1 |
/* ============================================================================ |
2 |
* Douglas Thrift's Search Engine License |
3 |
* |
4 |
* Copyright (C) 2002-2004, 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 Ranker |
46 |
// |
47 |
// Douglas Thrift |
48 |
// |
49 |
// $Id$ |
50 |
|
51 |
#include "Ranker.hpp" |
52 |
|
53 |
Ranker::Ranker(Page& page) : Page(page) |
54 |
{ |
55 |
value = 0; |
56 |
requiredValue = 0; |
57 |
excludedValue = 0; |
58 |
eitherOrValue = 0; |
59 |
allIn = all; |
60 |
} |
61 |
|
62 |
void Ranker::rank(vector<string> query) |
63 |
{ |
64 |
vector<string> prep; |
65 |
|
66 |
for (unsigned index = 0; index < query.size(); index++) |
67 |
{ |
68 |
if (query[index] == "allintitle:" && index == 0) |
69 |
{ |
70 |
allIn = title; |
71 |
} |
72 |
else if (query[index] == "allinurl:" && index == 0) |
73 |
{ |
74 |
allIn = url; |
75 |
} |
76 |
else if (query[index] == "allintext:" && index == 0) |
77 |
{ |
78 |
allIn = text; |
79 |
} |
80 |
else if (query[index].find("site:") == 0 && query[index].size() > 5) |
81 |
{ |
82 |
site = query[index].substr(5); |
83 |
} |
84 |
else if (query[index].find("intitle:") == 0 && query[index].size() > 8) |
85 |
{ |
86 |
prep.push_back("TITLE " + query[index].substr(8)); |
87 |
} |
88 |
else if (query[index].find("inurl:") == 0 && query[index].size() > 6) |
89 |
{ |
90 |
prep.push_back("URL " + query[index].substr(6)); |
91 |
} |
92 |
else if (query[index].find("intext:") == 0 && query[index].size() > 7) |
93 |
{ |
94 |
prep.push_back("TEXT " + query[index].substr(7)); |
95 |
} |
96 |
else |
97 |
{ |
98 |
prep.push_back(query[index]); |
99 |
} |
100 |
} |
101 |
|
102 |
if (prep.size() > 0) |
103 |
{ |
104 |
bool or_ = false; |
105 |
for (unsigned index = 0; index < prep.size(); index++) |
106 |
{ |
107 |
bool exclude = false; |
108 |
if (prep[index].find('+') == 0) |
109 |
{ |
110 |
prep[index].erase(0, 1); |
111 |
} |
112 |
else if (prep[index].find('-') == 0) |
113 |
{ |
114 |
exclude = true; |
115 |
prep[index].erase(0, 1); |
116 |
} |
117 |
|
118 |
if (or_) |
119 |
{ |
120 |
if (prep[index].find(" OR") == string::npos) |
121 |
{ |
122 |
or_ = false; |
123 |
} |
124 |
|
125 |
eitherOr[eitherOr.size() - 1] += ' ' + prep[index]; |
126 |
} |
127 |
else if (exclude) |
128 |
{ |
129 |
excluded.push_back(prep[index]); |
130 |
} |
131 |
else if (prep[index].find(" OR") != string::npos) |
132 |
{ |
133 |
or_ = true; |
134 |
eitherOr.push_back(prep[index]); |
135 |
} |
136 |
else |
137 |
{ |
138 |
required.push_back(prep[index]); |
139 |
} |
140 |
} |
141 |
} |
142 |
|
143 |
rank(); |
144 |
} |
145 |
|
146 |
void Ranker::setSample() |
147 |
{ |
148 |
map<unsigned, unsigned>::iterator itor; |
149 |
multimap<unsigned, map<unsigned, unsigned>::iterator> distances; |
150 |
|
151 |
for (itor = occurrencesText.begin(); itor != occurrencesText.end(); itor++) |
152 |
{ |
153 |
unsigned distance; |
154 |
|
155 |
if (++itor != occurrencesText.end()) |
156 |
{ |
157 |
unsigned next = itor->first; |
158 |
|
159 |
itor--; |
160 |
distance = next - (itor->first + itor->second); |
161 |
} |
162 |
else |
163 |
{ |
164 |
distance = string::npos; |
165 |
itor--; |
166 |
} |
167 |
|
168 |
distances.insert(pair<unsigned, map<unsigned, |
169 |
unsigned>::iterator>(distance, itor)); |
170 |
} |
171 |
|
172 |
if (distances.begin() != distances.end()) |
173 |
{ |
174 |
itor = distances.begin()->second; |
175 |
} |
176 |
|
177 |
string portion; |
178 |
unsigned sampleLength = 0, begin = 0, end = string::npos; |
179 |
|
180 |
while (sampleLength < sampleMax && itor != occurrencesText.end()) |
181 |
{ |
182 |
unsigned found = itor->first, length = itor->second; |
183 |
|
184 |
for (unsigned index = found; index > begin; index--) |
185 |
{ |
186 |
if (found - index >= sampleMax - sampleLength - length) |
187 |
{ |
188 |
for (; index < found; index++) |
189 |
{ |
190 |
if (isspace(getText()[index])) break; |
191 |
