1 | : Use /bin/sh |
---|
2 | # |
---|
3 | # $Id: munchlist.X,v 1.2 1997-12-13 16:07:36 ghudson Exp $ |
---|
4 | # |
---|
5 | # Copyright 1987, 1988, 1989, 1992, 1993, Geoff Kuenning, Granada Hills, CA |
---|
6 | # All rights reserved. |
---|
7 | # |
---|
8 | # Redistribution and use in source and binary forms, with or without |
---|
9 | # modification, are permitted provided that the following conditions |
---|
10 | # are met: |
---|
11 | # |
---|
12 | # 1. Redistributions of source code must retain the above copyright |
---|
13 | # notice, this list of conditions and the following disclaimer. |
---|
14 | # 2. Redistributions in binary form must reproduce the above copyright |
---|
15 | # notice, this list of conditions and the following disclaimer in the |
---|
16 | # documentation and/or other materials provided with the distribution. |
---|
17 | # 3. All modifications to the source code must be clearly marked as |
---|
18 | # such. Binary redistributions based on modified source code |
---|
19 | # must be clearly marked as modified versions in the documentation |
---|
20 | # and/or other materials provided with the distribution. |
---|
21 | # 4. All advertising materials mentioning features or use of this software |
---|
22 | # must display the following acknowledgment: |
---|
23 | # This product includes software developed by Geoff Kuenning and |
---|
24 | # other unpaid contributors. |
---|
25 | # 5. The name of Geoff Kuenning may not be used to endorse or promote |
---|
26 | # products derived from this software without specific prior |
---|
27 | # written permission. |
---|
28 | # |
---|
29 | # THIS SOFTWARE IS PROVIDED BY GEOFF KUENNING AND CONTRIBUTORS ``AS IS'' AND |
---|
30 | # ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
---|
31 | # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
---|
32 | # ARE DISCLAIMED. IN NO EVENT SHALL GEOFF KUENNING OR CONTRIBUTORS BE LIABLE |
---|
33 | # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL |
---|
34 | # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS |
---|
35 | # OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) |
---|
36 | # HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT |
---|
37 | # LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY |
---|
38 | # OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF |
---|
39 | # SUCH DAMAGE. |
---|
40 | # |
---|
41 | # Given a list of words for ispell, generate a reduced list |
---|
42 | # in which all possible affixes have been collapsed. The reduced |
---|
43 | # list will match the same list as the original. |
---|
44 | # |
---|
45 | # Usage: |
---|
46 | # |
---|
47 | # munchlist [-l lang] [-c lang] [-s hashfile] [-D] [-w chars] [-v] \ |
---|
48 | # [file] ... |
---|
49 | # |
---|
50 | # Options: |
---|
51 | # |
---|
52 | # -l lang Specifies the language table to be used. The default |
---|
53 | # is "$LIBDIR/!!DEFLANG!!". |
---|
54 | # -c lang Specifies "conversion" language table. If this option is |
---|
55 | # given, the input file(s) will be assumed to be described by |
---|
56 | # this table, rather than the table given in the -l option. |
---|
57 | # This may be used to convert between incompatible language |
---|
58 | # tables. (When in doubt, use this option -- it doesn't |
---|
59 | # hurt, and it may save you from creating a dictionary that has |
---|
60 | # illegal words in it). The default is no conversion. |
---|
61 | # -T suff Specifies that the source word lists are in the format |
---|
62 | # of a "suff"-suffixed file, rather than in the |
---|
63 | # canonical form. For example, "-T tex" specifies that |
---|
64 | # string characters in the word lists are in TeX format. |
---|
65 | # The string character conversions are taken from the language |
---|
66 | # table specified by the "-l" switch. |
---|
67 | # -s Remove any words that are already covered by the |
---|
68 | # dictionary in 'hashfile'. The words will be removed |
---|
69 | # only if all affixes are covered. This option should not be |
---|
70 | # specified when the main dictionary is being munched. |
---|
71 | # 'Hashfile' must have been created with the language |
---|
72 | # table given in the -l option, but this is not checked. |
---|
73 | # -D Leave temporary files for debugging purposes |
