 count (1+ count))
328 (if (/= prev-ratio (/ ratio 10))
329 (progn
330 (message "processed %2d%% %s ..." (/ ratio 10) kana)
331 (setq prev-ratio (/ ratio 10))))
332 (if (setq candidates
333 (skkdic-reduced-candidates skkbuf kana candidates))
334 (progn
335 (insert "\"" kana)
336 (while candidates
337 (insert " " (car candidates))
338 (setq candidates (cdr candidates)))
339 (insert "\"\n"))))
340 (setq l (cdr l))))
341 (insert ")\n\n")))
342
343 (defun skkdic-convert (filename &optional dirname)
344 "Convert SKK dictionary of FILENAME into the file \"skkdic.el\".
345 Optional argument DIRNAME if specified is the directory name under which
346 the generated \"skkdic.el\" is saved."
347 (interactive "FSKK dictionary file: ")
348 (message "Reading file \"%s\" ..." filename)
349 (let ((skkbuf(find-file-noselect (expand-file-name filename)))
350 (buf (get-buffer-create "*skkdic-work*")))
351 (save-excursion
352 ;; Setup and generate the header part of working buffer.
353 (set-buffer buf)
354 (erase-buffer)
355 (buffer-disable-undo)
356 (insert ";; skkdic.el -- dictionary for Japanese input method\n"
357 ";;\tGenerated by the command `skkdic-convert'\n"
358 ";;\tDate: " (current-time-string) "\n"
359 ";;\tOriginal SKK dictionary file: "
360 (file-name-nondirectory filename)
361 "\n\n"
362 ";;; Comment:\n\n"
363 ";; Do byte-compile this file again after any modification.\n\n"
364 ";;; Start of the header of the original SKK dictionary.\n\n")
365 (set-buffer skkbuf)
366 (widen)
367 (goto-char 1)
368 (let (pos)
369 (search-forward ";; okuri-ari")
370 (forward-line 1)
371 (setq pos (point))
372 (set-buffer buf)
373 (insert-buffer-substring skkbuf 1 pos))
374 (insert "\n"
375 ";;; Code:\n\n(eval-when-compile (require 'skkdic-cnv))\n\n")
376
377 ;; Generate the body part of working buffer.
378 (set-buffer skkbuf)
379 (let ((from (point))
380 to)
381 ;; Convert okuri-ari entries.
382 (search-forward ";; okuri-nasi")
383 (beginning-of-line)
384 (setq to (point))
385 (narrow-to-region from to)
386 (skkdic-convert-okuri-ari skkbuf buf)
387 (widen)
388
389 ;; Convert okuri-nasi postfix entries.
390 (goto-char to)
391 (forward-line 1)
392 (setq from (point))
393 (re-search-forward "^\\cH")
394 (setq to (match-beginning 0))
395 (narrow-to-region from to)
396 (skkdic-convert-postfix skkbuf buf)
397 (widen)
398
399 ;; Convert okuri-nasi prefix entries.
400 (goto-char to)
401 (skkdic-convert-prefix skkbuf buf)
402
403 ;;
404 (skkdic-collect-okuri-nasi)
405
406 ;; Convert okuri-nasi general entries.
407 (skkdic-convert-okuri-nasi skkbuf buf)
408
409 ;; Postfix
410 (save-excursion
411 (set-buffer buf)
412 (goto-char (point-max))
413 (insert ";;\n(provide 'skkdic)\n\n;; skkdic.el ends here\n")))
414
415 ;; Save the working buffer.
416 (set-buffer buf)
417 (set-visited-file-name (expand-file-name skkdic-filename dirname) t)
418 (set-buffer-file-coding-system 'iso-2022-7bit-short)
419 (save-buffer 0))
420 (kill-buffer skkbuf)
421 (switch-to-buffer buf)))
422
423 (defun batch-skkdic-convert ()
424 "Run `skkdic-convert' on the files remaining on the command line.
425 Use this from the command line, with `-batch';
426 it won't work in an interactive Emacs.
427 For example, invoke:
428 % emacs -batch -l skkdic-cnv -f batch-skkdic-convert SKK-JISYO.L
429 to generate \"skkdic.el\" from SKK dictionary file \"SKK-JISYO.L\".
430 To get complete usage, invoke:
431 % emacs -batch -l skkdic-cnv -f batch-skkdic-convert -h"
432 (defvar command-line-args-left) ; Avoid compiler warning.
