1 |
/* -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8 -*- */ |
2 |
|
3 |
/* |
4 |
* Copyright (C) 2006 Ryo SHIMIZU |
5 |
* |
6 |
* This program is free software; you can redistribute it and/or modify |
7 |
* it under the terms of the GNU General Public License as published by |
8 |
* the Free Software Foundation; either version 2, or (at your option) |
9 |
* any later version. |
10 |
* |
11 |
* This program is distributed in the hope that it will be useful, |
12 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
13 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
14 |
* GNU General Public License for more details. |
15 |
* |
16 |
* You should have received a copy of the GNU General Public License |
17 |
* along with this program; if not, write to the Free Software |
18 |
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
19 |
*/ |
20 |
|
21 |
#include <string.h> |
22 |
#include <ctype.h> |
23 |
#include <glib/gi18n.h> |
24 |
|
25 |
#include "kazehakase.h" |
26 |
#include "utils/utils.h" |
27 |
#include "glib-utils.h" |
28 |
#include "rast-search.h" |
29 |
#include "egg-pixbuf-thumbnail.h" |
30 |
|
31 |
#define HISTORY_INDEX "/."PACKAGE"/history_index.rast" |
32 |
|
33 |
#define RAST_URI "http://projects.netlab.jp/rast/" |
34 |
#define DTD "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.01//EN\" \"http://www.w3.org/TR/html4/strict.dtd\">" |
35 |
#define HEAD "<head>\n" \ |
36 |
" <title>Full-text search in history</title>\n" \ |
37 |
" <link rel=\"stylesheet\" type=\"text/css\" href=\"history-search:?css=search-result.css\">\n" \ |
38 |
"</head>\n" |
39 |
#define HEADER "" |
40 |
#define CONTENT "<div class=\"content\">\n" \ |
41 |
" <div class=\"header\"><span class=\"title\"><a href=\"%s\">%s</a></span></div>\n" \ |
42 |
" <div class=\"summary\"><img src=\"%s\" class=\"thumbnail\">\n" \ |
43 |
" <span class=\"sentence\">%s</span>\n" \ |
44 |
" </div>\n" \ |
45 |
" <div class=\"footer\">\n" \ |
46 |
" <span class=\"uri\">%s</span>\n" \ |
47 |
" <span class=\"cache\"><a href=\"%s\">cache</a></span>\n" \ |
48 |
" <span class=\"date\">%s</span>\n" \ |
49 |
" </div>\n" \ |
50 |
"</div>\n" |
51 |
#define FOOTER "<div class=\"footer\">\n" \ |
52 |
"Powered by <a href=\"%s\">Rast</a> version %s\n" \ |
53 |
"</div>\n" |
54 |
|
55 |
static gchar *rast_get_version (void); |
56 |
gchar *get_value (const gchar *line); |
57 |
|
58 |
static gboolean |
59 |
rast_execute_search_command(const gchar *search_text, gint *standard_output) |
60 |
{ |
61 |
gboolean ret; |
62 |
const gchar *rast_cmd = "rast search "; |
63 |
gchar *command; |
64 |
gint argc; |
65 |
gchar **argv = NULL; |
66 |
GSpawnFlags flags; |
67 |
GPid pid; |
68 |
gint err; |
69 |
gchar **split = NULL; |
70 |
gchar *join = NULL; |
71 |
gint max_results = 20, num_summary = 128; |
72 |
gchar *except_word; |
73 |
gchar **except_keywords = NULL; |
74 |
|
75 |