---|
74 | # -w Passed on to ispell (specify chars that are part of a word) |
---|
75 | # Unfortunately, special characters must be quoted twice |
---|
76 | # rather than once when invoking this script. Also, since |
---|
77 | # buildhash doesn't accept this option, the final ispell -l |
---|
78 | # step ignores it, making it somewhat less than useful. |
---|
79 | # -v Report progress to stderr. |
---|
80 | # |
---|
81 | # The given input files are merged, then processed by 'ispell -c' |
---|
82 | # to generate possible affix lists; these are then combined |
---|
83 | # and reduced. The final result is written to standard output. |
---|
84 | # |
---|
85 | # For portability to older systems, I have avoided getopt. |
---|
86 | # |
---|
87 | # Geoff Kuenning |
---|
88 | # 2/28/87 |
---|
89 | # |
---|
90 | # $Log: not supported by cvs2svn $ |
---|
91 | # Revision 1.1.1.1 1997/09/03 21:08:12 ghudson |
---|
92 | # Import of ispell 3.1.20 |
---|
93 | # |
---|
94 | # Revision 1.53 1995/01/08 23:23:36 geoff |
---|
95 | # Support variable hashfile suffixes for DOS purposes. |
---|
96 | # |
---|
97 | # Revision 1.52 1994/12/27 23:08:46 geoff |
---|
98 | # Dynamically determine how to pass backslashes to 'tr' so that it'll |
---|
99 | # work on any machine. Define LC_CTYPE to work around yet more |
---|
100 | # internationalized sort programs. Work around a bug in GNU uniq that |
---|
101 | # uses the wrong separator between counts and duplicated lines. |
---|
102 | # |
---|
103 | # Revision 1.51 1994/11/21 07:02:54 geoff |
---|
104 | # Correctly quote the arguments to 'tr' when detecting systems with |
---|
105 | # unsigned sorts. Be sure to provide a zero exit status on all systems, |
---|
106 | # even if MUNCHDEBUG is not set. |
---|
107 | # |
---|
108 | # Revision 1.50 1994/10/25 05:46:05 geoff |
---|
109 | # Export values for LANG and LOCALE in an attempt to override some |
---|
110 | # stupidly-internationalized sort programs. |
---|
111 | # |
---|
112 | # Revision 1.49 1994/10/04 03:51:30 geoff |
---|
113 | # Add the MUNCHMAIL feature. If the MUNCHMAIL environment variable is |
---|
114 | # set to an email address, debugging information about the munchlist run |
---|
115 | # will automatically be collected and mailed to that address. |
---|
116 | # |
---|
117 | # Revision 1.48 1994/05/17 06:32:06 geoff |
---|
118 | # Don't look for affix tables in LIBDIR if the name contains a slash |
---|
119 | # |
---|
120 | # Revision 1.47 1994/04/27 02:50:48 geoff |
---|
121 | # Fix some cosmetic flaws in the verbose-mode messages. |
---|
122 | # |
---|
123 | # Revision 1.46 1994/01/25 07:11:59 geoff |
---|
124 | # Get rid of all old RCS log lines in preparation for the 3.1 release. |
---|
125 | # |
---|
126 | # |
---|
127 | if [ "X$MUNCHMAIL" != X ] |
---|
128 | then |
---|
129 | exec 2> /tmp/munchlist.mail |
---|
130 | echo "munchlist $*" 1>&2 |
---|
131 | set -vx |
---|
132 | fi |
---|
133 | LIBDIR=!!LIBDIR!! |
---|
134 | TDIR=${TMPDIR-/var/tmp} |
---|
135 | TMP=${TDIR}/munch$$ |
---|
136 | SORTTMP="-T ${TDIR}" # !!SORTTMP!! |
---|
137 | if [ -r ./icombine ] |
---|
138 | then |
---|
139 | COMBINE=./icombine |
---|
140 | else |
---|
141 | COMBINE=icombine |
---|
142 | fi |
---|
143 | if [ -r ./ijoin ] |
---|
144 | then |
---|
145 | JOIN=./ijoin |
---|
146 | else |
---|
147 | JOIN=ijoin |
---|
148 | fi |
---|
149 | |
---|
150 | # |
---|
151 | # The following is necessary so that some internationalized versions of |
---|
152 | # sort(1) don't confuse things by sorting into a nonstandard order. |
---|
153 | # |
---|
154 | LANG=C |
---|
155 | LOCALE=C |
---|
156 | LC_CTYPE=C |
---|
157 | export LANG LOCALE LC_CTYPE |
---|
158 | |
---|
159 | debug=no |
---|
160 | dictopt= |
---|
161 | langtabs=${LIBDIR}/!!DEFLANG!! |
---|
162 | convtabs= |
---|
163 | strip=no |
---|
164 | icflags= |
---|
165 | verbose=false |
---|
166 | # The following value of "wchars" is necessary to prevent ispell from |
---|
167 | # receiving a null argument if -w is not specified. As long as "A" is |
---|
168 | # a member of the existing character set, ispell will ignore the argument. |
---|
169 | wchars=-wA |
---|
170 | while [ $# != 0 ] |
---|
171 | do |
---|
172 | case "$1" in |
---|
173 | -l) |
---|
174 | case "$2" in |
---|
175 | */*) |
---|
