1 |
/* ============================================================================ |
2 |
* Douglas Thrift's Search Engine License |
3 |
* |
4 |
* Copyright (C) 2002-2003, 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 |
// $Id: Outputer.cpp,v 1.10 2003/07/18 03:52:39 douglas Exp $ |
50 |
|
51 |
#include "Outputer.h" |
52 |
|
53 |
Outputer::Outputer(const string& headerFile, const string& bodyFile, const |
54 |
string& footerFile, const string& notfoundFile, const 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 << "\n" |
90 |
<< "results = " << results << "\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(const string& query, unsigned page, vector<string> |
118 |
common, bool and_, bool or_, const 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 |
#ifndef _OpenSSL_ |
136 |
tag(line, "<?version?>", programName + ' ' + programVersion + ' ' + |
137 |
platform()); |
138 |
#else |
139 |
tag(line, "<?version?>", programName + ' ' + programVersion + ' ' + |
140 |
platform() + ' ' + openssl()); |
141 |
#endif |
142 |
tag(line, "<?query?>", query); |
143 |
tag(line, "<?range?>", range(page)); |
144 |
tag(line, "<?total?>", total()); |
145 |
tag(line, "<?time?>", duration()); |
146 |
tag(line, "<?pages?>", pages(query, page)); |
147 |
tag(line, "<?ignore?>", ignore); |
148 |
tag(line, "<?common?>", common[0]); |
149 |
tag(line, "<?manycommon?>", manycommon(common)); |
150 |
|
151 |
cout << line << (fin.good() ? "\n" : ""); |
152 |
} |
153 |
|
154 |
fin.close(); |
155 |
} |
156 |
|
157 |
void Outputer::body() |
158 |
{ |
159 |
for (int index = 0; index < webpages.size(); index++) |
160 |
{ |
161 |
Ranker webpage = webpages[index]; |
162 |
string title = webpage.getTitle(); |
163 |
if (title == "") |
164 |
{ |
165 |
title = webpage.getURL(); |
166 |
entities(title, '&', "&"); |
167 |
entities(title, '\"', """); |
168 |
entities(title, '<', "<"); |
169 |
entities(title, '>', ">"); |
170 |
} |
171 |
string address = webpage.getURL(); |
172 |
string sample = webpage.getSample(); |
173 |
string description = webpage.getDescription(); |
174 |
ostringstream size; |
175 |
|
176 |
size.precision(1); |
177 |
size << double(webpage.getSize()) / double(1024) << "k"; |
178 |
|
179 |
entities(address, '&', "&"); |
180 |
entities(address, '\"', """); |
181 |
entities(address, '<', "<"); |
182 |
entities(address, '>', ">"); |
183 |
|
184 |
ifstream fin(bodyFile.c_str()); |
185 |
|
186 |
string line; |
187 |
while (fin.good()) |
188 |
{ |
189 |
getline(fin, line); |
190 |
|
191 |
conditional(line, fin, "<?ifdescription?>", description != ""); |
192 |
|
193 |
tag(line, "<?address?>", address); |
194 |
tag(line, "<?title?>", title); |
195 |
tag(line, "<?sample?>", sample); |
196 |
tag(line, "<?description?>", description); |
197 |
tag(line, "<?size?>", size.str()); |
198 |
|
199 |
cout << line << (fin.good() ? "\n" : ""); |
200 |
} |
201 |
|
202 |
fin.close(); |
203 |
} |
204 |
} |
205 |
|
206 |
void Outputer::footer(const string& query, unsigned page, vector<string> |
207 |
common, bool and_, bool or_, const string& ignore) |
208 |
{ |
209 |
ifstream fin(footerFile.c_str()); |
210 |
|
211 |
string line; |
212 |
while (fin.good()) |
213 |
{ |
214 |
getline(fin, line); |
215 |
|
216 |
conditional(line, fin, "<?ifquery?>", this->query); |
217 |
conditional(line, fin, "<?ifresults?>", results); |
218 |
conditional(line, fin, "<?ifor?>", or_); |
219 |
conditional(line, fin, "<?ifand?>", and_); |
220 |
conditional(line, fin, "<?ifignore?>", ignore != ""); |
221 |
conditional(line, fin, "<?ifcommon?>", common.size() == 1); |
222 |