KZ_CONF_GET("History", "num_summary", num_summary, INT); |
76 |
KZ_CONF_GET("History", "max_results", max_results, INT); |
77 |
|
78 |
split = g_strsplit(search_text, " ", -1); |
79 |
if (split) |
80 |
{ |
81 |
join = g_strjoinv(" & ", split); |
82 |
g_strfreev(split); |
83 |
} |
84 |
|
85 |
except_word = KZ_CONF_GET_STR("History", "except_keyword"); |
86 |
if (except_word && *except_word) |
87 |
{ |
88 |
except_keywords = g_strsplit(except_word, ",", -1); |
89 |
g_free(except_word); |
90 |
except_word = g_strjoinv(" - ", except_keywords); |
91 |
g_strfreev(except_keywords); |
92 |
|
93 |
command = g_strdup_printf("%s --num-items %d --summary-nchars %d '%s - %s' %s%s ", |
94 |
rast_cmd, |
95 |
max_results, |
96 |
num_summary, |
97 |
join, |
98 |
except_word, |
99 |
g_get_home_dir(), |
100 |
HISTORY_INDEX); |
101 |
g_free(except_word); |
102 |
} |
103 |
else |
104 |
{ |
105 |
command = g_strdup_printf("%s --num-items %d --summary-nchars %d '%s' %s%s", |
106 |
rast_cmd, |
107 |
max_results, |
108 |
num_summary, |
109 |
join, |
110 |
g_get_home_dir(), |
111 |
HISTORY_INDEX" "); |
112 |
} |
113 |
|
114 |
if (join) |
115 |
g_free(join); |
116 |
|
117 |
g_shell_parse_argv(command, |
118 |
&argc, |
119 |
&argv, |
120 |
NULL); |
121 |
|
122 |
flags = G_SPAWN_SEARCH_PATH; |
123 |
ret = g_spawn_async_with_pipes(NULL, |
124 |
argv, |
125 |
NULL, |
126 |
flags, |
127 |
NULL, |
128 |
NULL, |
129 |
&pid, |
130 |
NULL, |
131 |
standard_output, |
132 |
&err, |
133 |
NULL); |
134 |
g_strfreev(argv); |
135 |
g_free(command); |
136 |
|
137 |
return ret; |
138 |
} |
139 |
|
140 |
|
141 |
static gchar * |
142 |
rast_create_search_result_html (gint out, const gchar *text) |
143 |
{ |
144 |
GIOChannel *io; |
145 |
gchar *line; |
146 |
gsize length; |
147 |
gchar *title = NULL, *uri = NULL, *date = NULL, *desc = NULL; |
148 |
gchar *cache_link = NULL; |
149 |
gchar *rastversion = rast_get_version(); |
150 |
GString *html; |
151 |
|
152 |
io = g_io_channel_unix_new(out); |
153 |
g_io_channel_set_encoding(io, NULL, NULL); |
154 |
|
155 |
html = g_string_sized_new(0); |
156 |
|
157 |
g_string_append(html, DTD"\n"); |
158 |
g_string_append(html, "<html>\n"); |
159 |
g_string_append(html, HEAD); |
160 |
g_string_append(html, "<body>\n"); |
161 |
|
162 |
g_string_append_printf(html, "<h1>Search results for %s</h1>", |
163 |
text); |
164 |
|
165 |
while (g_io_channel_read_line(io, &line, &length, NULL, NULL) == G_IO_STATUS_NORMAL) |
166 |
{ |
167 |
if (g_str_has_prefix(line, "uri :")) |
168 |
{ |
169 |
size_t len; |
170 |
gchar *dirname = g_strconcat(g_get_home_dir(), |
171 |
HISTORY_DIR, |
172 |
NULL); |
173 |
len = strlen(dirname); |
174 |
|
175 |
cache_link = get_value(line); |
176 |
g_print("%s\n", cache_link); |
177 |
g_print("%s\n", dirname); |
178 |
uri = create_uri_from_filename(cache_link + |
179 |
strlen("file://") + |
180 |
len); |
181 |
g_free(dirname); |
182 |
} |
183 |