} |
192 |
|
193 |
begin = index + 1; |
194 |
|
195 |
break; |
196 |
} |
197 |
else if ((index > begin ? (isupper(getText()[index]) && |
198 |
!isalnum(getText()[index - 1])) : isupper(getText()[index])) && |
199 |
index != found) |
200 |
{ |
201 |
begin = index; |
202 |
|
203 |
break; |
204 |
} |
205 |
} |
206 |
|
207 |
if (end + 1 != begin) sample += " <strong>...</strong> "; |
208 |
|
209 |
portion = getText().substr(begin, found - begin); |
210 |
sampleLength += portion.length(); |
211 |
|
212 |
entities(portion, '&', "&"); |
213 |
entities(portion, '\"', """); |
214 |
entities(portion, '<', "<"); |
215 |
entities(portion, '>', ">"); |
216 |
|
217 |
sample += portion + "<strong>"; |
218 |
portion = getText().substr(found, length); |
219 |
sampleLength += portion.length(); |
220 |
|
221 |
entities(portion, '&', "&"); |
222 |
entities(portion, '\"', """); |
223 |
entities(portion, '<', "<"); |
224 |
entities(portion, '>', ">"); |
225 |
|
226 |
sample += portion + "</strong>"; |
227 |
begin = found + length; |
228 |
end = begin - 1; |
229 |
|
230 |
if (++itor != occurrencesText.end()) |
231 |
{ |
232 |
if (itor->first + itor->second < begin + sampleMax - sampleLength) |
233 |
{ |
234 |
portion = getText().substr(begin, itor->first - begin); |
235 |
sampleLength += portion.length(); |
236 |
|
237 |
entities(portion, '&', "&"); |
238 |
entities(portion, '\"', """); |
239 |
entities(portion, '<', "<"); |
240 |
entities(portion, '>', ">"); |
241 |
|
242 |
sample += portion; |
243 |
begin = itor->first; |
244 |
end = begin - 1; |
245 |
} |
246 |
else |
247 |
{ |
248 |
for (end = begin + sampleMax - sampleLength; end > begin; |
249 |
end--) |
250 |
{ |
251 |
if (isspace(getText()[end])) break; |
252 |
} |
253 |
|
254 |
portion = getText().substr(begin, end - begin + 1); |
255 |
sampleLength += portion.length(); |
256 |
|
257 |
entities(portion, '&', "&"); |
258 |
entities(portion, '\"', """); |
259 |
entities(portion, '<', "<"); |
260 |
entities(portion, '>', ">"); |
261 |
|
262 |
sample += portion + " <strong>...</strong>"; |
263 |
|
264 |
break; |
265 |
} |
266 |
} |
267 |
else |
268 |
{ |
269 |
for (end = begin + sampleMax - sampleLength; end > begin && (end + |
270 |
1 < getText().length()); end--) |
271 |
{ |
272 |
if (isspace(getText()[end])) break; |
273 |
} |
274 |
|
275 |
if (end >= getText().length()) end = getText().length() - 1; |
276 |
|
277 |
portion = getText().substr(begin, end - begin + 1); |
278 |
sampleLength += portion.length(); |
279 |
|
280 |
entities(portion, '&', "&"); |
281 |
entities(portion, '\"', """); |
282 |
entities(portion, '<', "<"); |
283 |
entities(portion, '>', ">"); |
284 |
|
285 |
sample += portion; |
286 |
|
287 |
if (end + 1 < getText().length()) |
288 |
{ |
289 |
sample += " <strong>...</strong>"; |
290 |
} |
291 |
|
292 |
break; |
293 |
} |
294 |
} |
295 |
|
296 |
if (sample == "") |
297 |
{ |
298 |
for (end = sampleMax; end > 0 && (end + 1 < getText().length()); end--) |
299 |
{ |
300 |
if (isspace(getText()[end])) break; |
301 |
} |
302 |
|
303 |
sample = getText().substr(0, end + 1); |
304 |
|
305 |
entities(sample, '&', "&"); |
306 |
entities(sample, '\"', """); |
307 |
entities(sample, '<', "<"); |
308 |
entities(sample, '>', ">"); |
309 |
|
310 |
if (end + 1 < getText().length()) |
311 |
{ |
312 |
sample += " <strong>...</strong>"; |
313 |
} |
314 |
} |
315 |
} |
316 |
|
317 |
string Ranker::getTitle() |
318 |
{ |
319 |
string title, portion; |
320 |
|
321 |
unsigned begin = 0; |
322 |
for (map<unsigned, unsigned>::iterator itor = occurrencesTitle.begin(); |
323 |
itor != occurrencesTitle.end(); itor++) |
324 |
{ |
325 |
unsigned found = itor->first; |
326 |
unsigned length = itor->second; |
327 |
|
328 |
portion = Page::getTitle().substr(begin, found - begin); |
329 |
|
330 |
entities(portion, '&', "&"); |
331 |
entities(portion, '\"', """); |
332 |
entities(portion, '<', "<"); |
333 |
entities(portion, '>', ">"); |
