1 |
/* ============================================================================ |
2 |
* Douglas Thrift's Search Engine License |
3 |
* |
4 |
* Copyright (C) 2002, Douglas Thrift. All Rights Reserved. |
5 |
* Redistribution and use in source and binary forms, with or without |
6 |
* modification, are permitted provided that the following conditions are met: |
7 |
* |
8 |
* 1. Redistributions of source code must retain the above copyright notice, |
9 |
* this list of conditions and the following disclaimer. |
10 |
* |
11 |
* 2. Redistributions in binary form must reproduce the above copyright notice, |
12 |
* this list of conditions and the following disclaimer in the documentation |
13 |
* and/or other materials provided with the distribution. |
14 |
* |
15 |
* 3. The end-user documentation included with the redistribution, if any, must |
16 |
* include the following acknowledgment: |
17 |
* |
18 |
* "This product includes software developed by Douglas Thrift |
19 |
* (http://computers.douglasthrift.net/searchengine/)." |
20 |
* |
21 |
* Alternately, this acknowledgment may appear in the software itself, if |
22 |
* and wherever such third-party acknowledgments normally appear. |
23 |
* |
24 |
* 4. The names "Douglas Thrift" and "Douglas Thrift's Search Engine" must not |
25 |
* be used to endorse or promote products derived from this software without |
26 |
* specific prior written permission. For written permission, please visit |
27 |
* http://www.douglasthrift.net/contact.cgi for contact information. |
28 |
* |
29 |
* 5. Products derived from this software may not be called "Douglas Thrift's |
30 |
* Search Engine", nor may "Douglas Thrift's Search Engine" appear in their |
31 |
* name, without prior written permission. |
32 |
* |
33 |
* THIS SOFTWARE IS PROVIDED "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, |
34 |
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND |
35 |
* FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE |
36 |
* COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, |
37 |
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT |
38 |
* LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, |
39 |
* OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF |
40 |
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING |
41 |
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, |
42 |
* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
43 |
* ============================================================================ |
44 |
*/ |
45 |
// Douglas Thrift's Search Engine Outputer |
46 |
// |
47 |
// Douglas Thrift |
48 |
// |
49 |
// Outputer.cpp |
50 |
|
51 |
#include "Outputer.h" |
52 |
|
53 |
Outputer::Outputer(string& headerFile, string& bodyFile, string& footerFile, |
54 |
string& notfoundFile, string& pagesFile) |
55 |
{ |
56 |
this->headerFile = headerFile; |
57 |
this->bodyFile = bodyFile; |
58 |
this->footerFile = footerFile; |
59 |
this->notfoundFile = notfoundFile; |
60 |
this->pagesFile = pagesFile; |
61 |
} |
62 |
|
63 |
void Outputer::output(Searcher& searcher, unsigned page) |
64 |
{ |
65 |
MultiSet pagesSet = searcher.getPages(); |
66 |