else if (g_str_has_prefix(line, "summary :")) |
184 |
{ |
185 |
gchar *thumb_filename, *thumb_uri; |
186 |
gchar *summary = get_value(line); |
187 |
|
188 |
desc = remove_tag(summary, g_strlen(summary)); |
189 |
thumb_filename = egg_pixbuf_get_thumb_filename(uri, |
190 |
EGG_PIXBUF_THUMB_LARGE); |
191 |
thumb_uri = g_strdup_printf("history-search:?image=%s", |
192 |
thumb_filename); |
193 |
g_string_append_printf(html, |
194 |
CONTENT, |
195 |
uri, |
196 |
title, |
197 |
thumb_uri, /* thumbnail */ |
198 |
desc, |
199 |
uri, |
200 |
cache_link, |
201 |
date); |
202 |
|
203 |
g_free(desc); |
204 |
g_free(title); |
205 |
g_free(uri); |
206 |
g_free(date); |
207 |
g_free(cache_link); |
208 |
g_free(summary); |
209 |
g_free(thumb_filename); |
210 |
g_free(thumb_uri); |
211 |
} |
212 |
else if (g_str_has_prefix(line, "title :")) |
213 |
{ |
214 |
title = get_value(line); |
215 |
} |
216 |
else if (g_str_has_prefix(line, "last_modified :")) |
217 |
{ |
218 |
date = get_value(line); |
219 |
} |
220 |
g_free(line); |
221 |
} |
222 |
g_io_channel_unref(io); |
223 |
g_string_append_printf(html, FOOTER, RAST_URI, rastversion); |
224 |
g_string_append(html, "</body></html>"); |
225 |
|
226 |
if (rastversion) |
227 |
g_free(rastversion); |
228 |
return g_string_free(html, FALSE); |
229 |
} |
230 |
|
231 |
|
232 |
gchar * |
233 |
rast_get_search_result (const gchar *text) |
234 |
{ |
235 |
gint out; |
236 |
|
237 |
if (!text) return NULL; |
238 |
if (!exists_search_cmd) return NULL; |
239 |
|
240 |
if (!rast_execute_search_command(text, &out)) |
241 |
return NULL; |
242 |
|
243 |
return rast_create_search_result_html(out, text); |
244 |
} |
245 |
|
246 |
gboolean |
247 |
rast_update_index (gpointer data) |
248 |
{ |
249 |
const gchar *rast_register = "rast register "; |
250 |
gchar *index_dir; |
251 |
gchar *command; |
252 |
gint argc; |
253 |
gchar **argv = NULL; |
254 |
GSpawnFlags flags; |
255 |
GPid pid; |
256 |
|
257 |
index_dir = g_strconcat(g_get_home_dir(), HISTORY_INDEX, NULL); |
258 |
|
259 |
command = g_strconcat(rast_register, |
260 |
index_dir, |
261 |
" ", |
262 |
(gchar*)data, |
263 |
NULL); |
264 |
g_free(index_dir); |
265 |
|
266 |
g_shell_parse_argv(command, |
267 |
&argc, |
268 |
&argv, |
269 |
NULL); |
270 |
|
271 |
flags = G_SPAWN_SEARCH_PATH | |
272 |
G_SPAWN_STDOUT_TO_DEV_NULL; |
273 |
g_spawn_async(NULL, |
274 |
argv, |
275 |
NULL, |
276 |
flags, |
277 |
NULL, |
278 |
NULL, |
279 |
&pid, |
280 |
NULL); |
281 |
g_strfreev(argv); |
282 |
g_free(command); |
283 |
|
284 |
g_free(data); |
285 |
|
286 |
return FALSE; |
287 |
} |
288 |
|
289 |
|
290 |
gboolean |
291 |
rast_purge_index (void) |
292 |
{ |
293 |
const gchar *estpurge = "rast delete"; |
294 |
gchar *command; |
295 |
gint argc; |
296 |
gchar **argv = NULL; |
297 |
GSpawnFlags flags; |
298 |
GPid pid; |
299 |
|
300 |
/* purge index */ |
301 |
flags = G_SPAWN_SEARCH_PATH | |
302 |
G_SPAWN_STDOUT_TO_DEV_NULL; |