334 |
|
335 |
title += portion + "<strong>"; |
336 |
|
337 |
portion = Page::getTitle().substr(found, length); |
338 |
|
339 |
entities(portion, '&', "&"); |
340 |
entities(portion, '\"', """); |
341 |
entities(portion, '<', "<"); |
342 |
entities(portion, '>', ">"); |
343 |
|
344 |
title += portion + "</strong>"; |
345 |
|
346 |
begin = found + length; |
347 |
} |
348 |
|
349 |
portion = Page::getTitle().substr(begin); |
350 |
|
351 |
entities(portion, '&', "&"); |
352 |
entities(portion, '\"', """); |
353 |
entities(portion, '<', "<"); |
354 |
entities(portion, '>', ">"); |
355 |
|
356 |
title += portion; |
357 |
|
358 |
return title; |
359 |
} |
360 |
|
361 |
string Ranker::getDescription() |
362 |
{ |
363 |
string description, portion; |
364 |
|
365 |
unsigned begin = 0; |
366 |
for (map<unsigned, unsigned>::iterator itor = |
367 |
occurrencesDescription.begin(); itor != occurrencesDescription.end(); |
368 |
itor++) |
369 |
{ |
370 |
unsigned found = itor->first; |
371 |
unsigned length = itor->second; |
372 |
|
373 |
portion = Page::getDescription().substr(begin, found - begin); |
374 |
|
375 |
entities(portion, '&', "&"); |
376 |
entities(portion, '\"', """); |
377 |
entities(portion, '<', "<"); |
378 |
entities(portion, '>', ">"); |
379 |
|
380 |
description += portion + "<strong>"; |
381 |
|
382 |
portion = Page::getDescription().substr(found, length); |
383 |
|
384 |
entities(portion, '&', "&"); |
385 |
entities(portion, '\"', """); |
386 |
entities(portion, '<', "<"); |
387 |
entities(portion, '>', ">"); |
388 |
|
389 |
description += portion + "</strong>"; |
390 |
|
391 |
begin = found + length; |
392 |
} |
393 |
|
394 |
portion = Page::getDescription().substr(begin); |
395 |
|
396 |
entities(portion, '&', "&"); |
397 |
entities(portion, '\"', """); |
398 |
entities(portion, '<', "<"); |
399 |
entities(portion, '>', ">"); |
400 |
|
401 |
description += portion; |
402 |
|
403 |
return description; |
404 |
} |
405 |
|
406 |
bool Ranker::operator==(const unsigned number) const |
407 |
{ |
408 |
return value == number; |
409 |
} |
410 |
|
411 |
bool Ranker::operator==(const Ranker& ranker) const |
412 |
{ |
413 |
return value == ranker.value; |
414 |
} |
415 |
|
416 |
bool Ranker::operator!=(const unsigned number) const |
417 |
{ |
418 |
return value != number; |
419 |
} |
420 |
|
421 |
bool Ranker::operator!=(const Ranker& ranker) const |
422 |
{ |
423 |
return value != ranker.value; |
424 |
} |
425 |
|
426 |
bool Ranker::operator<(const unsigned number) const |
427 |
{ |
428 |
return value < number; |
429 |
} |
430 |
|
431 |
bool Ranker::operator<(const Ranker& ranker) const |
432 |
{ |
433 |
return value < ranker.value; |
434 |
} |
435 |
|
436 |
bool Ranker::operator>(const unsigned number) const |
437 |
{ |
438 |
return value > number; |
439 |
} |
440 |
|
441 |
bool Ranker::operator >(const Ranker& ranker) const |
442 |
{ |
443 |
return value > ranker.value; |
444 |
} |
445 |
|
446 |
void Ranker::rank() |
447 |
{ |
448 |
lowerAddress = tolower(getAddress()); |
449 |
|
450 |
if (site == "" || lowerAddress.rfind(site) == lowerAddress.length() - |
451 |
site.length()) |
452 |
{ |
453 |
bool isRequired = required.size() > 0; |
454 |
bool isExcluded = excluded.size() > 0; |
455 |
bool isEitherOr = eitherOr.size() > 0; |
456 |
|
457 |
lowerURL = tolower(getURL()); |
458 |
lowerTitle = tolower(Page::getTitle()); |
459 |
lowerText = tolower(Page::getText()); |
460 |
|
461 |
if (isRequired) checkRequired(); |
462 |
if (isExcluded && (isRequired || isEitherOr)) checkExcluded(); |
463 |
if (isEitherOr) checkEitherOr(); |
464 |
|
465 |
if (isRequired && isExcluded && isEitherOr) |
466 |
{ |
467 |
value += requiredValue && !excludedValue && eitherOrValue ? |
468 |
requiredValue + eitherOrValue : 0; |
469 |
} |
470 |
else if (isRequired && isExcluded) |
471 |
{ |
472 |
value += requiredValue && !excludedValue ? requiredValue : 0; |
473 |