176 | langtabs=$2 |
---|
177 | ;; |
---|
178 | *) |
---|
179 | if [ -r "$2" ] |
---|
180 | then |
---|
181 | langtabs="$2" |
---|
182 | else |
---|
183 | langtabs="${LIBDIR}/$2" |
---|
184 | fi |
---|
185 | ;; |
---|
186 | esac |
---|
187 | if [ ! -r "$langtabs" ] |
---|
188 | then |
---|
189 | echo "Can't open language table '$2'" 1>&2 |
---|
190 | exit 1 |
---|
191 | fi |
---|
192 | shift |
---|
193 | ;; |
---|
194 | -c) |
---|
195 | if [ -r "$2" ] |
---|
196 | then |
---|
197 | convtabs="$2" |
---|
198 | elif [ -r "${LIBDIR}/$2" ] |
---|
199 | then |
---|
200 | convtabs="${LIBDIR}/$2" |
---|
201 | else |
---|
202 | echo "Can't open conversion language table '$2'" 1>&2 |
---|
203 | exit 1 |
---|
204 | fi |
---|
205 | shift |
---|
206 | ;; |
---|
207 | -s) |
---|
208 | dictopt="-d $2" |
---|
209 | strip=yes |
---|
210 | shift |
---|
211 | ;; |
---|
212 | -D) |
---|
213 | debug=yes |
---|
214 | ;; |
---|
215 | -T) |
---|
216 | icflags="-T $2" |
---|
217 | shift |
---|
218 | ;; |
---|
219 | -v) |
---|
220 | verbose=true |
---|
221 | ;; |
---|
222 | -w) |
---|
223 | wchars="-w$2" |
---|
224 | shift |
---|
225 | ;; |
---|
226 | --) |
---|
227 | shift |
---|
228 | break |
---|
229 | ;; |
---|
230 | -) |
---|
231 | break |
---|
232 | ;; |
---|
233 | -*) |
---|
234 | echo 'Usage: munchlist [-l lang] [-c lang] [-T suff] [-s hashfile] [-D] [-w chars] [-v] [file] ...' \ |
---|
235 | 1>&2 |
---|
236 | exit 2 |
---|
237 | ;; |
---|
238 | *) |
---|
239 | break |
---|
240 | ;; |
---|
241 | esac |
---|
242 | shift |
---|
243 | done |
---|
244 | if [ "X$MUNCHMAIL" != X ] |
---|
245 | then |
---|
246 | verbose=true |
---|
247 | debug=yes |
---|
248 | fi |
---|
249 | trap "/bin/rm -f ${TMP}*; exit 1" 1 2 13 15 |
---|
250 | # |
---|
251 | # Names of temporary files. This is just to make the code a little easier |
---|
252 | # to read. |
---|
253 | # |
---|
254 | EXPANDEDINPUT=${TMP}a |
---|
255 | STRIPPEDINPUT=${TMP}b |
---|
256 | CRUNCHEDINPUT=${TMP}c |
---|
257 | PRODUCTLIST=${TMP}d |
---|
258 | EXPANDEDPAIRS=${TMP}e |
---|
259 | LEGALFLAGLIST=${TMP}f |
---|
260 | JOINEDPAIRS=${TMP}g |
---|
261 | MINIMALAFFIXES=${TMP}h |
---|
262 | CROSSROOTS=${TMP}i |
---|
263 | CROSSEXPANDED=${TMP}j |
---|
264 | CROSSPAIRS=${TMP}k |
---|
265 | CROSSILLEGAL=${TMP}l |
---|
266 | ILLEGALCOMBOS=${TMP}m |
---|
267 | FAKEDICT=${TMP}n |
---|
268 | # Ispell insists that hash files have a "!!HASHSUFFIX!!" suffix |
---|
269 | FAKEHASH=${TMP}o!!HASHSUFFIX!! |
---|
270 | AWKSCRIPT=${TMP}p |
---|
271 | if [ "$debug" = yes ] |
---|
272 | then |
---|
273 | touch $EXPANDEDINPUT $STRIPPEDINPUT $CRUNCHEDINPUT $PRODUCTLIST \ |
---|
274 | $EXPANDEDPAIRS $LEGALFLAGLIST $JOINEDPAIRS $MINIMALAFFIXES \ |
---|
275 | $CROSSROOTS $CROSSEXPANDED $CROSSPAIRS $CROSSILLEGAL $ILLEGALCOMBOS \ |
---|
276 | $FAKEDICT $FAKEHASH $AWKSCRIPT |
---|
277 | rm -f ${TDIR}/EXPANDEDINPUT ${TDIR}/STRIPPEDINPUT ${TDIR}/CRUNCHEDINPUT \ |
---|
278 | ${TDIR}/PRODUCTLIST ${TDIR}/EXPANDEDPAIRS ${TDIR}/LEGALFLAGLIST \ |
---|
279 | ${TDIR}/JOINEDPAIRS ${TDIR}/MINIMALAFFIXES ${TDIR}/CROSSROOTS \ |
---|
280 | ${TDIR}/CROSSEXPANDED ${TDIR}/CROSSPAIRS ${TDIR}/CROSSILLEGAL \ |
---|
281 | ${TDIR}/ILLEGALCOMBOS ${TDIR}/FAKEDICT ${TDIR}/FAKEHASH!!HASHSUFFIX!! \ |
---|
282 | ${TDIR}/AWKSCRIPT ${TDIR}/CROSSROOTS.[0-9]* ${TDIR}/CROSSEXP.[0-9]* \ |
---|
283 | ${TDIR}/CROSSPAIRS.[0-9]* ${TDIR}/CROSSILLEGAL.[0-9]* |
---|
284 | ln $EXPANDEDINPUT ${TDIR}/EXPANDEDINPUT |
---|
285 | ln $STRIPPEDINPUT ${TDIR}/STRIPPEDINPUT |
---|
286 | ln $CRUNCHEDINPUT ${TDIR}/CRUNCHEDINPUT |
---|
287 | ln $PRODUCTLIST ${TDIR}/PRODUCTLIST |
---|
288 | ln $EXPANDEDPAIRS ${TDIR}/EXPANDEDPAIRS |
---|
289 | ln $LEGALFLAGLIST ${TDIR}/LEGALFLAGLIST |
---|
290 | ln $JOINEDPAIRS ${TDIR}/JOINEDPAIRS |
---|
291 | ln $MINIMALAFFIXES ${TDIR}/MINIMALAFFIXES |
---|
292 | ln $CROSSROOTS ${TDIR}/CROSSROOTS |
---|
293 | ln $CROSSEXPANDED ${TDIR}/CROSSEXPANDED |
---|
294 | ln $CROSSPAIRS ${TDIR}/CROSSPAIRS |
---|
295 | ln $CROSSILLEGAL ${TDIR}/CROSSILLEGAL |
---|
296 | ln $ILLEGALCOMBOS ${TDIR}/ILLEGALCOMBOS |
---|
297 | ln $FAKEDICT ${TDIR}/FAKEDICT |
---|
298 | ln $FAKEHASH ${TDIR}/FAKEHASH!!HASHSUFFIX!! |
---|
299 | ln $AWKSCRIPT ${TDIR}/AWKSCRIPT |
---|
300 | fi |
---|
301 | # |
---|
302 | # Create a dummy dictionary to hold a compiled copy of the language |