303 |
command = g_strconcat(estpurge, |
304 |
g_get_home_dir(), |
305 |
HISTORY_INDEX, |
306 |
NULL); |
307 |
|
308 |
g_shell_parse_argv(command, |
309 |
&argc, |
310 |
&argv, |
311 |
NULL); |
312 |
flags = G_SPAWN_SEARCH_PATH | |
313 |
G_SPAWN_STDOUT_TO_DEV_NULL; |
314 |
|
315 |
g_spawn_async(NULL, |
316 |
argv, |
317 |
NULL, |
318 |
flags, |
319 |
NULL, |
320 |
NULL, |
321 |
&pid, |
322 |
NULL); |
323 |
|
324 |
g_strfreev(argv); |
325 |
g_free(command); |
326 |
|
327 |
return FALSE; |
328 |
} |
329 |
|
330 |
GPid |
331 |
rast_optimize_index (void) |
332 |
{ |
333 |
const gchar *estoptimize = "rast optimize "; |
334 |
gchar *command; |
335 |
gint argc; |
336 |
gchar **argv = NULL; |
337 |
GSpawnFlags flags; |
338 |
GPid pid; |
339 |
|
340 |
/* optimize index process */ |
341 |
command = g_strconcat(estoptimize, |
342 |
g_get_home_dir(), |
343 |
HISTORY_INDEX, |
344 |
NULL); |
345 |
|
346 |
g_shell_parse_argv(command, |
347 |
&argc, |
348 |
&argv, |
349 |
NULL); |
350 |
flags = G_SPAWN_SEARCH_PATH | |
351 |
G_SPAWN_STDOUT_TO_DEV_NULL; |
352 |
|
353 |
g_spawn_async(NULL, |
354 |
argv, |
355 |
NULL, |
356 |
flags, |
357 |
NULL, |
358 |
NULL, |
359 |
&pid, |
360 |
NULL); |
361 |
g_strfreev(argv); |
362 |
g_free(command); |
363 |
|
364 |
return pid; |
365 |
} |
366 |
|
367 |
static gchar* |
368 |
rast_get_version (void) |
369 |
{ |
370 |
gchar *version; |
371 |
const gchar *rastversion = "rast-config --version"; |
372 |
gint argc; |
373 |
gchar **argv = NULL; |
374 |
GSpawnFlags flags; |
375 |
GPid pid; |
376 |
gint out, err; |
377 |
gboolean ret; |
378 |
GIOChannel *io; |
379 |
gsize length; |
380 |
|
381 |
if (!exists_search_cmd) return NULL; |
382 |
|
383 |
g_shell_parse_argv(rastversion, |
384 |
&argc, |
385 |
&argv, |
386 |
NULL); |
387 |
|
388 |
flags = G_SPAWN_SEARCH_PATH; |
389 |
ret = g_spawn_async_with_pipes(NULL, |
390 |
argv, |
391 |
NULL, |
392 |
flags, |
393 |
NULL, |
394 |
NULL, |
395 |
&pid, |
396 |
NULL, |
397 |
&out, |
398 |
&err, |
399 |
NULL); |
400 |
g_strfreev(argv); |
401 |
if (!ret) return NULL; |
402 |
|
403 |
io = g_io_channel_unix_new(out); |
404 |
g_io_channel_set_encoding(io, NULL, NULL); |
405 |
g_io_channel_read_line(io, &version, &length, NULL, NULL); |
406 |
g_io_channel_shutdown(io, TRUE, NULL); |
407 |
g_io_channel_unref(io); |
408 |
|
409 |
return version; |
410 |
} |
411 |
|
412 |
static KzBookmark * |
413 |
rast_create_search_result_bookmark (gint out, const gchar *text) |
414 |
{ |
415 |
GIOChannel *io; |
416 |
gchar *line; |
417 |
gsize length; |
418 |
gchar *title = NULL, *uri = NULL, *desc = NULL; |
419 |
KzBookmark *result; |
420 |
|
421 |
io = g_io_channel_unix_new(out); |
422 |
g_io_channel_set_encoding(io, NULL, NULL); |
423 |
|
424 |
result = kz_bookmark_pure_folder_new(); |
425 |
|
426 |
while (g_io_channel_read_line(io, &line, &length, NULL, NULL) == G_IO_STATUS_NORMAL) |
427 |
{ |
428 |