} |
474 |
else if (isRequired && isEitherOr) |
475 |
{ |
476 |
value += requiredValue && eitherOrValue ? requiredValue + |
477 |
eitherOrValue : 0; |
478 |
} |
479 |
else if (isExcluded && isEitherOr) |
480 |
{ |
481 |
value += !excludedValue && eitherOrValue ? eitherOrValue : 0; |
482 |
} |
483 |
else if (isRequired) |
484 |
{ |
485 |
value += requiredValue; |
486 |
} |
487 |
else if (isEitherOr) |
488 |
{ |
489 |
value += eitherOrValue; |
490 |
} |
491 |
else |
492 |
{ |
493 |
// do nothing this is a bad search and warrants no results |
494 |
} |
495 |
|
496 |
if (value > 0) |
497 |
{ |
498 |
string lowerDescription = tolower(Page::getDescription()); |
499 |
|
500 |
for (unsigned index = 0; index < required.size(); index++) |
501 |
{ |
502 |
if (required[index].find("URL ") == 0) |
503 |
{ |
504 |
value += find(required[index].substr(4), lowerDescription, |
505 |
occurrencesDescription); |
506 |
} |
507 |
else if (required[index].find("TITLE ") == 0) |
508 |
{ |
509 |
value += find(required[index].substr(6), lowerDescription, |
510 |
occurrencesDescription); |
511 |
} |
512 |
else if (required[index].find("TEXT ") == 0) |
513 |
{ |
514 |
value += find(required[index].substr(5), lowerDescription, |
515 |
occurrencesDescription); |
516 |
} |
517 |
else |
518 |
{ |
519 |
value += find(required[index], lowerDescription, |
520 |
occurrencesDescription); |
521 |
} |
522 |
} |
523 |
|
524 |
for (unsigned index1 = 0; index1 < eitherOr.size(); index1++) |
525 |
{ |
526 |
vector<string> words; |
527 |
|
528 |
unsigned begin = 0, found; |
529 |
do |
530 |
{ |
531 |
found = eitherOr[index1].find(" OR ", begin); |
532 |
|
533 |
if (found != string::npos) |
534 |
{ |
535 |
words.push_back(eitherOr[index1].substr(begin, found - |
536 |
begin)); |
537 |
} |
538 |
else |
539 |
{ |
540 |
words.push_back(eitherOr[index1].substr(begin)); |
541 |
} |
542 |
|
543 |
begin = found + 4; |
544 |
} |
545 |
while (begin < eitherOr[index1].length() && found != |
546 |
string::npos); |
547 |
|
548 |
for (unsigned number = 0; number < words.size(); number++) |
549 |
{ |
550 |
if (words[index1].find("URL ") == 0) |
551 |
{ |
552 |
value += find(words[index1].substr(4), |
553 |
lowerDescription, occurrencesDescription); |
554 |
} |
555 |
else if (words[index1].find("TITLE ") == 0) |
556 |
{ |
557 |
value += find(words[index1].substr(6), |
558 |
lowerDescription, occurrencesDescription); |
559 |
} |
560 |
else if (words[index1].find("TEXT ") == 0) |
561 |
{ |
562 |
value += find(words[index1].substr(5), |
563 |
lowerDescription, occurrencesDescription); |
564 |
} |
565 |
else |
566 |
{ |
567 |
value += find(words[index1], lowerDescription, |
568 |
occurrencesDescription); |
569 |
} |
570 |
} |
571 |
} |
572 |
|
573 |
for (unsigned index2 = 0; index2 < getHeadings().size(); index2++) |
574 |
{ |
575 |
string lowerHeading = string(getHeadings()[index2].length(), |
576 |
' '); |
577 |
for (unsigned number = 0; number < |
578 |
getHeadings()[index2].length(); number++) |
579 |
{ |
580 |
lowerHeading[number] = tolower( |
581 |
getHeadings()[index2][number]); |
582 |
} |
583 |
|
584 |
for (unsigned number0 = 0; number0 < required.size(); number0++) |
585 |
{ |
586 |
if (required[number0].find("URL ") == 0) |
587 |
{ |
588 |
value += find(required[number0].substr(4), |
589 |
lowerHeading); |
590 |
} |
591 |
else if (required[number0].find("TITLE ") == 0) |
592 |
{ |
593 |
value += find(required[number0].substr(6), |
594 |
lowerHeading); |
595 |
} |
596 |
else if (required[number0].find("TEXT ") == 0) |
597 |
{ |
598 |
value += find(required[number0].substr(5), |
599 |
lowerHeading); |
600 |
} |
601 |
else |
602 |
{ |
603 |
value += find(required[number0], lowerHeading); |
604 |
} |
605 |
} |
606 |
|
607 |
for (unsigned number1 = 0; number1 < eitherOr.size(); number1++) |
608 |
{ |
609 |
vector<string> words; |
610 |
|
611 |
unsigned begin = 0, found; |
612 |
do |
613 |
{ |
614 |