conditional(line, fin, "<?ifmanycommon?>", common.size() > 1); |
223 |
|
224 |
#ifndef _OpenSSL_ |
225 |
tag(line, "<?version?>", programName + ' ' + programVersion + ' ' + |
226 |
platform()); |
227 |
#else |
228 |
tag(line, "<?version?>", programName + ' ' + programVersion + ' ' + |
229 |
platform() + ' ' + openssl()); |
230 |
#endif |
231 |
tag(line, "<?query?>", query); |
232 |
tag(line, "<?range?>", range(page)); |
233 |
tag(line, "<?total?>", total()); |
234 |
tag(line, "<?time?>", duration()); |
235 |
tag(line, "<?pages?>", pages(query, page)); |
236 |
tag(line, "<?ignore?>", ignore); |
237 |
tag(line, "<?common?>", common[0]); |
238 |
tag(line, "<?manycommon?>", manycommon(common)); |
239 |
|
240 |
cout << line << (fin.good() ? "\n" : ""); |
241 |
} |
242 |
|
243 |
fin.close(); |
244 |
} |
245 |
|
246 |
void Outputer::notfound(const string& query, unsigned keywords) |
247 |
{ |
248 |
ifstream fin(notfoundFile.c_str()); |
249 |
|
250 |
string line; |
251 |
while (fin.good()) |
252 |
{ |
253 |
getline(fin, line); |
254 |
|
255 |
conditional(line, fin, "<?ifmany?>", keywords > 1); |
256 |
|
257 |
tag(line, "<?query?>", query); |
258 |
|
259 |
cout << line << (fin.good() ? "\n" : ""); |
260 |
} |
261 |
|
262 |
fin.close(); |
263 |
} |
264 |
|
265 |
string Outputer::pages(string query, unsigned page) |
266 |
{ |
267 |
entities(query, "<", '<'); |
268 |
entities(query, ">", '>'); |
269 |
entities(query, """, '\"'); |
270 |
entities(query, "&", '&'); |
271 |
|
272 |
entities(query, '%', "%25"); |
273 |
entities(query, '\t', "%09"); |
274 |
entities(query, ' ', "%20"); |
275 |
entities(query, '\"', "%22"); |
276 |
entities(query, '#', "%23"); |
277 |
entities(query, '$', "%24"); |
278 |
entities(query, '&', "%26"); |
279 |
entities(query, '\'', "%27"); |
280 |
entities(query, '+', "%2B"); |
281 |
entities(query, ',', "%2C"); |
282 |
entities(query, '/', "%2F"); |
283 |
entities(query, ':', "%3A"); |
284 |
entities(query, ';', "%3B"); |
285 |
entities(query, '<', "%3C"); |
286 |
entities(query, '=', "%3D"); |
287 |
entities(query, '>', "%3E"); |
288 |
entities(query, '?', "%3F"); |
289 |
entities(query, '@', "%40"); |
290 |
entities(query, '[', "%5B"); |
291 |
entities(query, ']', "%5D"); |
292 |
entities(query, '\\', "%5C"); |
293 |
entities(query, '^', "%5E"); |
294 |
entities(query, '`', "%60"); |
295 |
entities(query, '{', "%7B"); |
296 |
entities(query, '|', "%7C"); |
297 |
entities(query, '}', "%7D"); |
298 |
entities(query, '~', "%7E"); |
299 |
|
300 |
string lines; |
301 |
|
302 |
ifstream fin(pagesFile.c_str()); |
303 |
|
304 |
string line; |
305 |
while (fin.good()) |
306 |
{ |
307 |
getline(fin, line); |
308 |
conditional(line, fin, "<?ifprevious?>", page >= 1); |
309 |
conditional(line, fin, "<?ifpage?>", false); |
310 |
conditional(line, fin, "<?ifnum?>", false); |
311 |
conditional(line, fin, "<?ifnext?>", false); |
312 |
|
313 |
ostringstream previous; |
314 |
|
315 |
previous << page; |
316 |
|
317 |
tag(line, "<?query?>", query); |
318 |
tag(line, "<?previous?>", previous.str()); |
319 |
|
320 |
lines += line + (fin.good() ? "\n" : ""); |
321 |
} |
322 |
|
323 |
fin.close(); |
324 |
fin.clear(); |
325 |
|
326 |
for (int index = 0; index < numPages; index++) |
327 |
{ |
328 |
fin.open(pagesFile.c_str()); |
329 |
|
330 |
while (fin.good()) |
331 |
{ |
332 |
getline(fin, line); |
333 |
if (index == page) |
334 |
{ |
335 |
conditional(line, fin, "<?ifprevious?>", false); |
336 |
conditional(line, fin, "<?ifpage?>", true); |
337 |
conditional(line, fin, "<?ifnum?>", false); |
338 |
conditional(line, fin, "<?ifnext?>", false); |
339 |
|
340 |