433 (if (not noninteractive)
434 (error "`batch-skkdic-convert' should be used only with -batch"))
435 (if (string= (car command-line-args-left) "-h")
436 (progn
437 (message "To convert SKK-JISYO.L into skkdic.el:")
438 (message " %% emacs -batch -l skkdic-conv -f batch-skkdic-convert SKK-JISYO.L")
439 (message "To convert SKK-JISYO.L into DIR/skkdic.el:")
440 (message " %% emacs -batch -l skkdic-conv -f batch-skkdic-convert -dir DIR SKK-JISYO.L"))
441 (let (targetdir filename)
442 (if (string= (car command-line-args-left) "-dir")
443 (progn
444 (setq command-line-args-left (cdr command-line-args-left))
445 (setq targetdir (expand-file-name (car command-line-args-left)))
446 (setq command-line-args-left (cdr command-line-args-left))))
447 (setq filename (expand-file-name (car command-line-args-left)))
448 (message "Converting %s to skkdic.el ..." filename)
449 (message "It takes around 10 minutes even on Sun SS20.")
450 (skkdic-convert filename targetdir)
451 (message "Do byte-compile the created file by:")
452 (message " %% emacs -batch -f batch-byte-compile skkdic.el")
453 ))
454 (kill-emacs 0))
455
456
457 ;; The following macros are expanded at byte-compiling time so that
458 ;; compiled code can be loaded quickly.
459
460 (defun skkdic-get-kana-compact-codes (kana)
461 (let* ((len (length kana))
462 (vec (make-vector len 0))
463 (i 0)
464 ch)
465 (while (< i len)
466 (setq ch (aref kana i))
467 (aset vec i
468 (if (< ch 128) ; CH is an ASCII letter for OKURIGANA,
469 (- ch) ; represented by a negative code.
470 (if (= ch ?\e$B!<\e(B) ; `\e$B!<\e(B' is represented by 0.
471 0
472 (- (nth 2 (split-char ch)) 32))))
473 (setq i (1+ i)))
474 vec))
475
476 (defun skkdic-extract-conversion-data (entry)
477 (string-match "^\\cj+[a-z]* " entry)
478 (let ((kana (substring entry (match-beginning 0) (1- (match-end 0))))
479 (i (match-end 0))
480 candidates)
481 (while (string-match "[^ ]+" entry i)
482 (setq candidates (cons (match-string 0 entry) candidates))
483 (setq i (match-end 0)))
484 (cons (skkdic-get-kana-compact-codes kana) candidates)))
485
486 (defmacro skkdic-set-okuri-ari (&rest entries)
487 `(defconst skkdic-okuri-ari
488 ',(let ((l entries)
489 (map '(skkdic-okuri-ari))
490 entry)
491 (while l
492 (setq entry (skkdic-extract-conversion-data (car l)))
493 (set-nested-alist (car entry) (cdr entry) map)
494 (setq l (cdr l)))
495 map)))
496
497 (defmacro skkdic-set-postfix (&rest entries)
498 `(defconst skkdic-postfix
499 ',(let ((l entries)
500 (map '(nil))
501 (longest 1)
502 len entry)
503 (while l
504 (setq entry (skkdic-extract-conversion-data (car l)))
505 (setq len (length (car entry)))
506 (if (> len longest)
507 (setq longest len))
508 (let ((entry2 (lookup-nested-alist (car entry) map nil nil t)))
509 (if (consp (car entry2))
510 (let ((conversions (cdr entry)))
511 (while conversions
512 (if (not (member (car conversions) (car entry2)))
513 (setcar entry2 (cons (car conversions) (car entry2))))
514 (setq conversions (cdr conversions))))
515 (set-nested-alist (car entry) (cdr entry) map)))
516 (setq l (cdr l)))
517 (setcar map longest)
518 map)))
519
520 (defmacro skkdic-set-prefix (&rest entries)
521 `(defconst skkdic-prefix
522 ',(let ((l entries)
523 (map '(nil))
524 (longest 1)
525 len entry)
526 (while l
527 (setq entry (skkdic-extract-conversion-data (car l)))
528 (setq len (length (car entry)))
529 (if (> len longest)
530 (setq longest len))
531 (let ((entry2 (lookup-nested-alist (car entry) map len nil t)))
532 (if (consp (car entry2))
533 (let ((conversions (cdr entry)))
534 (while conversions
535 (if (not (member (car conversions) (car entry2)))
536 (setcar entry2 (cons (car conversions) (car entry2))))
537 (setq conversions (cdr conversions))))
538 (set-nested-alist (car entry) (cdr entry) map len)))
539 (setq l (cdr l)))
540 (setcar map longest)
541 map)))
542
543 (defmacro skkdic-set-okuri-nasi (&rest entries)
544 `(defconst skkdic-okuri-nasi
545 ',(let ((l entries)
546 (map '(skdic-okuri-nasi))
547 (count 0)
548 entry)
549 (while l
550 (setq count (1+ count))
551 (if (= (% count 10) 0)
552 (message (format "%d entries" count)))
553 (setq entry (skkdic-extract-conversion-data (car l)))
554 (set-nested-alist (car entry) (cdr entry) map)
555 (setq l (cdr l)))
556 map)))
557
558 (provide 'skkdic-cnv)
559
560 ;; skkdic-cnv.el ends here