found = eitherOr[number1].find(" OR ", begin); |
615 |
|
616 |
if (found != string::npos) |
617 |
{ |
618 |
words.push_back(eitherOr[number1].substr(begin, |
619 |
found - begin)); |
620 |
} |
621 |
else |
622 |
{ |
623 |
words.push_back(eitherOr[number1].substr(begin)); |
624 |
} |
625 |
|
626 |
begin = found + 4; |
627 |
} |
628 |
while (begin < eitherOr[number1].length() && found != |
629 |
string::npos); |
630 |
|
631 |
for (unsigned number = 0; number < words.size(); number++) |
632 |
{ |
633 |
if (words[number].find("URL ") == 0) |
634 |
{ |
635 |
value += find(words[number].substr(4), |
636 |
lowerHeading); |
637 |
} |
638 |
else if (words[number].find("TITLE ") == 0) |
639 |
{ |
640 |
value += find(words[number].substr(6), |
641 |
lowerHeading); |
642 |
} |
643 |
else if (words[number].find("TEXT ") == 0) |
644 |
{ |
645 |
value += find(words[number].substr(5), |
646 |
lowerHeading); |
647 |
} |
648 |
else |
649 |
{ |
650 |
value += find(words[number], lowerHeading); |
651 |
} |
652 |
} |
653 |
} |
654 |
} |
655 |
} |
656 |
} |
657 |
} |
658 |
|
659 |
void Ranker::checkRequired() |
660 |
{ |
661 |
vector<unsigned> inURLs, inTitles, inTexts; |
662 |
|
663 |
for (unsigned index = 0; index < required.size(); index++) |
664 |
{ |
665 |
unsigned inURL = 0, inTitle = 0, inText = 0; |
666 |
|
667 |
if (required[index].find("URL ") == 0) |
668 |
{ |
669 |
inURL = find(required[index].substr(4), lowerURL.substr(7)); |
670 |
|
671 |
if (inURL) |
672 |
{ |
673 |
inTitle = find(required[index].substr(4), lowerTitle, |
674 |
occurrencesTitle); |
675 |
inText = find(required[index].substr(4), lowerText, |
676 |
occurrencesText); |
677 |
|
678 |
if (!inTitle) inTitle++; |
679 |
if (!inText) inText++; |
680 |
} |
681 |
} |
682 |
else if (required[index].find("TITLE ") == 0) |
683 |
{ |
684 |
inTitle = find(required[index].substr(6), lowerTitle, |
685 |
occurrencesTitle); |
686 |
|
687 |
if (inTitle) |
688 |
{ |
689 |
inURL = find(required[index].substr(6), lowerURL.substr(7)); |
690 |
inText = find(required[index].substr(6), lowerText, |
691 |
occurrencesText); |
692 |
|
693 |
if (!inURL) inURL++; |
694 |
if (!inText) inText++; |
695 |
} |
696 |
} |
697 |
else if (required[index].find("TEXT ") == 0) |
698 |
{ |
699 |
inText = find(required[index].substr(5), lowerText, |
700 |
occurrencesText); |
701 |
|
702 |
if (inText) |
703 |
{ |
704 |
inURL = find(required[index].substr(5), lowerURL.substr(7)); |
705 |
inTitle = find(required[index].substr(5), lowerTitle, |
706 |
occurrencesTitle); |
707 |
|
708 |
if (!inURL) inURL++; |
709 |
if (!inTitle) inTitle++; |
710 |
} |
711 |
} |
712 |
else |
713 |
{ |
714 |
inURL = find(required[index], lowerURL.substr(7)); |
715 |
inTitle = find(required[index], lowerTitle, occurrencesTitle); |
716 |
inText = find(required[index], lowerText, occurrencesText); |
717 |
} |
718 |
|
719 |
inURLs.push_back(inURL); |
720 |
inTitles.push_back(inTitle); |
721 |
inTexts.push_back(inText); |
722 |
} |
723 |
|
724 |
unsigned inURL = evaluate(inURLs); |
725 |
unsigned inTitle = evaluate(inTitles); |
726 |
unsigned inText = evaluate(inTexts); |
727 |
|
728 |
requiredValue += (inURL && (allIn == url)) || (inTitle && (allIn == title)) |
729 |
|| (inText && ((allIn == text) || (allIn == all))) ? inURL + inTitle + |
730 |
inText : 0; |
731 |
} |
732 |
|
733 |
void Ranker::checkExcluded() |
734 |
{ |
735 |
vector<unsigned> inURLs, inTitles, inTexts; |
736 |
|
737 |
for (unsigned index = 0; index < excluded.size(); index++) |
738 |
{ |
739 |
unsigned inURL = 0, inTitle = 0, inText = 0; |
740 |
|
741 |
inURL = find(excluded[index], lowerURL.substr(7)); |
742 |
inTitle = find(excluded[index], lowerTitle); |
743 |
inText = find(excluded[index], lowerText); |
744 |
|
745 |
inURLs.push_back(inURL); |
746 |
inTitles.push_back(inTitle); |
747 |
inTexts.push_back(inText); |
748 |
} |
749 |
|
750 |
unsigned inURL = evaluate(inURLs); |
751 |