---|
303 | # table. Initially, it holds the conversion table, if it exists. |
---|
304 | # |
---|
305 | case "X$convtabs" in |
---|
306 | X) |
---|
307 | convtabs="$langtabs" |
---|
308 | ;; |
---|
309 | esac |
---|
310 | echo 'QQQQQQQQ' > $FAKEDICT |
---|
311 | buildhash -s $FAKEDICT $convtabs $FAKEHASH \ |
---|
312 | || (echo "Couldn't create fake hash file" 1>&2; /bin/rm -f ${TMP}*; exit 1) \ |
---|
313 | || exit 1 |
---|
314 | # |
---|
315 | # Figure out how 'sort' sorts signed fields, for arguments to ijoin. |
---|
316 | # This is a little bit of a tricky pipe, but the result is that SIGNED |
---|
317 | # is set to "-s" if characters with the top bit set sort before those |
---|
318 | # without, and "-u" if the reverse is true. How does it work? The |
---|
319 | # first "tr" step generates two lines, one containing "-u", the other |
---|
320 | # with the same but with the high-order bit set. The second "tr" |
---|
321 | # changesthe high-bit "-u" back to "-s". If the high-bit "-u" was |
---|
322 | # sorted first, the sed step will select "-s" for SIGNED; otherwise |
---|
323 | # it'll pick "-u". We have to be careful about backslash quoting |
---|
324 | # conventions, because some systems differ. |
---|
325 | # |
---|
326 | backslash=\\ |
---|
327 | for i in 0 1 2 3 |
---|
328 | do |
---|
329 | if [ `echo a | tr "${backslash}141" b` = b ] |
---|
330 | then |
---|
331 | break |
---|
332 | fi |
---|
333 | backslash="$backslash$backslash" |
---|
334 | done |
---|
335 | SIGNED=`echo '-s |
---|
336 | -u' | tr s "${backslash}365" | sort | tr "${backslash}365" s | sed 1q` |
---|
337 | # |
---|
338 | # Collect all the input and expand all the affix options (ispell -e), |
---|
339 | # and preserve (sorted) for later joining in EXPANDEDINPUT. The icombine |
---|
340 | # step is to make sure that unneeded capitalizations (e.g., Farmer and farmer) |
---|
341 | # are weeded out. The first sort must be folded for icombine; the second |
---|
342 | # must be unfolded for join. |
---|
343 | # |
---|
344 | $verbose && echo "Collecting input." 1>&2 |
---|
345 | if [ $# -eq 0 ] |
---|
346 | then |
---|
347 | ispell "$wchars" -e1 -d $FAKEHASH -p /dev/null | tr " " ' |
---|
348 | ' |
---|
349 | else |
---|
350 | cat "$@" | ispell "$wchars" -e1 -d $FAKEHASH -p /dev/null | tr " " ' |
---|
351 | ' |
---|
352 | fi \ |
---|
353 | | sort $SORTTMP -u +0f -1 +0 \ |
---|
354 | | $COMBINE $icflags $langtabs \ |
---|
355 | | sort $SORTTMP -u > $EXPANDEDINPUT |
---|
356 | # |
---|
357 | # If a conversion table existed, recreate the fake hash file with the |
---|
358 | # "real" language table. |
---|
359 | # |
---|
360 | case "$convtabs" in |
---|
361 | $langtabs) |
---|
362 | ;; |
---|
363 | *) |
---|
364 | buildhash -s $FAKEDICT $langtabs $FAKEHASH \ |
---|
365 | || (echo "Couldn't create fake hash file" 1>&2; \ |
---|
366 | /bin/rm -f ${TMP}*; exit 1) \ |
---|
367 | || exit 1 |
---|
368 | ;; |
---|
369 | esac |
---|
370 | /bin/rm -f ${FAKEDICT}* |
---|
371 | # |
---|
372 | # If the -s (strip) option was specified, remove all |
---|
373 | # expanded words that are covered by the dictionary. This produces |
---|
374 | # the final list of expanded words that this dictionary must cover. |
---|
375 | # Leave the list in STRIPPEDINPUT. |
---|
376 | # |
---|
377 | if [ "X$strip" = "Xno" ] |
---|
378 | then |
---|
379 | rm -f $STRIPPEDINPUT |
---|
380 | ln $EXPANDEDINPUT $STRIPPEDINPUT |
---|
381 | if [ "$debug" = yes ] |
---|
382 | then |
---|
383 | rm -f ${TDIR}/STRIPPEDINPUT |
---|
384 | ln $STRIPPEDINPUT ${TDIR}/STRIPPEDINPUT |
---|
385 | fi |
---|
386 | else |
---|
387 | $verbose && echo "Stripping words already in the dictionary." 1>&2 |
---|
388 | ispell "$wchars" -l $dictopt -p /dev/null < $EXPANDEDINPUT \ |
---|
389 | > $STRIPPEDINPUT |
---|
390 | fi |
---|
391 | # |
---|
392 | # Figure out what the flag-marking character is. |
---|
393 | # |
---|
394 | $verbose && echo "Finding flag marker." 1>&2 |
---|
395 | flagmarker=`ispell -D -d $FAKEHASH \ |
---|
396 | | sed -n '/^flagmarker/s/flagmarker //p'` |
---|
397 | case "$flagmarker" in |
---|
398 | \\*) |
---|
399 | flagmarker=`expr "$flagmarker" : '.\(.\)'` |
---|
400 | ;; |
---|
401 | esac |
---|
402 | # |
---|
403 | # Munch the input to generate roots and affixes (ispell -c). We are |