numWebpages = pagesSet.size(); |
67 |
numPages = (numWebpages + 9) / 10; |
68 |
string query = searcher.getQueryString(); |
69 |
vector<string> common = searcher.getCommonUsed(); |
70 |
|
71 |
MultiSetIterator itor = pagesSet.begin(); |
72 |
|
73 |
for (int count = 0; count < page * 10 && itor != pagesSet.end(); count++) |
74 |
{ |
75 |
itor++; |
76 |
} |
77 |
|
78 |
for (int index = 0; index < 10 && itor != pagesSet.end(); index++, itor++) |
79 |
{ |
80 |
webpages.push_back(*itor); |
81 |
} |
82 |
|
83 |
this->query = searcher.getQuery().size() > 0; |
84 |
results = webpages.size() > 0; |
85 |
time = searcher.time(); |
86 |
|
87 |
if (debug) |
88 |
{ |
89 |
cerr << "query = " << (this->query ? "true" : "false") << "\n" |
90 |
<< "results = " << (results ? "true" : "false") << "\n" |
91 |
<< "time = " << duration() << "\n"; |
92 |
} |
93 |
|
94 |
entities(query, '&', "&"); |
95 |
entities(query, '\"', """); |
96 |
entities(query, '<', "<"); |
97 |
entities(query, '>', ">"); |
98 |
|
99 |
string ignore = searcher.getIgnore(); |
100 |
|
101 |
header(query, page, common, searcher.getAnd(), searcher.getOr(), |
102 |
ignore); |
103 |
|
104 |
if (results) |
105 |
{ |
106 |
body(); |
107 |
} |
108 |
else if (this->query) |
109 |
{ |
110 |
notfound(query, searcher.getQuery().size()); |
111 |
} |
112 |
|
113 |
footer(query, page, common, searcher.getAnd(), searcher.getOr(), |
114 |
ignore); |
115 |
} |
116 |
|
117 |
void Outputer::header(string& query, unsigned page, vector<string> common, |
118 |
bool and_, bool or_, string& ignore) |
119 |
{ |
120 |
ifstream fin(headerFile.c_str()); |
121 |
|
122 |
string line; |
123 |
while (fin.good()) |
124 |
{ |
125 |
getline(fin, line); |
126 |
|
127 |
conditional(line, fin, "<?ifquery?>", this->query); |
128 |
conditional(line, fin, "<?ifresults?>", results); |
129 |
conditional(line, fin, "<?ifor?>", or_); |
130 |
conditional(line, fin, "<?ifand?>", and_); |
131 |
conditional(line, fin, "<?ifignore?>", ignore != ""); |
132 |
conditional(line, fin, "<?ifcommon?>", common.size() == 1); |
133 |
conditional(line, fin, "<?ifmanycommon?>", common.size() > 1); |
134 |
|
135 |
tag(line, "<?query?>", query); |
136 |
|
137 |
string fred = range(page); |
138 |
|
139 |
tag(line, "<?range?>", fred); |
140 |
|
141 |
string martha = total(); |
142 |
|
143 |
tag(line, "<?total?>", martha); |
144 |
|
145 |
string george = duration(); |
146 |
|
147 |
tag(line, "<?time?>", george); |
148 |
|
149 |
string derf = pages(query, page); |
150 |
|
151 |
tag(line, "<?pages?>", derf); |
152 |
tag(line, "<?ignore?>", ignore); |
153 |
tag(line, "<?common?>", common[0]); |
154 |
|
155 |
string elaine = manycommon(common); |
156 |
|
157 |
tag(line, "<?manycommon?>", elaine); |
158 |
|
159 |
cout << line << "\n"; |
160 |
} |
161 |
|
162 |
fin.close(); |
163 |
} |
164 |
|
165 |
void Outputer::body() |
166 |
{ |
167 |
for (int index = 0; index < webpages.size(); index++) |
168 |
{ |
169 |
Ranker webpage = webpages[index]; |
170 |
string title = webpage.getTitle(); |
171 |
if (title == "") |
172 |
{ |
173 |
title = webpage.getURL(); |
174 |
entities(title, '&', "&"); |
175 |
entities(title, '\"', """); |
176 |
entities(title, '<', "<"); |