unsigned inTitle = evaluate(inTitles); |
752 |
unsigned inText = evaluate(inTexts); |
753 |
|
754 |
excludedValue += (inURL && (allIn == url)) || (inTitle && (allIn == title)) |
755 |
|| (inText && ((allIn == text) || (allIn == all))) ? inURL + inTitle + |
756 |
inText : 0; |
757 |
} |
758 |
|
759 |
void Ranker::checkEitherOr() |
760 |
{ |
761 |
vector<unsigned> inURLs, inTitles, inTexts; |
762 |
|
763 |
for (unsigned index = 0; index < eitherOr.size(); index++) |
764 |
{ |
765 |
vector<unsigned> inURLz, inTitlez, inTextz; |
766 |
unsigned inURL = 0, inTitle = 0, inText = 0; |
767 |
vector<string> words; |
768 |
|
769 |
unsigned begin = 0, found; |
770 |
do |
771 |
{ |
772 |
found = eitherOr[index].find(" OR ", begin); |
773 |
|
774 |
if (found != string::npos) |
775 |
{ |
776 |
words.push_back(eitherOr[index].substr(begin, found - begin)); |
777 |
} |
778 |
else |
779 |
{ |
780 |
words.push_back(eitherOr[index].substr(begin)); |
781 |
} |
782 |
|
783 |
begin = found + 4; |
784 |
} |
785 |
while (begin < eitherOr[index].length() && found != string::npos); |
786 |
|
787 |
for (unsigned number = 0; number < words.size(); number++) |
788 |
{ |
789 |
unsigned inURL = 0, inTitle = 0, inText = 0; |
790 |
|
791 |
if (words[number].find("URL ") == 0) |
792 |
{ |
793 |
inURL = find(words[number].substr(4), lowerURL.substr(7)); |
794 |
|
795 |
if (inURL) |
796 |
{ |
797 |
inTitle = find(words[number].substr(4), lowerTitle, |
798 |
occurrencesTitle); |
799 |
inText = find(words[number].substr(4), lowerText, |
800 |
occurrencesText); |
801 |
|
802 |
if (!inTitle) inTitle++; |
803 |
if (!inText) inText++; |
804 |
} |
805 |
} |
806 |
else if (words[number].find("TITLE ") == 0) |
807 |
{ |
808 |
inTitle = find(words[number].substr(6), lowerTitle, |
809 |
occurrencesTitle); |
810 |
|
811 |
if (inTitle) |
812 |
{ |
813 |
inURL = find(words[number].substr(6), lowerURL.substr(7)); |
814 |
inText = find(words[number].substr(6), lowerText, |
815 |
occurrencesText); |
816 |
|
817 |
if (!inURL) inURL++; |
818 |
if (!inText) inText++; |
819 |
} |
820 |
} |
821 |
else if (words[number].find("TEXT ") == 0) |
822 |
{ |
823 |
inText = find(words[number].substr(5), lowerText, |
824 |
occurrencesText); |
825 |
|
826 |
if (inText) |
827 |
{ |
828 |
inURL = find(words[number].substr(5), lowerURL.substr(7)); |
829 |
inTitle = find(words[number].substr(5), lowerTitle, |
830 |
occurrencesTitle); |
831 |
|
832 |
if (!inURL) inURL++; |
833 |
if (!inTitle) inTitle++; |
834 |
} |
835 |
} |
836 |
else |
837 |
{ |
838 |
inURL = find(words[number], lowerURL.substr(7)); |
839 |
inTitle = find(words[number], lowerTitle, occurrencesTitle); |
840 |
inText = find(words[number], lowerText, occurrencesText); |
841 |
} |
842 |
|
843 |
inURLz.push_back(inURL); |
844 |
inTitlez.push_back(inTitle); |
845 |
inTextz.push_back(inText); |
846 |
} |
847 |
|
848 |
for (unsigned number0 = 0; number0 < inURLz.size(); number0++) |
849 |
{ |
850 |
inURL += inURLz[number0]; |
851 |
} |
852 |
|
853 |
for (unsigned number1 = 0; number1 < inTitlez.size(); number1++) |
854 |
{ |
855 |
inTitle += inTitlez[number1]; |
856 |
} |
857 |
|
858 |
for (unsigned number2 = 0; number2 < inTextz.size(); number2++) |
859 |
{ |
860 |
inText += inTextz[number2]; |
861 |
} |
862 |
|
863 |
inURLs.push_back(inURL); |
864 |
inTitles.push_back(inTitle); |
865 |
inTexts.push_back(inText); |
866 |
|
867 |
inURLz.clear(); |
868 |
inTitlez.clear(); |
869 |
inTextz.clear(); |
870 |
words.clear(); |
871 |
} |
872 |
|
873 |
unsigned inURL = evaluate(inURLs); |
874 |
unsigned inTitle = evaluate(inTitles); |
875 |
unsigned inText = evaluate(inTexts); |
876 |
|
877 |
eitherOrValue += (inURL && (allIn == url)) || (inTitle && (allIn == title)) |
878 |
|| (inText && ((allIn == text) || (allIn == all))) ? inURL + inTitle + |
879 |
inText : 0; |
880 |
} |
881 |
|
882 |
unsigned Ranker::find(string word, const string& where) |
883 |
{ |
884 |
unsigned value = 0; |