---|
404 | # only interested in words that have at least one affix (egrep $flagmarker); |
---|
405 | # the next step will pick up the rest. Some of the roots are illegal. We |
---|
406 | # use join to restrict the output to those root words that are found |
---|
407 | # in the original dictionary. |
---|
408 | # |
---|
409 | $verbose && echo "Generating roots and affixes." 1>&2 |
---|
410 | ispell "$wchars" -c -W0 -d $FAKEHASH -p /dev/null < $STRIPPEDINPUT \ |
---|
411 | | tr " " ' |
---|
412 | ' \ |
---|
413 | | egrep "$flagmarker" | sort $SORTTMP -u "-t$flagmarker" +0 -1 +1 \ |
---|
414 | | $JOIN $SIGNED "-t$flagmarker" - $EXPANDEDINPUT > $CRUNCHEDINPUT |
---|
415 | # |
---|
416 | # We now have a list of legal roots, and of affixes that apply to the |
---|
417 | # root words. However, it is possible for some affix flags to generate more |
---|
418 | # than one output word. For example, with the flag table entry |
---|
419 | # |
---|
420 | # flag R: . > ER |
---|
421 | # . > ERS |
---|
422 | # |
---|
423 | # the input "BOTHER" will generate an entry "BOTH/R" in CRUNCHEDINPUT. But |
---|
424 | # this will accept "BOTHER" and "BOTHERS" in the dictionary, which is |
---|
425 | # wrong (in this case, though it's good English). |
---|
426 | # |
---|
427 | # To cure this problem, we first have to know which flags generate which |
---|
428 | # expansions. We use ispell -e3 to expand the flags (the second e causes |
---|
429 | # the root and flag to be included in the output), and get pairs |
---|
430 | # suitable for joining. In the example above, we would get |
---|
431 | # |
---|
432 | # BOTH/R BOTHER |
---|
433 | # BOTH/R BOTHERS |
---|
434 | # |
---|
435 | # We save this in EXPANDEDPAIRS for the next step. |
---|
436 | # |
---|
437 | $verbose && echo 'Expanding dictionary into EXPANDEDPAIRS.' 1>&2 |
---|
438 | ispell "$wchars" -e3 -d $FAKEHASH -p /dev/null < $CRUNCHEDINPUT \ |
---|
439 | | sort $SORTTMP +1 > $EXPANDEDPAIRS |
---|
440 | # |
---|
441 | # Now we want to extract the lines in EXPANDEDPAIRS in which the second field |
---|
442 | # is *not* listed in the original dictionary EXPANDEDINPUT; these illegal |
---|
443 | # lines contain the flags we cannot include without accepting illegal words. |
---|
444 | # It is somewhat easier to extract those which actually are listed (with |
---|
445 | # join), and then use comm to strip these from EXPANDEDPAIRS to get the |
---|
446 | # illegal expansions, together with the flags that generate them (we must |
---|
447 | # re-sort EXPANDEDPAIRS before running comm). Sed |
---|
448 | # gets rid of the expansion and uniq gets rid of duplicates. Comm then |
---|
449 | # selects the remainder of the list from CRUNCHEDINPUT and puts it in |
---|
450 | # LEGALFLAGLIST. The final step is to use a sort and icombine to put |
---|
451 | # the list into a one-entry-per-root format. |
---|
452 | # |
---|
453 | # BTW, I thought of using cut for the sed step (on systems that have it), |
---|
454 | # but it turns out that sed is faster! |
---|
455 | # |
---|
456 | $JOIN -j1 2 -o 1.1 1.2 $SIGNED $EXPANDEDPAIRS $EXPANDEDINPUT \ |
---|
457 | | sort $SORTTMP -u > $JOINEDPAIRS |
---|
458 | |
---|
459 | sort $SORTTMP -o $EXPANDEDPAIRS $EXPANDEDPAIRS |
---|
460 | sort $SORTTMP -o $CRUNCHEDINPUT $CRUNCHEDINPUT |
---|
461 | |
---|
462 | $verbose && echo 'Creating list of legal roots/flags.' 1>&2 |
---|
463 | comm -13 $JOINEDPAIRS $EXPANDEDPAIRS \ |
---|
464 | | (sed -e 's; .*$;;' ; /bin/rm -f $JOINEDPAIRS $EXPANDEDPAIRS) \ |
---|
465 | | uniq \ |
---|
466 | | (comm -13 - $CRUNCHEDINPUT ; /bin/rm -f $CRUNCHEDINPUT) \ |
---|
467 | | sort $SORTTMP -u "-t$flagmarker" +0f -1 +0 \ |
---|
468 | | $COMBINE $langtabs > $LEGALFLAGLIST |
---|
469 | |
---|
470 | # |
---|
471 | # LEGALFLAGLIST now contains root/flag combinations that, when expanded, |
---|
472 | # produce only words from EXPANDEDPAIRS. However, there is still a |
---|
473 | # problem if the language tables have any cross-product flags. A legal |
---|
474 | # root may appear in LEGALFLAGLIST with two flags that participate |
---|
475 | # in cross-products. When such a dictionary entry is expanded, |
---|
476 | # the cross-products will generate some extra words that may not |
---|
477 | # be in EXPANDEDPAIRS. We need to remove these from LEGALFLAGLIST. |