177 |
entities(title, '>', ">"); |
178 |
} |
179 |
string address = webpage.getURL(); |
180 |
string sample = webpage.getSample(); |
181 |
string description = webpage.getDescription(); |
182 |
|
183 |
char* csize = new char[1024]; |
184 |
sprintf(csize, "%.0fk", (double(webpage.getSize()) / double(1024))); |
185 |
|
186 |
string size = csize; |
187 |
|
188 |
delete [] csize; |
189 |
|
190 |
entities(address, '&', "&"); |
191 |
entities(address, '\"', """); |
192 |
entities(address, '<', "<"); |
193 |
entities(address, '>', ">"); |
194 |
|
195 |
ifstream fin(bodyFile.c_str()); |
196 |
|
197 |
string line; |
198 |
while (fin.good()) |
199 |
{ |
200 |
getline(fin, line); |
201 |
|
202 |
conditional(line, fin, "<?ifdescription?>", description != ""); |
203 |
|
204 |
tag(line, "<?address?>", address); |
205 |
tag(line, "<?title?>", title); |
206 |
tag(line, "<?sample?>", sample); |
207 |
tag(line, "<?description?>", description); |
208 |
tag(line, "<?size?>", size); |
209 |
|
210 |
cout << line << "\n"; |
211 |
} |
212 |
|
213 |
fin.close(); |
214 |
} |
215 |
} |
216 |
|
217 |
void Outputer::footer(string& query, unsigned page, vector<string> common, |
218 |
bool and_, bool or_, string& ignore) |
219 |
{ |
220 |
ifstream fin(footerFile.c_str()); |
221 |
|
222 |
string line; |
223 |
while (fin.good()) |
224 |
{ |
225 |
getline(fin, line); |
226 |
|
227 |
conditional(line, fin, "<?ifquery?>", this->query); |
228 |
conditional(line, fin, "<?ifresults?>", results); |
229 |
conditional(line, fin, "<?ifor?>", or_); |
230 |
conditional(line, fin, "<?ifand?>", and_); |
231 |
conditional(line, fin, "<?ifignore?>", ignore != ""); |
232 |
conditional(line, fin, "<?ifcommon?>", common.size() == 1); |
233 |
conditional(line, fin, "<?ifmanycommon?>", common.size() > 1); |
234 |
|
235 |
tag(line, "<?query?>", query); |
236 |
|
237 |
string fred = range(page); |
238 |
|
239 |
tag(line, "<?range?>", fred); |
240 |
|
241 |
string martha = total(); |
242 |
|
243 |
tag(line, "<?total?>", martha); |
244 |
|
245 |
string george = duration(); |
246 |
|
247 |
tag(line, "<?time?>", george); |
248 |
|
249 |
string derf = pages(query, page); |
250 |
|
251 |
tag(line, "<?pages?>", derf); |
252 |
tag(line, "<?ignore?>", ignore); |
253 |
tag(line, "<?common?>", common[0]); |
254 |
|
255 |
string elaine = manycommon(common); |
256 |
|
257 |
tag(line, "<?manycommon?>", elaine); |
258 |
|
259 |
cout << line << "\n"; |
260 |
} |
261 |
|
262 |
fin.close(); |
263 |
} |
264 |
|
265 |
void Outputer::notfound(string& query, unsigned keywords) |
266 |
{ |
267 |
ifstream fin(notfoundFile.c_str()); |
268 |
|
269 |
string line; |
270 |
while (fin.good()) |
271 |
{ |
272 |
getline(fin, line); |
273 |
|
274 |
conditional(line, fin, "<?ifmany?>", keywords > 1); |
275 |
|
276 |
tag(line, "<?query?>", query); |
277 |
|
278 |
cout << line << "\n"; |
279 |
} |
280 |
|
281 |
fin.close(); |
282 |
} |
283 |
|
284 |
string Outputer::pages(string query, unsigned page) |
285 |
{ |
286 |
entities(query, "<", '<'); |
287 |
entities(query, ">", '>'); |
288 |
entities(query, """, '\"'); |
289 |
entities(query, "&", '&'); |
290 |
|
291 |
entities(query, '%', "%25"); |