885 |
|
886 |
decrap(word); |
887 |
|
888 |
if (word == "") |
889 |
{ |
890 |
// this can happen if a word is all crap characters |
891 |
value++; |
892 |
} |
893 |
else if (word.find_first_of(" \n ") == string::npos) |
894 |
{ |
895 |
unsigned begin = 0, found; |
896 |
do |
897 |
{ |
898 |
found = where.find(word, begin); |
899 |
|
900 |
if (found != string::npos) |
901 |
{ |
902 |
bool isBefore, isAfter, before = false, after = false; |
903 |
isBefore = found - 1 > 0; |
904 |
isAfter = found + word.length() < where.length(); |
905 |
|
906 |
if (isBefore) before = isalnum(where[found - 1]) != 0; |
907 |
if (isAfter) after = isalnum(where[found + word.length()]) != 0; |
908 |
|
909 |
if (!before && !after) |
910 |
{ |
911 |
value++; |
912 |
} |
913 |
} |
914 |
|
915 |
begin = found + word.length(); |
916 |
} |
917 |
while (found != string::npos && begin < where.length()); |
918 |
} |
919 |
else |
920 |
{ |
921 |
value = phrase(word, where); |
922 |
} |
923 |
|
924 |
return value; |
925 |
} |
926 |
|
927 |
unsigned Ranker::find(string word, const string& where, map<unsigned, |
928 |
unsigned>& occurrences) |
929 |
{ |
930 |
unsigned value = 0; |
931 |
|
932 |
decrap(word); |
933 |
|
934 |
if (word == "") |
935 |
{ |
936 |
// this can happen if a word is all crap characters |
937 |
value++; |
938 |
} |
939 |
else if (word.find_first_of(" \n ") == string::npos) |
940 |
{ |
941 |
unsigned begin = 0, found; |
942 |
do |
943 |
{ |
944 |
found = where.find(word, begin); |
945 |
|
946 |
if (found != string::npos) |
947 |
{ |
948 |
bool isBefore, isAfter, before = false, after = false; |
949 |
isBefore = found - 1 > 0; |
950 |
isAfter = found + word.length() < where.length(); |
951 |
|
952 |
if (isBefore) before = isalnum(where[found - 1]) != 0; |
953 |
if (isAfter) after = isalnum(where[found + word.length()]) != 0; |
954 |
|
955 |
if (!before && !after) |
956 |
{ |
957 |
value++; |
958 |
|
959 |
occurrences.insert(pair<unsigned, unsigned>(found, |
960 |
word.length())); |
961 |
} |
962 |
} |
963 |
|
964 |
begin = found + word.length(); |
965 |
} |
966 |
while (found != string::npos && begin < where.length()); |
967 |
} |
968 |
else |
969 |
{ |
970 |
value = phrase(word, where, occurrences); |
971 |
} |
972 |
|
973 |
return value; |
974 |
} |
975 |
|
976 |
unsigned Ranker::phrase(const string& phrase, const string& where) |
977 |
{ |
978 |
unsigned value = 0; |
979 |
vector<string> words; |
980 |
|
981 |
unsigned begin = 0, space; |
982 |
do |
983 |
{ |
984 |
space = phrase.find(' ', begin); |
985 |
|
986 |
words.push_back(phrase.substr(begin, space - begin)); |
987 |
|
988 |
begin = space + 1; |
989 |
} |
990 |
while (space != string::npos && begin < phrase.length()); |
991 |
|
992 |
begin = 0; |
993 |
unsigned counter = 0; |
994 |
do |
995 |
{ |
996 |
value += this->phrase(words, 0, begin, true, where); |
997 |
} |
998 |
while (begin < where.length()); |
999 |
|
1000 |
return value; |
1001 |
} |
1002 |
|
1003 |
unsigned Ranker::phrase(const string& phrase, const string& where, |
1004 |
map<unsigned, unsigned>& occurrences) |
1005 |
{ |
1006 |
unsigned value = 0; |
1007 |
vector<string> words; |
1008 |
|
1009 |
unsigned begin = 0, space; |
1010 |
do |
1011 |
{ |
1012 |
space = phrase.find(' ', begin); |
1013 |
|
1014 |
words.push_back(phrase.substr(begin, space - begin)); |
1015 |
|
1016 |
begin = space + 1; |
1017 |
} |
1018 |
while (space != string::npos && begin < phrase.length()); |
1019 |
|
1020 |
begin = 0; |
1021 |
do |
1022 |
{ |
1023 |
value += this->phrase(words, 0, begin, true, where, occurrences); |
1024 |
} |
1025 |
while (begin < where.length()); |
1026 |
|
1027 |
return value; |
1028 |
} |
1029 |
|
1030 |
unsigned Ranker::phrase(const vector<string>& words, unsigned word, unsigned& |
1031 |
begin, bool start, const string& where) |
1032 |
{ |
1033 |
unsigned value = 0; |
1034 |
bool end = !(word + 1 < words.size()); |
1035 |
