ViewVC Help
View File | Revision Log | Show Annotations | Download File | View Changeset | Root Listing
root/proj/trunk/Search/Outputer.cpp
Revision: 312
Committed: 2004-01-01T15:00:34-08:00 (21 years, 5 months ago) by douglas
File size: 12592 byte(s)
Log Message: 
Updated copyright years.

File Contents

# Content
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 Outputer
46 //
47 // Douglas Thrift
48 //
49 // $Id: Outputer.cpp,v 1.13 2004/01/01 23:00:34 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, '&', "&amp;");
95 entities(query, '\"', "&quot;");
96 entities(query, '<', "&lt;");
97 entities(query, '>', "&gt;");
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, '&', "&amp;");
167 entities(title, '\"', "&quot;");
168 entities(title, '<', "&lt;");
169 entities(title, '>', "&gt;");
170 }
171 string address = webpage.getURL();
172 string sample = webpage.getSample();
173 string description = webpage.getDescription();
174 ostringstream size;
175
176 size.precision(0);
177 size.setf(ios_base::fixed, ios_base::floatfield);
178 size << double(webpage.getSize()) / double(1024) << "k";
179
180 entities(address, '&', "&amp;");
181 entities(address, '\"', "&quot;");
182 entities(address, '<', "&lt;");
183 entities(address, '>', "&gt;");
184
185 ifstream fin(bodyFile.c_str());
186
187 string line;
188 while (fin.good())
189 {
190 getline(fin, line);
191
192 conditional(line, fin, "<?ifdescription?>", description != "");
193
194 tag(line, "<?address?>", address);
195 tag(line, "<?title?>", title);
196 tag(line, "<?sample?>", sample);
197 tag(line, "<?description?>", description);
198 tag(line, "<?size?>", size.str());
199
200 cout << line << (fin.good() ? "\n" : "");
201 }
202
203 fin.close();
204 }
205 }
206
207 void Outputer::footer(const string& query, unsigned page, vector<string>
208 common, bool and_, bool or_, const string& ignore)
209 {
210 ifstream fin(footerFile.c_str());
211
212 string line;
213 while (fin.good())
214 {
215 getline(fin, line);
216
217 conditional(line, fin, "<?ifquery?>", this->query);
218 conditional(line, fin, "<?ifresults?>", results);
219 conditional(line, fin, "<?ifor?>", or_);
220 conditional(line, fin, "<?ifand?>", and_);
221 conditional(line, fin, "<?ifignore?>", ignore != "");
222 conditional(line, fin, "<?ifcommon?>", common.size() == 1);
223 conditional(line, fin, "<?ifmanycommon?>", common.size() > 1);
224
225 #ifndef _OpenSSL_
226 tag(line, "<?version?>", programName + ' ' + programVersion + ' ' +
227 platform());
228 #else
229 tag(line, "<?version?>", programName + ' ' + programVersion + ' ' +
230 platform() + ' ' + openssl());
231 #endif
232 tag(line, "<?query?>", query);
233 tag(line, "<?range?>", range(page));
234 tag(line, "<?total?>", total());
235 tag(line, "<?time?>", duration());
236 tag(line, "<?pages?>", pages(query, page));
237 tag(line, "<?ignore?>", ignore);
238 tag(line, "<?common?>", common[0]);
239 tag(line, "<?manycommon?>", manycommon(common));
240
241 cout << line << (fin.good() ? "\n" : "");
242 }
243
244 fin.close();
245 }
246
247 void Outputer::notfound(const string& query, unsigned keywords)
248 {
249 ifstream fin(notfoundFile.c_str());
250
251 string line;
252 while (fin.good())
253 {
254 getline(fin, line);
255
256 conditional(line, fin, "<?ifmany?>", keywords > 1);
257
258 tag(line, "<?query?>", query);
259
260 cout << line << (fin.good() ? "\n" : "");
261 }
262
263 fin.close();
264 }
265
266 string Outputer::pages(string query, unsigned page)
267 {
268 entities(query, "&lt;", '<');
269 entities(query, "&gt;", '>');
270 entities(query, "&quot;", '\"');
271 entities(query, "&amp;", '&');
272
273 entities(query, '%', "%25");
274 entities(query, '\t', "%09");
275 entities(query, ' ', "%20");
276 entities(query, '\"', "%22");
277 entities(query, '#', "%23");
278 entities(query, '$', "%24");
279 entities(query, '&', "%26");
280 entities(query, '\'', "%27");
281 entities(query, '+', "%2B");
282 entities(query, ',', "%2C");
283 entities(query, '/', "%2F");
284 entities(query, ':', "%3A");
285 entities(query, ';', "%3B");
286 entities(query, '<', "%3C");
287 entities(query, '=', "%3D");
288 entities(query, '>', "%3E");
289 entities(query, '?', "%3F");
290 entities(query, '@', "%40");
291 entities(query, '[', "%5B");