292 |
entities(query, '\t', "%09"); |
293 |
entities(query, ' ', "%20"); |
294 |
entities(query, '\"', "%22"); |
295 |
entities(query, '#', "%23"); |
296 |
entities(query, '$', "%24"); |
297 |
entities(query, '&', "%26"); |
298 |
entities(query, '\'', "%27"); |
299 |
entities(query, '+', "%2B"); |
300 |
entities(query, ',', "%2C"); |
301 |
entities(query, '/', "%2F"); |
302 |
entities(query, ':', "%3A"); |
303 |
entities(query, ';', "%3B"); |
304 |
entities(query, '<', "%3C"); |
305 |
entities(query, '=', "%3D"); |
306 |
entities(query, '>', "%3E"); |
307 |
entities(query, '?', "%3F"); |
308 |
entities(query, '@', "%40"); |
309 |
entities(query, '[', "%5B"); |
310 |
entities(query, ']', "%5D"); |
311 |
entities(query, '\\', "%5C"); |
312 |
entities(query, '^', "%5E"); |
313 |
entities(query, '`', "%60"); |
314 |
entities(query, '{', "%7B"); |
315 |
entities(query, '|', "%7C"); |
316 |
entities(query, '}', "%7D"); |
317 |
entities(query, '~', "%7E"); |
318 |
|
319 |
string lines; |
320 |
|
321 |
ifstream fin(pagesFile.c_str()); |
322 |
|
323 |
string line; |
324 |
while (fin.good()) |
325 |
{ |
326 |
getline(fin, line); |
327 |
conditional(line, fin, "<?ifprevious?>", page >= 1); |
328 |
conditional(line, fin, "<?ifpage?>", false); |
329 |
conditional(line, fin, "<?ifnum?>", false); |
330 |
conditional(line, fin, "<?ifnext?>", false); |
331 |
|
332 |
char* cprevious = new char[1024]; |
333 |
|
334 |
sprintf(cprevious, "%u", page); |
335 |
|
336 |
string previous = cprevious; |
337 |
|
338 |
delete [] cprevious; |
339 |
|
340 |
tag(line, "<?query?>", query); |
341 |
tag(line, "<?previous?>", previous); |
342 |
|
343 |
lines += line + "\n"; |
344 |
} |
345 |
|
346 |
fin.close(); |
347 |
fin.clear(); |
348 |
|
349 |
for (int index = 0; index < numPages; index++) |
350 |
{ |
351 |
fin.open(pagesFile.c_str()); |
352 |
|
353 |
while (fin.good()) |
354 |
{ |
355 |
getline(fin, line); |
356 |
if (index == page) |
357 |
{ |
358 |
conditional(line, fin, "<?ifprevious?>", false); |
359 |
conditional(line, fin, "<?ifpage?>", true); |
360 |
conditional(line, fin, "<?ifnum?>", false); |
361 |
conditional(line, fin, "<?ifnext?>", false); |
362 |
|
363 |
char* cpage = new char[1024]; |
364 |
|
365 |
sprintf(cpage, "%u", (index + 1)); |
366 |
|
367 |
string spage = cpage; |
368 |
|
369 |
delete [] cpage; |
370 |
|
371 |
tag(line, "<?page?>", spage); |
372 |
} |
373 |
else |
374 |
{ |
375 |
conditional(line, fin, "<?ifprevious?>", false); |
376 |
conditional(line, fin, "<?ifpage?>", false); |
377 |
conditional(line, fin, "<?ifnum?>", true); |
378 |
conditional(line, fin, "<?ifnext?>", false); |
379 |
|
380 |
char* cnum = new char[1024]; |
381 |
|
382 |
sprintf(cnum, "%u", (index + 1)); |
383 |
|
384 |
string num = cnum; |
385 |
|
386 |
delete [] cnum; |
387 |
|
388 |
tag(line, "<?query?>", query); |
389 |
tag(line, "<?num?>", num); |
390 |
} |
391 |
|
392 |
lines += line + "\n"; |
393 |
} |
394 |
|
395 |
fin.close(); |
396 |
fin.clear(); |
397 |
} |
398 |
|
399 |
fin.open(pagesFile.c_str()); |
400 |
|
401 |
while (fin.good()) |
402 |
{ |
403 |
getline(fin, line); |
404 |
conditional(line, fin, "<?ifprevious?>", false); |
405 |