unsigned found = where.find(words[word], begin); |
1036 |
unsigned newBegin = found + words[word].length(); |
1037 |
|
1038 |
if (found != string::npos) |
1039 |
{ |
1040 |
bool isBefore, isAfter, before = false, after = false; |
1041 |
isBefore = found - 1 > 0; |
1042 |
isAfter = found + words[word].length() < where.length(); |
1043 |
|
1044 |
if (isBefore) before = isalnum(where[found - 1]) != 0; |
1045 |
if (isAfter) after = isalnum(where[found + words[word].length()]) != 0; |
1046 |
|
1047 |
if (!before && !after) |
1048 |
{ |
1049 |
bool between = true; |
1050 |
if (!start) |
1051 |
{ |
1052 |
for (unsigned index = begin + 1; index < found - 1; index++) |
1053 |
{ |
1054 |
if (isalnum(where[index])) |
1055 |
{ |
1056 |
between = false; |
1057 |
break; |
1058 |
} |
1059 |
} |
1060 |
} |
1061 |
|
1062 |
if (between) |
1063 |
{ |
1064 |
if (end) |
1065 |
{ |
1066 |
begin = newBegin; |
1067 |
value = 1; |
1068 |
} |
1069 |
else |
1070 |
{ |
1071 |
value = phrase(words, (word + 1), newBegin, false, where); |
1072 |
} |
1073 |
} |
1074 |
} |
1075 |
} |
1076 |
|
1077 |
if (start) |
1078 |
{ |
1079 |
if (found != string::npos) |
1080 |
{ |
1081 |
begin = newBegin; |
1082 |
} |
1083 |
else |
1084 |
{ |
1085 |
begin = string::npos; |
1086 |
} |
1087 |
} |
1088 |
|
1089 |
return value; |
1090 |
} |
1091 |
|
1092 |
unsigned Ranker::phrase(const vector<string>& words, unsigned word, unsigned& |
1093 |
begin, bool start, const string& where, map<unsigned, unsigned>& |
1094 |
occurrences) |
1095 |
{ |
1096 |
unsigned value = 0; |
1097 |
bool end = !(word + 1 < words.size()); |
1098 |
unsigned found = where.find(words[word], begin); |
1099 |
unsigned newBegin = found + words[word].length(); |
1100 |
|
1101 |
if (found != string::npos) |
1102 |
{ |
1103 |
bool isBefore, isAfter, before = false, after = false; |
1104 |
isBefore = found - 1 > 0; |
1105 |
isAfter = found + words[word].length() < where.length(); |
1106 |
|
1107 |
if (isBefore) before = isalnum(where[found - 1]) != 0; |
1108 |
if (isAfter) after = isalnum(where[found + words[word].length()]) != 0; |
1109 |
|
1110 |
if (!before && !after) |
1111 |
{ |
1112 |
bool between = true; |
1113 |
if (!start) |
1114 |
{ |
1115 |
for (unsigned index = begin + 1; index < found - 1; index++) |
1116 |
{ |
1117 |
if (isalnum(where[index])) |
1118 |
{ |
1119 |
between = false; |
1120 |
break; |
1121 |
} |
1122 |
} |
1123 |
} |
1124 |
|
1125 |
if (between) |
1126 |
{ |
1127 |
occurrences.insert(pair<unsigned, unsigned>(found, |
1128 |
words[word].length())); |
1129 |
|
1130 |
if (end) |
1131 |
{ |
1132 |
begin = newBegin; |
1133 |
value = 1; |
1134 |
} |
1135 |
else |
1136 |
{ |
1137 |
value = phrase(words, (word + 1), newBegin, false, where, |
1138 |
occurrences); |
1139 |
} |
1140 |
} |
1141 |
} |
1142 |
} |
1143 |
|
1144 |
if (start) |
1145 |
{ |
1146 |
if (found != string::npos) |
1147 |
{ |
1148 |
begin = newBegin; |
1149 |
} |
1150 |
else |
1151 |
{ |
1152 |
begin = string::npos; |
1153 |
} |
1154 |
} |
1155 |
|
1156 |
return value; |
1157 |
} |
1158 |
|
1159 |
unsigned Ranker::evaluate(vector<unsigned>& ins) |
1160 |
{ |
1161 |
unsigned in = 0; |
1162 |
|
1163 |
for (unsigned index = 0; index < ins.size(); index++) |
1164 |
{ |
1165 |
if (ins[index] > 0) |
1166 |
{ |
1167 |
in += ins[index]; |
1168 |
} |
1169 |
else |
1170 |
{ |
1171 |
in = 0; |
1172 |
break; |
1173 |
} |
1174 |
} |
1175 |
|
1176 |
return in; |
1177 |
} |
1178 |
|
1179 |
void Ranker::decrap(string& crap) |
1180 |
{ |
1181 |
unsigned begin = 0, found; |
1182 |
do |
1183 |
{ |
1184 |
// &, _, +, and # are not considered crap |
1185 |
found = crap.find_first_of("!\"$%\'()*,-./:;<=>?@[\\]^`{|}~", begin); |
1186 |
|
1187 |
if (found != string::npos) |
1188 |
{ |
1189 |
crap[found] = ' '; |
1190 |
} |
1191 |
|
1192 |
begin = found + 1; |
1193 |
} |
1194 |
while (found != string::npos && begin < crap.length()); |
1195 |
|
1196 |
normalize(crap); |
1197 |
} |