292 entities(query, ']', "%5D");
293 entities(query, '\\', "%5C");
294 entities(query, '^', "%5E");
295 entities(query, '`', "%60");
296 entities(query, '{', "%7B");
297 entities(query, '|', "%7C");
298 entities(query, '}', "%7D");
299 entities(query, '~', "%7E");
300
301 string lines;
302
303 ifstream fin(pagesFile.c_str());
304
305 string line;
306 while (fin.good())
307 {
308 getline(fin, line);
309 conditional(line, fin, "<?ifprevious?>", page >= 1);
310 conditional(line, fin, "<?ifpage?>", false);
311 conditional(line, fin, "<?ifnum?>", false);
312 conditional(line, fin, "<?ifnext?>", false);
313
314 ostringstream previous;
315
316 previous << page;
317
318 tag(line, "<?query?>", query);
319 tag(line, "<?previous?>", previous.str());
320
321 lines += line + (fin.good() ? "\n" : "");
322 }
323
324 fin.close();
325 fin.clear();
326
327 for (int index = 0; index < numPages; index++)
328 {
329 fin.open(pagesFile.c_str());
330
331 while (fin.good())
332 {
333 getline(fin, line);
334 if (index == page)
335 {
336 conditional(line, fin, "<?ifprevious?>", false);
337 conditional(line, fin, "<?ifpage?>", true);
338 conditional(line, fin, "<?ifnum?>", false);
339 conditional(line, fin, "<?ifnext?>", false);
340
341 ostringstream current;
342
343 current << index + 1;
344
345 tag(line, "<?page?>", current.str());
346 }
347 else
348 {
349 conditional(line, fin, "<?ifprevious?>", false);
350 conditional(line, fin, "<?ifpage?>", false);
351 conditional(line, fin, "<?ifnum?>", true);
352 conditional(line, fin, "<?ifnext?>", false);
353
354 ostringstream num;
355
356 num << index + 1;
357
358 tag(line, "<?query?>", query);
359 tag(line, "<?num?>", num.str());
360 }
361
362 lines += line + (fin.good() ? "\n" : "");
363 }
364
365 fin.close();
366 fin.clear();
367 }
368
369 fin.open(pagesFile.c_str());
370
371 while (fin.good())
372 {
373 getline(fin, line);
374 conditional(line, fin, "<?ifprevious?>", false);
375 conditional(line, fin, "<?ifpage?>", false);
376 conditional(line, fin, "<?ifnum?>", false);
377 conditional(line, fin, "<?ifnext?>", page + 2 <= numPages);
378
379 ostringstream next;
380
381 next << page + 2;
382
383 tag(line, "<?query?>", query);
384 tag(line, "<?next?>", next.str());
385
386 lines += line + (fin.good() ? "\n" : "");
387 }
388
389 fin.close();
390
391 return lines;
392 }
393
394 string Outputer::range(unsigned page)
395 {
396 unsigned bottom = page * 10 + 1;
397 unsigned top = numWebpages > page * 10 + 10 ? page * 10 + 10 : numWebpages;
398 ostringstream range;
399
400 range << bottom << " - " << top;
401
402 return range.str();
403 }
404
405 string Outputer::total()
406 {
407 ostringstream total;
408
409 total << numWebpages;
410
411 return total.str();
412 }
413
414 string Outputer::duration()
415 {
416 ostringstream duration;
417
418 duration.precision(2);
419 duration.setf(ios_base::fixed, ios_base::floatfield);
420 duration << time;
421
422 return duration.str();
423 }
424
425 string Outputer::manycommon(vector<string> common)
426 {
427 string line;
428
429 for (int index = 0; index < common.size(); index++)
430 {
431 line += common[index];
432
433 if (index != common.size() - 1) line += ' ';
434 }
435
436 return line;
437 }
438
439 void Outputer::tag(string& line, char* tag, const string& replacement)
440 {
441 int begin = 0;
442 while (begin < line.length())
443 {
444 int spot = line.find(tag, begin);
445
446 if (spot != string::npos)
447 {
448 line.replace(spot, strlen(tag), replacement);
449 }
450 else
451 {
452 break;
453 }
454
455 begin = spot + replacement.length();
456 }
457 }
458
459 void Outputer::conditional(string& line, ifstream& fin, char* tag, bool
460 condition)
461 {
462 unsigned begin = 0;
463 while (begin < line.length())
464 {
465 unsigned start = line.find(tag, begin);
466 unsigned finish = line.find("<?endif?>", start);
467
468 if (start == string::npos) break;
469
470 string next;
471 while (finish == string::npos)
472 {
473 getline(fin, next);
474 line += '\n' + next;
475 finish = line.find("<?endif?>", start);
476 }
477
478 if (condition)
479 {
480 line.erase(start, strlen(tag));
481 line.erase(finish - strlen(tag), 9);
482
483 begin = finish - strlen(tag) - 9;
484 }
485 else
486 {
487 line.erase(start, finish - start + 9);
488
489 begin = start;
490 }
491 }
492 }