---|
478 | # |
---|
479 | # The first step is to collect the names of the flags that participate |
---|
480 | # in cross-products. Ispell will dump the language tables for us, and |
---|
481 | # sed is a pretty handy way to strip out extra information. We use |
---|
482 | # uniq -c and a numerical sort to put the flags in approximate order of how |
---|
483 | # "productive" they are (in terms of how likely they are to generate a lot |
---|
484 | # of output words). The least-productive flags are given last and will |
---|
485 | # be removed first. |
---|
486 | # |
---|
487 | $verbose \ |
---|
488 | && echo 'Creating list of flags that participate in cross-products.' 1>&2 |
---|
489 | ispell -D -d $FAKEHASH \ |
---|
490 | | sed -n '1,$s/:.*$// |
---|
491 | /^flagmarker/d |
---|
492 | /^prefixes/,/^suffixes/s/^ flag \*/p /p |
---|
493 | /^suffixes/,$s/^ flag \*/s /p' \ |
---|
494 | | sort $SORTTMP \ |
---|
495 | | uniq -c \ |
---|
496 | | tr ' ' ' ' \ |
---|
497 | | sort $SORTTMP +0rn -1 +2 > $PRODUCTLIST |
---|
498 | |
---|
499 | if [ `egrep ' p ' $PRODUCTLIST | wc -l` -gt 0 \ |
---|
500 | -a `egrep ' s ' $PRODUCTLIST | wc -l` -gt 0 ] |
---|
501 | then |
---|
502 | # |
---|
503 | # The language tables allow cross products. See if LEGALFLAGLIST has |
---|
504 | # any roots with multiple cross-product flags. Put them in CROSSROOTS. |
---|
505 | # |
---|
506 | $verbose && echo 'Finding prefix and suffix flags.' 1>&2 |
---|
507 | preflags=`sed -n 's/^[ 0-9]*p //p' $PRODUCTLIST | tr -d ' |
---|
508 | '` |
---|
509 | sufflags=`sed -n 's/^[ 0-9]*s //p' $PRODUCTLIST | tr -d ' |
---|
510 | '` |
---|
511 | egrep "$flagmarker.*[$preflags].*[$sufflags]|$flagmarker.*[$sufflags].*[$preflags]" \ |
---|
512 | $LEGALFLAGLIST \ |
---|
513 | > $CROSSROOTS |
---|
514 | |
---|
515 | # |
---|
516 | # We will need an awk script; it's so big that it core-dumps my shell |
---|
517 | # under certain conditions. The rationale behind the script is commented |
---|
518 | # where the script is used. Note that you may want to change this |
---|
519 | # script for languages other than English. |
---|
520 | # |
---|
521 | case "$flagmarker" in |
---|
522 | /) |
---|
523 | sedchar=: |
---|
524 | ;; |
---|
525 | *) |
---|
526 | sedchar=/ |
---|
527 | ;; |
---|
528 | esac |
---|
529 | $verbose && echo 'Creating awk script.' 1>&2 |
---|
530 | sed -e "s/PREFLAGS/$preflags/" -e "s/SUFFLAGS/$sufflags/" \ |
---|
531 | -e "s;ILLEGALCOMBOS;$ILLEGALCOMBOS;" \ |
---|
532 | -e "s${sedchar}FLAGMARKER${sedchar}$flagmarker${sedchar}" \ |
---|
533 | > $AWKSCRIPT << 'ENDOFAWKSCRIPT' |
---|
534 | BEGIN \ |
---|
535 | { |
---|
536 | preflags = "PREFLAGS" |
---|
537 | sufflags = "SUFFLAGS" |
---|
538 | illegalcombos = "ILLEGALCOMBOS" |
---|
539 | flagmarker = "FLAGMARKER" |
---|
540 | pflaglen = length (preflags) |
---|
541 | for (i = 1; i <= pflaglen; i++) |
---|
542 | pflags[i] = substr (preflags, i, 1); |
---|
543 | sflaglen = length (sufflags) |
---|
544 | for (i = 1; i <= sflaglen; i++) |
---|
545 | sflags[i] = substr (sufflags, i, 1); |
---|
546 | } |
---|
547 | { |
---|
548 | len = length ($2) |
---|
549 | pnew2 = "" |
---|
550 | snew2 = "" |
---|
551 | pbad = "" |
---|
552 | sbad = "" |
---|
553 | sufs = 0 |
---|
554 | pres = 0 |
---|
555 | for (i = 1; i <= len; i++) |
---|
556 | { |
---|
557 | curflag = substr ($2, i, 1) |
---|
558 | for (j = 1; j <= pflaglen; j++) |
---|
559 | { |
---|
560 | if (pflags[j] == curflag) |
---|
561 | { |
---|
562 | pres++ |
---|
563 | pnew2 = substr ($2, 1, i - 1) substr ($2, i + 1) |
---|
564 | pbad = curflag |
---|
565 | } |
---|
566 | } |
---|
567 | for (j = 1; j <= sflaglen; j++) |
---|
568 | { |
---|
569 | if (sflags[j] == curflag) |
---|
570 | { |
---|
571 | sufs++ |
---|
572 | snew2 = substr ($2, 1, i - 1) substr ($2, i + 1) |
---|
573 | sbad = curflag |
---|
574 | } |
---|
575 | } |
---|
576 | } |
---|
577 | if (pres == 1) |
---|
578 | { |
---|
579 | print $1 flagmarker pnew2 |
---|
580 | print $1 flagmarker pbad >> illegalcombos |
---|
581 | } |
---|
582 | else if (sufs == 1) |
---|
583 | { |
---|
584 | print $1 flagmarker snew2 |
---|
585 | print $1 flagmarker sbad >> illegalcombos |
---|
586 | } |
---|
587 | else if (pres > 0) |
---|
588 | { |
---|
589 | print $1 flagmarker pnew2 |