if (g_str_has_prefix(line, "</document>")) |
429 |
{ |
430 |
KzBookmark *child; |
431 |
child = kz_bookmark_new_with_attrs(title, uri, desc); |
432 |
kz_bookmark_append(result, child); |
433 |
g_object_unref(child); |
434 |
g_free(desc); |
435 |
g_free(title); |
436 |
g_free(uri); |
437 |
} |
438 |
else if (g_str_has_prefix(line, "<uri>")) |
439 |
{ |
440 |
gchar *dirname, *orig_uri; |
441 |
gchar *link; |
442 |
size_t len; |
443 |
link = xml_get_attr(line, "uri"); |
444 |
dirname = g_strconcat(g_get_home_dir(), |
445 |
HISTORY_DIR, |
446 |
NULL); |
447 |
len = strlen(dirname); |
448 |
orig_uri = create_uri_from_filename(link + strlen("file://") + len); |
449 |
uri = url_decode(orig_uri); |
450 |
g_free(orig_uri); |
451 |
g_free(dirname); |
452 |
g_free(link); |
453 |
} |
454 |
else if (g_str_has_prefix(line, "<title>")) |
455 |
{ |
456 |
title = xml_get_content(line); |
457 |
} |
458 |
else if (g_str_has_prefix(line, "<summary")) |
459 |
{ |
460 |
gchar *summary = xml_get_content(line); |
461 |
desc = remove_tag(summary, g_strlen(summary)); |
462 |
g_free(summary); |
463 |
} |
464 |
g_free(line); |
465 |
} |
466 |
g_io_channel_unref(io); |
467 |
|
468 |
return result; |
469 |
} |
470 |
|
471 |
KzBookmark * |
472 |
rast_get_search_result_bookmark (const gchar *text) |
473 |
{ |
474 |
gint out; |
475 |
|
476 |
if (!text) return NULL; |
477 |
if (!exists_search_cmd) return NULL; |
478 |
|
479 |
if (!rast_execute_search_command(text, &out)) |
480 |
return NULL; |
481 |
|
482 |
return rast_create_search_result_bookmark(out, text); |
483 |
} |
484 |
|
485 |
void |
486 |
rast_make_index(void) |
487 |
{ |
488 |
const gchar *rast_create = "rast create --preserve-text --property=title:string:search "; |
489 |
gchar *command; |
490 |
gint argc; |
491 |
gchar **argv = NULL; |
492 |
GSpawnFlags flags; |
493 |
GPid pid; |
494 |
|
495 |
command = g_strconcat(rast_create, |
496 |
g_get_home_dir(), |
497 |
HISTORY_INDEX" ", |
498 |
g_get_home_dir(), |
499 |
HISTORY_DIR, |
500 |
NULL); |
501 |
|
502 |
g_shell_parse_argv(command, |
503 |
&argc, |
504 |
&argv, |
505 |
NULL); |
506 |
flags = G_SPAWN_SEARCH_PATH | |
507 |
G_SPAWN_STDOUT_TO_DEV_NULL; |
508 |
|
509 |
g_spawn_async(NULL, |
510 |
argv, |
511 |
NULL, |
512 |
flags, |
513 |
NULL, |
514 |
NULL, |
515 |
&pid, |
516 |
NULL); |
517 |
g_strfreev(argv); |
518 |
g_free(command); |
519 |
} |
520 |
|
521 |
gboolean |
522 |
rast_exist_index_dir(void) |
523 |
{ |
524 |
gchar *index_dir; |
525 |
gboolean exist = FALSE; |
526 |
|
527 |
index_dir = g_build_filename(g_get_home_dir(), |
528 |
HISTORY_INDEX, NULL); |
529 |
exist = g_file_test(index_dir, G_FILE_TEST_IS_DIR); |
530 |
g_free(index_dir); |
531 |
|
532 |
return exist; |
533 |
} |
534 |
|
535 |
gchar *get_value (const gchar *line) |
536 |
{ |
537 |
gchar *p; |
538 |
|
539 |
p = strchr(line, ':'); |
540 |
p += 2; |
541 |
|
542 |
return g_strchomp(g_strdup(p)); |
543 |
} |