ostringstream current; |
341 |
|
342 |
current << index + 1; |
343 |
|
344 |
tag(line, "<?page?>", current.str()); |
345 |
} |
346 |
else |
347 |
{ |
348 |
conditional(line, fin, "<?ifprevious?>", false); |
349 |
conditional(line, fin, "<?ifpage?>", false); |
350 |
conditional(line, fin, "<?ifnum?>", true); |
351 |
conditional(line, fin, "<?ifnext?>", false); |
352 |
|
353 |
ostringstream num; |
354 |
|
355 |
num << index + 1; |
356 |
|
357 |
tag(line, "<?query?>", query); |
358 |
tag(line, "<?num?>", num.str()); |
359 |
} |
360 |
|
361 |
lines += line + (fin.good() ? "\n" : ""); |
362 |
} |
363 |
|
364 |
fin.close(); |
365 |
fin.clear(); |
366 |
} |
367 |
|
368 |
fin.open(pagesFile.c_str()); |
369 |
|
370 |
while (fin.good()) |
371 |
{ |
372 |
getline(fin, line); |
373 |
conditional(line, fin, "<?ifprevious?>", false); |
374 |
conditional(line, fin, "<?ifpage?>", false); |
375 |
conditional(line, fin, "<?ifnum?>", false); |
376 |
conditional(line, fin, "<?ifnext?>", page + 2 <= numPages); |
377 |
|
378 |
ostringstream next; |
379 |
|
380 |
next << page + 2; |
381 |
|
382 |
tag(line, "<?query?>", query); |
383 |
tag(line, "<?next?>", next.str()); |
384 |
|
385 |
lines += line + (fin.good() ? "\n" : ""); |
386 |
} |
387 |
|
388 |
fin.close(); |
389 |
|
390 |
return lines; |
391 |
} |
392 |
|
393 |
string Outputer::range(unsigned page) |
394 |
{ |
395 |
unsigned bottom = page * 10 + 1; |
396 |
unsigned top = numWebpages > page * 10 + 10 ? page * 10 + 10 : numWebpages; |
397 |
ostringstream range; |
398 |
|
399 |
range << bottom << " - " << top; |
400 |
|
401 |
return range.str(); |
402 |
} |
403 |
|
404 |
string Outputer::total() |
405 |
{ |
406 |
ostringstream total; |
407 |
|
408 |
total << numWebpages; |
409 |
|
410 |
return total.str(); |
411 |
} |
412 |
|
413 |
string Outputer::duration() |
414 |
{ |
415 |
ostringstream duration; |
416 |
|
417 |
duration.precision(2); |
418 |
duration << time; |
419 |
|
420 |
return duration.str(); |
421 |
} |
422 |
|
423 |
string Outputer::manycommon(vector<string> common) |
424 |
{ |
425 |
string line; |
426 |
|
427 |
for (int index = 0; index < common.size(); index++) |
428 |
{ |
429 |
line += common[index]; |
430 |
|
431 |
if (index != common.size() - 1) line += ' '; |
432 |
} |
433 |
|
434 |
return line; |
435 |
} |
436 |
|
437 |
void Outputer::tag(string& line, char* tag, const string& replacement) |
438 |
{ |
439 |
int begin = 0; |
440 |
while (begin < line.length()) |
441 |
{ |
442 |
int spot = line.find(tag, begin); |
443 |
|
444 |
if (spot != string::npos) |
445 |
{ |
446 |
line.replace(spot, strlen(tag), replacement); |
447 |
} |
448 |
else |
449 |
{ |
450 |
break; |
451 |
} |
452 |
|
453 |
begin = spot + replacement.length(); |
454 |
} |
455 |
} |
456 |
|
457 |
void Outputer::conditional(string& line, ifstream& fin, char* tag, bool |
458 |
condition) |
459 |
{ |
460 |
unsigned begin = 0; |
461 |
while (begin < line.length()) |
462 |
{ |
463 |
unsigned start = line.find(tag, begin); |
464 |
unsigned finish = line.find("<?endif?>", start); |
465 |
|
466 |
if (start == string::npos) break; |
467 |
|
468 |
string next; |
469 |
while (finish == string::npos) |
470 |
{ |
471 |
getline(fin, next); |
472 |
line += '\n' + next; |
473 |
finish = line.find("<?endif?>", start); |
474 |
} |
475 |
|
476 |
if (condition) |
477 |
{ |
478 |
line.erase(start, strlen(tag)); |
479 |
line.erase(finish - strlen(tag), 9); |
480 |
|
481 |
begin = finish - strlen(tag) - 9; |
482 |
} |
483 |
else |
484 |
{ |
485 |
line.erase(start, finish - start + 9); |
486 |
|
487 |
begin = start; |
488 |
} |
489 |
} |
490 |
} |