---|
590 | print $1 flagmarker pbad >> illegalcombos |
---|
591 | } |
---|
592 | else |
---|
593 | { |
---|
594 | print $1 flagmarker snew2 |
---|
595 | print $1 flagmarker sbad >> illegalcombos |
---|
596 | } |
---|
597 | } |
---|
598 | ENDOFAWKSCRIPT |
---|
599 | : > $ILLEGALCOMBOS |
---|
600 | dbnum=0 |
---|
601 | while [ -s $CROSSROOTS ] |
---|
602 | do |
---|
603 | # |
---|
604 | # CROSSROOTS contains the roots whose cross-product expansions |
---|
605 | # might be illegal. We now need to locate the actual illegal ones. |
---|
606 | # We do this in much the same way we created LEGALFLAGLIST from |
---|
607 | # CRUNCHEDINPUT. First we make CROSSEXPANDED, which is analogous |
---|
608 | # to EXPANDEDPAIRS. |
---|
609 | # |
---|
610 | $verbose && echo "Creating cross expansions (pass $dbnum)." 1>&2 |
---|
611 | ispell "$wchars" -e3 -d $FAKEHASH -p /dev/null < $CROSSROOTS \ |
---|
612 | | sort $SORTTMP +1 > $CROSSEXPANDED |
---|
613 | # |
---|
614 | # Now we join CROSSEXPANDED against EXPANDEDINPUT to produce |
---|
615 | # CROSSPAIRS, and then comm that against CROSSEXPANDED to |
---|
616 | # get CROSSILLEGAL, the list of illegal cross-product flag |
---|
617 | # combinations. |
---|
618 | # |
---|
619 | $JOIN -j1 2 -o 1.1 1.2 $SIGNED $CROSSEXPANDED $EXPANDEDINPUT \ |
---|
620 | | sort $SORTTMP -u > $CROSSPAIRS |
---|
621 | |
---|
622 | sort $SORTTMP -u -o $CROSSEXPANDED $CROSSEXPANDED |
---|
623 | |
---|
624 | $verbose \ |
---|
625 | && echo "Finding illegal cross expansions (pass $dbnum)." 1>&2 |
---|
626 | comm -13 $CROSSPAIRS $CROSSEXPANDED \ |
---|
627 | | sed -e 's; .*$;;' \ |
---|
628 | | uniq > $CROSSILLEGAL |
---|
629 | |
---|
630 | if [ "$debug" = yes ] |
---|
631 | then |
---|
632 | mv $CROSSROOTS $TDIR/CROSSROOTS.$dbnum |
---|
633 | ln $CROSSEXPANDED $TDIR/CROSSEXP.$dbnum |
---|
634 | ln $CROSSPAIRS $TDIR/CROSSPAIRS.$dbnum |
---|
635 | ln $CROSSILLEGAL $TDIR/CROSSILLEGAL.$dbnum |
---|
636 | fi |
---|
637 | # |
---|
638 | # Now it is time to try to clear up the illegalities. For |
---|
639 | # each word in the illegal list, remove one of the cross-product |
---|
640 | # flags. The flag chosen is selected in an attempt to cure the |
---|
641 | # problem quickly, as follows: (1) if there is only one suffix |
---|
642 | # flag or only one prefix flag, we remove that. (2) If there is |
---|
643 | # a prefix flag, we remove the "least desirable" (according to |
---|
644 | # the order of preflags). (This may be pro-English prejudice, |
---|
645 | # and you might want to change this if your language is prefix-heavy). |
---|
646 | # (3) Otherwise we remove the least-desirable suffix flag |
---|
647 | # |
---|
648 | # The output of the awk script becomes the new CROSSROOTS. In |
---|
649 | # addition, we add the rejected flags to ILLEGALCOMBOS (this is done |
---|
650 | # inside the awk script) so they can be removed from LEGALFLAGLIST |
---|
651 | # later. |
---|
652 | # |
---|
653 | awk "-F$flagmarker" -f $AWKSCRIPT $CROSSILLEGAL > $CROSSROOTS |
---|
654 | if [ "$debug" = yes ] |
---|
655 | then |
---|
656 | /bin/rm -f $CROSSEXPANDED $CROSSPAIRS $CROSSILLEGAL |
---|
657 | fi |
---|
658 | dbnum=`expr $dbnum + 1` |
---|
659 | done |
---|
660 | /bin/rm -f $CROSSEXPANDED $CROSSPAIRS $CROSSILLEGAL $AWKSCRIPT |
---|
661 | # |
---|
662 | # Now we have, in ILLEGALCOMBOS, a list of root/flag combinations |
---|
663 | # that must be removed from LEGALFLAGLIST to get the final list |
---|
664 | # of truly legal flags. ILLEGALCOMBOS has one flag per line, so |
---|
665 | # by turning LEGALFLAGLIST into this form (sed), it's an |
---|
666 | # easy task for comm. We have to recombine flags again after the |
---|
667 | # extraction, to get all flags for a given root on the same line so that |
---|
668 | # cross-products will come out right. |
---|
669 | # |
---|
670 | if [ -s $ILLEGALCOMBOS ] |
---|
671 | then |
---|
672 | sort $SORTTMP -u -o $ILLEGALCOMBOS $ILLEGALCOMBOS |
---|
673 | $verbose && echo 'Finding roots of cross expansions.' 1>&2 |
---|
674 | sort $SORTTMP $LEGALFLAGLIST \ |
---|
675 | | sed '/\/../{ |
---|
676 | s;^\(.*\)/\(.\)\(.*\);\1/\2\ |
---|
677 | \1/\3; |
---|
678 | P |
---|
679 | D |
---|
680 | }' \ |
---|
681 | | comm -23 - $ILLEGALCOMBOS \ |
---|
682 | | sort $SORTTMP -u "-t$flagmarker" +0f -1 +0 \ |