conditional(line, fin, "<?ifpage?>", false); |
406 |
conditional(line, fin, "<?ifnum?>", false); |
407 |
conditional(line, fin, "<?ifnext?>", page + 2 <= numPages); |
408 |
|
409 |
char* cnext = new char[1024]; |
410 |
|
411 |
sprintf(cnext, "%u", (page + 2)); |
412 |
|
413 |
string next = cnext; |
414 |
|
415 |
delete [] cnext; |
416 |
|
417 |
tag(line, "<?query?>", query); |
418 |
tag(line, "<?next?>", next); |
419 |
|
420 |
lines += line + "\n"; |
421 |
} |
422 |
|
423 |
fin.close(); |
424 |
|
425 |
return lines; |
426 |
} |
427 |
|
428 |
string Outputer::range(unsigned page) |
429 |
{ |
430 |
unsigned bottom = page * 10 + 1; |
431 |
unsigned top = numWebpages > page * 10 + 10 ? page * 10 + 10 : numWebpages; |
432 |
|
433 |
char* cbottom = new char[1024]; |
434 |
char* ctop = new char[1024]; |
435 |
|
436 |
sprintf(cbottom, "%u", bottom); |
437 |
sprintf(ctop, "%u", top); |
438 |
|
439 |
string range = string(cbottom) + " - " + ctop; |
440 |
|
441 |
delete [] cbottom; |
442 |
delete [] ctop; |
443 |
|
444 |
return range; |
445 |
} |
446 |
|
447 |
string Outputer::total() |
448 |
{ |
449 |
char* ctotal = new char[1024]; |
450 |
|
451 |
sprintf(ctotal, "%u", numWebpages); |
452 |
|
453 |
string total = ctotal; |
454 |
|
455 |
delete [] ctotal; |
456 |
|
457 |
return total; |
458 |
} |
459 |
|
460 |
string Outputer::duration() |
461 |
{ |
462 |
char* ctime = new char[1024]; |
463 |
sprintf(ctime, "%.2f", time); |
464 |
|
465 |
string duration = ctime; |
466 |
|
467 |
delete [] ctime; |
468 |
|
469 |
return duration; |
470 |
} |
471 |
|
472 |
string Outputer::manycommon(vector<string> common) |
473 |
{ |
474 |
string line; |
475 |
|
476 |
for (int index = 0; index < common.size(); index++) |
477 |
{ |
478 |
line += common[index]; |
479 |
|
480 |
if (index != common.size() - 1) line += ' '; |
481 |
} |
482 |
|
483 |
return line; |
484 |
} |
485 |
|
486 |
void Outputer::tag(string& line, char* tag, string& replacement) |
487 |
{ |
488 |
int begin = 0; |
489 |
while (begin < line.length()) |
490 |
{ |
491 |
int spot = line.find(tag, begin); |
492 |
|
493 |
if (spot != string::npos) |
494 |
{ |
495 |
line.replace(spot, strlen(tag), replacement); |
496 |
} |
497 |
else |
498 |
{ |
499 |
break; |
500 |
} |
501 |
|
502 |
begin = spot + replacement.length(); |
503 |
} |
504 |
} |
505 |
|
506 |
void Outputer::conditional(string& line, ifstream& fin, char* tag, bool |
507 |
condition) |
508 |
{ |
509 |
unsigned begin = 0; |
510 |
while (begin < line.length()) |
511 |
{ |
512 |
unsigned start = line.find(tag, begin); |
513 |
unsigned finish = line.find("<?endif?>", start); |
514 |
|
515 |
if (start == string::npos) break; |
516 |
|
517 |
string next; |
518 |
while (finish == string::npos) |
519 |
{ |
520 |
getline(fin, next); |
521 |
line += '\n' + next; |
522 |
finish = line.find("<?endif?>", start); |
523 |
} |
524 |
|
525 |
if (condition) |
526 |
{ |
527 |
line.erase(start, strlen(tag)); |
528 |
line.erase(finish - strlen(tag), 9); |
529 |
|
530 |
begin = finish - strlen(tag) - 9; |
531 |
} |
532 |
else |
533 |
{ |
534 |
line.erase(start, finish - start + 9); |
535 |
|
536 |
begin = start; |
537 |
} |
538 |
} |
539 |
} |