---|
683 | | $COMBINE $langtabs > $CROSSROOTS |
---|
684 | mv $CROSSROOTS $LEGALFLAGLIST |
---|
685 | if [ "$debug" = yes ] |
---|
686 | then |
---|
687 | rm -f ${TDIR}/LEGALFLAGLIST1 |
---|
688 | ln $LEGALFLAGLIST ${TDIR}/LEGALFLAGLIST1 |
---|
689 | fi |
---|
690 | fi |
---|
691 | fi |
---|
692 | /bin/rm -f $PRODUCTLIST $CROSSROOTS $ILLEGALCOMBOS $EXPANDEDINPUT |
---|
693 | # |
---|
694 | |
---|
695 | # We now have (in LEGALFLAGLIST) a list of roots and flags which will |
---|
696 | # accept words taken from EXPANDEDINPUT and no others (though some of |
---|
697 | # EXPANDEDINPUT is not covered by this list). However, many of the |
---|
698 | # expanded words can be generated in more than one way. For example, |
---|
699 | # "bather" can be generated from "bath/R" and "bathe/R". This wastes |
---|
700 | # unnecessary space in the raw dictionary and, in some cases, in the |
---|
701 | # hash file as well. The solution is to list the various ways of |
---|
702 | # getting a given word and choose exactly one. All other things being |
---|
703 | # equal, we want to choose the one with the highest expansion length |
---|
704 | # to root length ratio. The ispell -e4 option takes care of this by |
---|
705 | # providing us with a field to sort on. |
---|
706 | # |
---|
707 | # The ispell/awk combination is similar to the ispell/sed pipe used to |
---|
708 | # generate EXPANDEDPAIRS, except that ispell adds an extra field |
---|
709 | # giving the sort order. The first sort gets things in order so the |
---|
710 | # first root listed is the one we want, and the second sort (-um) then |
---|
711 | # selects that first root. Sed strips the expansion from the root, |
---|
712 | # and a final sort -u generates MINIMALAFFIXES, the final list of |
---|
713 | # affixes that (more or less) minimally covers what it can from |
---|
714 | # EXPANDEDINPUT. |
---|
715 | # |
---|
716 | $verbose && echo 'Eliminating non-optimal affixes.' 1>&2 |
---|
717 | ispell "$wchars" -e4 -d $FAKEHASH -p /dev/null < $LEGALFLAGLIST \ |
---|
718 | | sort $SORTTMP +1 -2 +2rn -3 +0 -1 \ |
---|
719 | | sort $SORTTMP -um +1 -2 \ |
---|
720 | | sed -e 's; .*$;;' \ |
---|
721 | | sort $SORTTMP -u "-t$flagmarker" +0f -1 +0 > $MINIMALAFFIXES |
---|
722 | /bin/rm -f $LEGALFLAGLIST |
---|
723 | # |
---|
724 | # Now we're almost done. MINIMALAFFIXES covers some (with luck, most) |
---|
725 | # of the words in STRIPPEDINPUT. Now we must create a list of the remaining |
---|
726 | # words (those omitted by MINIMALAFFIXES) and add it to MINIMALAFFIXES. |
---|
727 | # The best way to do this is to actually build a partial dictionary from |
---|
728 | # MINIMALAFFIXES in FAKEHASH, and then use ispell -l to list the words that |
---|
729 | # are not covered by this dictionary. This must then be combined with the |
---|
730 | # reduced version of MINIMALAFFIXES and sorted to produce the final result. |
---|
731 | # |
---|
732 | $verbose && echo "Generating output word list." 1>&2 |
---|
733 | if [ -s $MINIMALAFFIXES ] |
---|
734 | then |
---|
735 | buildhash -s $MINIMALAFFIXES $langtabs $FAKEHASH > /dev/null \ |
---|
736 | || (echo "Couldn't create intermediate hash file" 1>&2; |
---|
737 | /bin/rm -f ${TMP}*; |
---|
738 | exit 1) \ |
---|
739 | || exit 1 |
---|
740 | if [ "$debug" = yes ] |
---|
741 | then |
---|
742 | rm -f ${TDIR}/MINAFFIXES.!!COUNTSUFFIX!! \ |
---|
743 | ${TDIR}/MINAFFIXES!!STATSUFFIX!! |
---|
744 | ln $MINIMALAFFIXES.!!COUNTSUFFIX!! ${TDIR}/MINAFFIXES.!!COUNTSUFFIX!! |
---|
745 | ln $MINIMALAFFIXES!!STATSUFFIX!! ${TDIR}/MINAFFIXES!!STATSUFFIX!! |
---|
746 | fi |
---|
747 | (ispell "$wchars" -l -d $FAKEHASH -p /dev/null < $STRIPPEDINPUT; \ |
---|
748 | $COMBINE $langtabs < $MINIMALAFFIXES) \ |
---|
749 | | sort $SORTTMP "-t$flagmarker" -u +0f -1 +0 |
---|
750 | else |
---|
751 | # MINIMALAFFIXES is empty; just produce a sorted version of STRIPPEDINPUT |
---|
752 | sort $SORTTMP "-t$flagmarker" -u +0f -1 +0 $STRIPPEDINPUT |
---|
753 | fi |
---|
754 | /bin/rm -f ${TMP}* |
---|
755 | if [ "X$MUNCHMAIL" != X ] |
---|
756 | then |
---|
757 | ( |
---|
758 | ls -ld ${TDIR}/[A-Z]* |
---|
759 | cat /tmp/munchlist.mail |
---|
760 | ) | mail "$MUNCHMAIL" |
---|
761 | /bin/rm -f /tmp/munchlist.mail |
---|
762 | fi |
---|
763 | exit 0 |
---|